From: Andrei Heilper Newsgroups: alt.folklore.computers Subject: How was a ROM constructed before the transistor era? Date: Sun, 13 Dec 1998 14:52:39 +0200 Organization: IBM Lines: 9 Message-ID: <3673B896.F2C58E2B@vnet.ibm.com> NNTP-Posting-Host: gvul.haifa.il.ibm.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: news.ibm.net.il!ibm.net 913553628 13058 (none) 192.115.216.67 X-Complaints-To: postmaster@ibm.net X-Mailer: Mozilla 4.05 [en] (Win95; I) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newsfeed.uk.ibm.net!news.ibm.net.il!ibm.net!not-for-mail I remember that in the old days, in order to start a minicomputer you had to set some switches on the panel. This sequence of bits represented a very short loader (by the way, these were tricky self-modifying programs, and it seems quite interesting to have a collection of them). My question is if it was another solution for larger computers? Andrei Heilper ###### From: "Carl R. Friend" Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Sun, 13 Dec 1998 09:21:09 -0500 Organization: as little as possible! Lines: 35 Message-ID: <3673CD55.2D800B7E@stoneweb.com> References: <3673B896.F2C58E2B@vnet.ibm.com> NNTP-Posting-Host: zephyr.ultranet.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Complaints-To: abuse@ultra.net X-Ultra-Time: 13 Dec 1998 14:21:10 GMT X-Mailer: Mozilla 2.0 (X11; I; Linux 2.0.29 i586) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!18.24.4.11!newsswitch.lcs.mit.edu!news.ultranet.com!not-for-mail Andrei Heilper wrote: > > I remember that in the old days, in order to start a minicomputer you > had to set some switches on the panel. [...] My question is if it > was another solution for larger computers? ROMs existed for quite a while before the ROM "chip" of today was developed. At the simplest, a ROM can be an array of switches of which rows are read in sequence. Some computers used this technique for the initial program load (IPL) sequence. Diode arrays were also used and were programmed by snipping a diode from the array. Boards of this ilk were used in DEC's PDP-8 machine, for instance. Core rope, as used in the Apollo guidance and navigation computers, is another approach. Core rope works like a bunch of 1-turn pulse transformers, and how the rope is "strung" (drive wire inside of, or outside of the core) determines whether the bit is a 1 or a 0. Early Interdata computers used a variation of this for their microcode store. Core memory, itself, can be used as a ROM if you simply don't write to it. Core is non-volatile so whatever you put into it stays there while the power's off. -- ______________________________________________________________________ | | | | Carl Richard Friend (UNIX Sysadmin) | West Boylston | | Minicomputer Collector / Enthusiast | Massachusetts, USA | | mailto:carl.friend@stoneweb.com | | | http://www.ultranet.com/~crfriend/museum | ICBM: N42:22 W71:47 | |________________________________________________|_____________________| ###### From: "Robert Schuldenfrei" Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Sun, 13 Dec 1998 09:45:41 -0500 Organization: The Internet Access Company, Inc. Lines: 23 Message-ID: <750jro$1o2@news-central.tiac.net> References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> NNTP-Posting-Host: sailboat.tiac.net X-Newsreader: Microsoft Outlook Express 4.72.3110.1 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-feed1.tiac.net!posterchild2!news@tiac.net In fact IIRC when parts of the Challenger were brought up from the ocean, the programs and data in core memory were still in tact and could be read. Talk about the ultimate black box:) Robert Schuldenfrei - E-Mail: bob@s-i-inc.com S. I. Inc. - Voice: (781) 329-4828 32 Ridley Road - FAX: (781) 329-1696 Dedham, MA 02026 - http://www.tiac.net/users/tangaroa Carl R. Friend wrote in message <3673CD55.2D800B7E@stoneweb.com>... > Core rope, as used in the Apollo guidance and navigation computers, >is another approach. Core rope works like a bunch of 1-turn pulse >transformers, and how the rope is "strung" (drive wire inside of, or >outside of the core) determines whether the bit is a 1 or a 0. Early >Interdata computers used a variation of this for their microcode store. > > Core memory, itself, can be used as a ROM if you simply don't >write to it. Core is non-volatile so whatever you put into it stays >there while the power's off. > ###### From: jeverett@wwa.DEFEAT.UCE.BOTS.com (John Everett) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 13 Dec 1998 16:04:22 GMT Organization: Everett Associates Lines: 15 Message-ID: <750oi6$ccj$2@hirame.wwa.com> References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> NNTP-Posting-Host: poolf2-020.wwa.com Mime-Version: 1.0 Content-Type: Text/Plain; charset=US-ASCII X-Newsreader: WinVN 0.99.8 (x86 32bit) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.cwix.com!207.241.0.194!news.wwa.com!not-for-mail In article <3673CD55.2D800B7E@stoneweb.com>, carl.friend@stoneweb.com says... > > Core memory, itself, can be used as a ROM if you simply don't >write to it. Core is non-volatile so whatever you put into it stays >there while the power's off. True, in fact the Honeywell H-200's (circa mid-'60s) control memory was core ROM. -- jeverett@wwa.DEFEAT.UCE.BOTS.com (John Everett) http://www.wwa.com/~jeverett ^^^^^^^^^^^^^^^ Things have gotten so bad I feel the need to disguise my email address. And I don't like this explanation because I just hate long signatures. ###### From: hancock4@bbs.cpcn.com (Lisa or Jeff) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 13 Dec 1998 17:44:14 GMT Organization: Net Access BBS Lines: 18 Message-ID: <750ude$8va@netaxs.com> NNTP-Posting-Host: bbs.cpcn.com Originator: root@bbs.cpcn.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news-peer.gip.net!news.gsl.net!gip.net!news-xfer.newsread.com!netaxs.com!newsread.com!netaxs.newsread.com!bbs.cpcn.com!root > I remember that in the old days, in order to start a minicomputer you > had to set some switches on the panel. This sequence of bits > represented a very short loader (by the way, these were tricky > self-modifying programs, and it seems quite interesting to have a > collection of them). My question is if it was another solution for > larger computers? I believe on the older mainframes there existed a hardware read function to bring in a record from a card reader (or later disk drive); a process known as "IPL or Initial Program Load". That first record would initiate the loading sequence of the control program. On the S/360, we merely had to dial in the address of our IPL source. There were three rotary dials to provide a three character hex address (normally these were left in one spot since we usually loaded from our system disk; but they could be switched to the card reader). We pressed LOAD and the system did the rest, loading and starting up the operating system. ###### Reply-To: "Dennis J. Minette" From: "Dennis J. Minette" References: <3673B896.F2C58E2B@vnet.ibm.com> Subject: Re: How was a ROM constructed before the transistor era? Date: Sun, 13 Dec 1998 18:50:56 -0500 Lines: 15 Organization: Minette Data Systems, Inc. X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-Mimeole: Produced By Microsoft MimeOLE V4.72.3155.0 Message-ID: Newsgroups: alt.folklore.computers NNTP-Posting-Host: 1Cust115.tnt2.sarasota.fl.da.uu.net [208.252.44.115] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!207.68.152.14!upnetnews04!upnetnews03 On Honeywell 200 circa 1970: Put bootstrap card in reader and hit switches (boot, boot, run, run, run.) Andrei Heilper wrote in message <3673B896.F2C58E2B@vnet.ibm.com>... >I remember that in the old days, in order to start a minicomputer you >had to set some switches on the panel. This sequence of bits >represented a very short loader (by the way, these were tricky >self-modifying programs, and it seems quite interesting to have a >collection of them). My question is if it was another solution for >larger computers? > >Andrei Heilper > ###### From: WStreett@shell.monmouth.com (Wilbur Streett) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Sun, 13 Dec 1998 19:13:52 GMT Organization: Monmouth Internet Lines: 28 Message-ID: <3674100e.220383965@news.monmouth.com> References: <750ude$8va@netaxs.com> NNTP-Posting-Host: pa-tc-ppp34.monmouth.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newspeer.monmouth.com!news.monmouth.com!not-for-mail I'm amazed that people don't know this one.. The first time that I got to see a computer, in 1967, my father showed me the different strips of paper tape that they had setup to initialize the computer with different hardware configurations. They would load the paper tape into the system default reader and when they powered on the system it would load and execute whatever was on the paper tape by default. The paper tape would end up on the floor, and they would pick it up and tape it to the wall for the next time that they would startup the system. Eventually they figured out the dirt on the floor was messing up the tape reader, and they figured out that loading the paper tape for every load meant that they had to walk across the room and would take more time. They then looped the paper tape by taping the beginning to the end, and the loop sat on the tape reader since the tape had pretty much be replaced by the magnetic tape and cards for I/O. They also figured out that reconfiguring the system required quite a bit of manual intervention, so they would configure the system with all the appropriate devices and leave them all connected even if they were only doing a card sort. Eventually, the loop of paper tape, which looked like a loop the the back of a pair of military boots, was identified with the process of putting on the boots in the morning, or "picking yourself up by your own bootstraps".. or more simply "booting up" The Papertape was Read Only Memory or ROM, and we still "boot" computers up when we start them. Wilbur ###### From: "Joel C. Ewing" Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Sun, 13 Dec 1998 20:32:01 -0600 Content-Transfer-Encoding: 7bit References: <3673B896.F2C58E2B@vnet.ibm.com> X-Posted-Path-Was: not-for-mail Content-Type: text/plain; charset=us-ascii X-ELN-Date: 14 Dec 1998 03:12:54 GMT X-ELN-Insert-Date: Sun Dec 13 19:15:10 1998 Organization: EarthLink Network, Inc. Lines: 62 Mime-Version: 1.0 NNTP-Posting-Host: 2cust11.tnt8.dfw5.da.uu.net Message-ID: <367478A1.B588E1A2@acm.org> X-Mailer: Mozilla 4.05 [en] (Win95; U) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.berkeley.edu!newsfeed1.earthlink.net!nntp.earthlink.net!posted-from-earthlink!not-for-mail Andrei Heilper wrote: > I remember that in the old days, in order to start a minicomputer you > had to set some switches on the panel. This sequence of bits > represented a very short loader (by the way, these were tricky > self-modifying programs, and it seems quite interesting to have a > collection of them). My question is if it was another solution for > larger computers? > Andrei Heilper The IBM 360 and successor families typically (always?) operated at two different levels: (1) the real physical hardware level, which typically had little resemblance to the 360 architecture known by the customer, and (2) the 360 architecture level, the view seen by the customer, with all the standard 360 instruction and "hardware" register behavior. The bridge between these two levels was "microprogramming": a highly specialized form of programming, unique to a specific processor model, which instructed all the gates and registers of the real hardware how to interact to produce the effect of a machine capable of satisfying the external behavior of a "360". This separation of the physical structure from the architecture seen by application code is one of the reasons for the longevity of the 360 family and successors, as it allows for drastic changes and improvements in the physical implementation of the hardware and the use of increased bus sizes and higher parallelism for faster (& more expensive) models, all while providing the customer with a machine which satisfies the same architecture (machine instruction) definitions. The microcode that made these machines function as a "360" was stored in Read Only Storage (called ROS, not ROM), and several different techniques were used: CCROS (Card Capacitor ROS, used in 360/30, where 720 bits were encoded via punches in a mylar IBM card); Balanced Capacitor ROS (models 50, 60, 65), with bits photo etched into copper-clad epoxy glass plates; Transformer ROS (TROS) (models 20, 40), in which bits were encoded by punching holes in "word lines" that were photo etched on flexible Mylar tapes. While this ROS could be replaced after construction, the process was non-trivial. Successors to the 360 family have taken a more flexible approach of relying instead on volatile high-speed memory that is initialized at hardware power-up by a process known as either "IML" (Initial Microcode Load) or more recently "Power-On Reset", during which the necessary microcode is downloaded from a separate support processor, which appears to retain this information in the form of readily changed files on some storage media (the internals of this support processor are not part of the defined user interface). On the 360/370/390 architecture level there is no read-only storage involved. The architecture definition provides for an IPL (Initial Program Load) sequence in which a standard I/O channel read from an operator specified external device is initiated to read data into a fixed location in memory followed a startup of a central processor on the resulting data. The resulting bootstrap process eventually results in the loading and initialization of the entire operating system. As an aside, the ENIAC had read-only storage as early as 1943, implemented as Resistive Read-Only Storage Arrays (arrays of resistors), for the storage of functional tables. Historical source: Microprogramming: Principles and Practices, Samir S. Husson, Prentice Hall 1970. -- Joel C. Ewing, Fort Smith, AR jcewing@acm.org ###### From: johnl@radix.net (John A. Limpert) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Sun, 13 Dec 1998 20:39:34 GMT Organization: RadixNet Internet Services Lines: 21 Message-ID: <367523f7.103418437@news1.radix.net> References: <3673B896.F2C58E2B@vnet.ibm.com> NNTP-Posting-Host: port38.annex4.radix.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!ptdnetP!newsgate.ptd.net!news1.radix.net!not-for-mail On Sun, 13 Dec 1998 14:52:39 +0200, Andrei Heilper wrote: >I remember that in the old days, in order to start a minicomputer you >had to set some switches on the panel. This sequence of bits >represented a very short loader (by the way, these were tricky >self-modifying programs, and it seems quite interesting to have a >collection of them). My question is if it was another solution for >larger computers? Some old DEC systems used diode matrix ROM boards that were programmed with wire cutters. I've used/programmed several old computer systems that had paper tape readers with controllers that could independently DMA data into core. You would halt the main computer, read a program from paper tape into core, and restart the main computer at a fixed address. The paper tape (actually mylar tape) would contain the operating software or a bootstrap for a mag tape or other device. ###### Reply-To: "Dennis J. Minette" From: "Dennis J. Minette" References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 00:31:13 -0500 Lines: 32 Organization: Minette Data Systems, Inc. X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-Mimeole: Produced By Microsoft MimeOLE V4.72.3155.0 Message-ID: Newsgroups: alt.folklore.computers NNTP-Posting-Host: 1Cust164.tnt1.sarasota.fl.da.uu.net [153.37.162.164] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!207.68.152.14!upnetnews04!upnetnews03 The Bakers wrote in message <367485B4.367384D3@no.spam>... >"Dennis J. Minette" wrote: > >> On Honeywell 200 circa 1970: >> Put bootstrap card in reader and hit switches (boot, boot, run, run, run.) > >Following remarks reference a machine with console typewriter as opposed to a >"boot" switch. > >First you'd always hit the stop & initialize buttons. > >IIRC, your remarks sound like the tape boot.....B 40 000000 at the console >twice, and then hit run three times. Wasn't the intent to get past the tape >header ??? > >The card boot for an object deck w/loader was B 41 000317 and then hit run, >right ? I remember the address but don't remember why it was special. > >The disk boot was B 44 000000, wasn't it ? > >That was a great machine. Alwasy glad to see someone mention it here :-) > > I see you have a better memory of the H200 than I :-) I was more than happy to move on to the H6000 after HISI acquired the GE computer line. ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 00:47:54 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 77 Message-ID: <3674114f.1588945@news.vip.net> References: <3673B896.F2C58E2B@vnet.ibm.com> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.16 X-Trace: 913596175 A01OARAUVD410CCD1C usenet53.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!feeder.qis.net!newsfeed-east.supernews.com!supernews.com!Supernews69!not-for-mail Andrei Heilper wrote: >I remember that in the old days, in order to start a minicomputer you >had to set some switches on the panel. This sequence of bits >represented a very short loader (by the way, these were tricky >self-modifying programs, and it seems quite interesting to have a >collection of them). My question is if it was another solution for >larger computers? You'll see a bunch of other responses which I agree with. There could also be an instruction set designed to make it easy. Ever heard of CARDIAC? CARDIAC stands for "CARDboard Illustrative Aid to Computing". It was a model of a simple processor. The manual mentioned a bootstrap loader at address 00 which contains 001. The manual did not go into details. It took me years before I figured it out. Given execution starting at 00, a CARDIAC boot image would be: 002 800 zero or more occurrence of 0aa [address to load at] vvvv [value to load (possibly signed)] 8aa [aa=start address] I've tested this and it works. There is a loop from 00 to 02: input to 01, execute whatever that instruction is (0aa is input into aa), and loop back. If you load a jump instead of an input instruction, you get the jump to the starting address. On a simple program to add two numbers: aa Obj Instruction org 10 10: 020 start inp augend 11: 021 inp addend 12: 120 cla augend 13: 221 add addend 14: 622 sto sum 15: 522 out sum 16: 900 hrs 00 augend ds 1 addend ds 1 end start The boot image is: 002 800 010 020 011 021 012 120 013 221 014 622 015 522 016 900 810 You can also hack this bootstrap slightly if tight for space. If execution starts at 02, you have 02's instruction be loaded last. Instead of looping after the load, the program starts. Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 03:20:07 GMT Organization: Daedalus Consulting Lines: 23 Message-ID: <752057$pnc$1@news.wlg.netlink.net.nz> References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-dc.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Carl R. Friend wrote: > Core memory, itself, can be used as a ROM if you simply don't >write to it. Core is non-volatile so whatever you put into it stays >there while the power's off. This isn't actually a good idea, because to read from core memory you in fact have to erase it. A core read sequence goes: 1: Saturate core in "zero" direction 2: If core was in one state, a pulse is induced into the sense wire as it saturates 3: If pulse received, re-saturate core in "one" direction. If power is lost between steps 1 and 3, you lose all the one bits in the word being accessed. But there's nothing wrong with toggling the primary boot into core once and leaving it there. If it gets stomped on by a power failure during booting, you just toggle it in again. You just don't want to be entirely reliant on a core-resident bootstrap. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: The Bakers Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 03:28:08 GMT Organization: AT&T WorldNet Services Lines: 22 Message-ID: <367485B4.367384D3@no.spam> References: <3673B896.F2C58E2B@vnet.ibm.com> Reply-To: donteventry@no.spam NNTP-Posting-Host: 208.250.113.212 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.5 [en] (Win95; I) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!wn3feed!worldnet.att.net!135.173.83.225!attworldnet!newsadm "Dennis J. Minette" wrote: > On Honeywell 200 circa 1970: > Put bootstrap card in reader and hit switches (boot, boot, run, run, run.) Following remarks reference a machine with console typewriter as opposed to a "boot" switch. First you'd always hit the stop & initialize buttons. IIRC, your remarks sound like the tape boot.....B 40 000000 at the console twice, and then hit run three times. Wasn't the intent to get past the tape header ??? The card boot for an object deck w/loader was B 41 000317 and then hit run, right ? I remember the address but don't remember why it was special. The disk boot was B 44 000000, wasn't it ? That was a great machine. Alwasy glad to see someone mention it here :-) ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 07:00:59 +0000 Organization: Tanglewood Message-ID: <3674B7AB.3543FBDC@tnglwood.demon.co.uk> References: <3673B896.F2C58E2B@vnet.ibm.com> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913621300 nnrp-11:1404 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 22 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Andrei Heilper wrote: > collection of them). My question is if it was another solution for > larger computers? The IBM1130 used a "Cold Start" card which was read in by setting the DMA into a weird mode, in which one card was read into memory at a hard wired address. Basically the 12 rows on the card were wired to the 16 bus bits, in such a way the the 4 missing bits were either supplied as zeroes, or in one case by using a bit twice. This was IIRC the two bits which selected the index register, so you could either use index 0 or 3, but not 1 or 2. The cold start program could then read the OS boot stuff off a disk cartridge. The sequence was started by a large blue button engraved IPL, somewhere in the lower right of the panel. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: jtNOSPAM@epix.net (Julian Thomas) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 09:58:58 -0500 Organization: epix Internet Services Lines: 26 Message-ID: <36752844$1$wg$mr2ice@news.epix.net> References: <750ude$8va@netaxs.com> NNTP-Posting-Host: itha-125ppp32.epix.net X-Newsreader: MR/2 Internet Cruiser Edition for OS/2 v1.51 b51 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!news-xfer.epix.net!news1.epix.net!epix-news In <750ude$8va@netaxs.com>, on 12/13/98 at 05:44 PM, hancock4@bbs.cpcn.com (Lisa or Jeff) said: >On the S/360, we merely had to dial in the address of our IPL source. >There were three rotary dials to provide a three character hex address >(normally these were left in one spot since we usually loaded from our >system disk; but they could be switched to the card reader). We pressed >LOAD and the system did the rest, loading and starting up the operating >system. And everyone pressed System Reset before pressing Load - even though the first thing that the Load button did was to invoke the system reset function. It then went into a fairly lengthy microprogram that wrote a very short channel program starting at location 0 and then simulated a SIO to the lIPL device. (I wrote the load microprogram for the mod 50). -- Julian Thomas: jt at epix dot net http://home.epix.net/~jt Boardmember of POSSI.org - Phoenix OS/2 Society, Inc http://www.possi.org In the beautiful Finger Lakes Wine Country of New York State! -- -- Whaddya mean my computer is old -- my car is a lot older! ###### From: mwilson@interlog.com (Mel Wilson) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 11:56:38 -0500 Organization: Interlog Internet Services -Voice (416) 975-2655 -Data 515-1414 Lines: 24 Message-ID: References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> NNTP-Posting-Host: shell1.interlog.com NNTP-Posting-Time: 14 Dec 1998 16:56:40 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news.sprintlink.net!news.maxwell.syr.edu!rill.news.pipex.net!pipex!newsfeed.interlog.com!news.interlog.com!not-for-mail In article <367485B4.367384D3@no.spam>, The Bakers wrote: >"Dennis J. Minette" wrote: > >> On Honeywell 200 circa 1970: >> Put bootstrap card in reader and hit switches (boot, boot, run, run, run.) > >Following remarks reference a machine with console typewriter as opposed to a >"boot" switch. >First you'd always hit the stop & initialize buttons. >IIRC, your remarks sound like the tape boot.....B 40 000000 at the console >twice, and then hit run three times. Wasn't the intent to get past the tape >header ??? >The card boot for an object deck w/loader was B 41 000317 and then hit run, >right ? I remember the address but don't remember why it was special. >The disk boot was B 44 000000, wasn't it ? >That was a great machine. Alwasy glad to see someone mention it here :-) From hanging around with the Field Engineers, I seem to remember hearing that some of the H-200s used core rope for some internal logic, as well. (?) Regards. Mel. ###### From: jones@cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 15:58:39 GMT Organization: The University of Iowa Lines: 45 Message-ID: <753cjf$nog$1@flood.weeg.uiowa.edu> References: <750ude$8va@netaxs.com> NNTP-Posting-Host: pyrite.cs.uiowa.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!feeder1.wwnet.net!streamer1.cleveland.iagnet.net!NewsNG.Chicago.Qual.Net!news.uiowa.edu!not-for-mail From article <750ude$8va@netaxs.com>, by hancock4@bbs.cpcn.com (Lisa or Jeff): > On the S/360, we merely had to dial in the address of our IPL source. ... But the IPL process itself was not trivial, and it took a bit of ROM (in the microcode) to hold the IPL logic. IBM used a number of ROM technologies in the 360. The most interesting was the capacative ROM used in the 360/40. This used punched cards as the storage medium; specifically, mylar punched cards. The ROM words were 80 bits long, and there were 12 words per card. The readout was capacative, based on the fact that the dielectric constants of mylar and air differ considerably. So, to read a word, you'd apply a pulse to the conductor that ran along the word line, and see how well this coupled to each bit line. Strong coupling implies no hole implies a zero. Weak coupling implies a hole implies a one. The punched cards were held between two printed circuit boards, one with the (horizontal, relative to the card) word lines, and one with the (vertical) bit lines. The boards were pressed against the card by air pressure -- basically, an elastic bladder inflated behine one of the PC boards to hold it tight. Loss of air pressure could cause the machine to behave very strangely! This was one of the first machines to be user microprogrammed. After all, card punches were commonplace around computers, and all you needed to begin microprogramming experiments was a stock of mylar cardstock. But, the dominant early ROM technology involved diode matrices. This is the technology Maurice Wilkes assumed in his original work on microprogramming. You just string word wires horizontally on a frame and string bit wires vertically on the frame, then solder diodes from word wire to bit wire wherever you want a one (or zero, depenging on the sense of your logic). Later, by the late 1950's, pinboard diode matrices were in use. In these, the diodes were manufactured in the shape of pins that you would stuff into holes in the board. The tip of the pin made contact with a bit line while the tail of the pin contacted the word line. Standard microwave diode packages from as far back as World War II were, conveniently and accidentally, well suited to the requirements of such pinboard ROMs. The bit density on the pinboard was about 1 bit per 1/4 inch square. Doug Jones jones@cs.uiowa.edu ###### From: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 16:51:25 GMT Organization: The National Capital FreeNet Lines: 10 Message-ID: <753fmd$g13@freenet-news.carleton.ca> References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> Reply-To: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) NNTP-Posting-Host: freenet6.carleton.ca X-Given-Sender: ab528@freenet6.carleton.ca (Heinz W. Wiggeshoff) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newsfeed.cwix.com!198.138.0.5!newshub.northeast.verio.net!newsserver.jvnc.net!xcski.com!freenet-news.carleton.ca!FreeNet.Carleton.CA!ab528 Douglas W. Jones,201H MLH,3193350740,3193382879 (jones@cs.uiowa.edu) writes: [snip] > > This was one of the first machines to be user microprogrammed. After > all, card punches were commonplace around computers, and all you needed > to begin microprogramming experiments was a stock of mylar cardstock. Which made the -40 an ideal machine to try an APL processor. (And I'll bet that keypunching errors weren't cheap!) ###### Date: 14 Dec 98 17:16:56 -0800 From: "Charlie Gibbs" Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <3674B7AB.3543FBDC@tnglwood.demon.co.uk> <3675A44B.D0694F63@no.spam> Message-ID: <1303.652T2816T10365721@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 26 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 14 Dec 1998 17:50:42 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news-pen-3.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.106 In article <3675A44B.D0694F63@no.spam> donteventry@no.spam (The Bakers) writes: >Don't forget that a couple of the sense switches located along the >front of the console were used to select which disk drive got IPL'ed. >Never saw an 1130 with more than one disk drive, so always made sure >all the switches were off or the bootstrap would hang. But it provides a great opportunity for a practical joke. When I worked in a shop with a card-based Univac 9200, I once set the data entry switches (from which it got its IPL address) to point to the read/punch unit instead of the card reader. These switches were normally not used by anyone except a serviceman (or a hacker like me), so they were just left at the standard IPL address and people seldom even bothered looking at them. So the operator came along, dropped an object deck into the reader, and did the usual CLEAR/LOAD/RUN procedure. Nothing happened for a couple of seconds, because the read/punch's drive motor (which was fairly quiet) was starting up. Just about the time the operator began to suspect that something was wrong, the read/punch fed a card with a CLANK that was loud enough to startle him before he realized what was happening. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### Message-ID: <3675B940.C470D74F@jkmicro.com> From: Jim Stewart Organization: http://www.jkmicro.com X-Mailer: Mozilla 4.04 [en] (Win95; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <367587b9.20293582@news.prosurfr.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 23 NNTP-Posting-Host: 168.150.253.123 X-Trace: news13.ispnews.com 913684879 168.150.253.123 (Mon, 14 Dec 1998 20:21:19 EDT) NNTP-Posting-Date: Mon, 14 Dec 1998 20:21:19 EDT Date: Mon, 14 Dec 1998 17:20:00 -0800 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!hub1.ispnews.com!news13.ispnews.com.POSTED!not-for-mail John Savard wrote: > > Andrei Heilper wrote, in part: > > >I remember that in the old days, in order to start a minicomputer you > >had to set some switches on the panel. This sequence of bits > >represented a very short loader (by the way, these were tricky > >self-modifying programs, and it seems quite interesting to have a > >collection of them). My question is if it was another solution for > >larger computers? > > You could get a ROM for the PDP-8 which was a big printed circuit > board with one diode soldered in for each bit set a certain way (no, I > don't know if it was 1 or 0). > > And some versions of the IBM 360 used, for their microprogram ROM, a > kind of punched card stuck between two sets of metal strips, one down, > one across. Didn't the early ESS central office switches have a really perverse memory composed of aluminium plates with ferrite chips embedded in them? I seem to recall that the plates could be written in an external device then installed in the switch where they were read-only. ###### Message-ID: <3675DD2E.37EA@pacbell.net> From: John Mullan X-Mailer: Mozilla 2.02 (OS/2; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 42 Date: Mon, 14 Dec 1998 19:53:18 -0800 NNTP-Posting-Host: 207.214.148.114 X-Complaints-To: abuse@pacbell.net X-Trace: typhoon-sf.snfc21.pbi.net 913694133 207.214.148.114 (Mon, 14 Dec 1998 19:55:33 PDT) NNTP-Posting-Date: Mon, 14 Dec 1998 19:55:33 PDT Organization: SBC Internet Services Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!cyclone.swbell.net!typhoon-sf.snfc21.pbi.net.POSTED!not-for-mail The Bakers wrote: > > "Dennis J. Minette" wrote: > > > On Honeywell 200 circa 1970: > > Put bootstrap card in reader and hit switches (boot, boot, run, run, run.) > > Following remarks reference a machine with console typewriter as opposed to a > "boot" switch. > > First you'd always hit the stop & initialize buttons. > > IIRC, your remarks sound like the tape boot.....B 40 000000 at the console > twice, and then hit run three times. Wasn't the intent to get past the tape > header ??? > > The card boot for an object deck w/loader was B 41 000317 and then hit run, > right ? I remember the address but don't remember why it was special. > > The disk boot was B 44 000000, wasn't it ? > > That was a great machine. Alwasy glad to see someone mention it here :-) Ahhh, the good old days. My first job out of college was as a computer operator (trainee) in Honeywell's Technical Computer Center in Waltham, where we had 1 (or more) each of the various series 200/2000 machines. Funny, with PC's, they get moved all the time, but I remember a 110 that was used as a print machine that was relocated from one end of the computer room to the other and it took 2 or more weeks to get it back up and running. One of the machines had a special console that was done for the Air Force where none of the push buttons functioned, but were all done through the console, and were logged on the flexo. I recently found my notebook with all my notes on running the different machines and operating systems. Mod I, Mod I DM, Mod II, Mod IV, OS/200 (and OS/2000), HOS, each one was different, using different jcl and keyboard commands. Talk about noisy... eeriest silence I ever experienced was when everything was powered down. When the printers started up and the noise level went up even further. jm ###### From: jcmorris@mwunix.mitre.org (Joe Morris) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 20:24:12 GMT Organization: The MITRE Corporation Lines: 25 Message-ID: <753s5c$6e9@top.mitre.org> References: <750ude$8va@netaxs.com> <36752844$1$wg$mr2ice@news.epix.net> NNTP-Posting-Host: mwunix.mitre.org Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news.sprintlink.net!news.maxwell.syr.edu!newspeer1.nac.net!logbridge.uoregon.edu!news.fas.harvard.edu!news.tufts.edu!blanket.mitre.org!news.mitre.org!mwunix!jcmorris jtNOSPAM@epix.net (Julian Thomas) writes: >And everyone pressed System Reset before pressing Load - even though the >first thing that the Load button did was to invoke the system reset >function. It then went into a fairly lengthy microprogram that wrote a >very short channel program starting at location 0 and then simulated a SIO >to the lIPL device. >(I wrote the load microprogram for the mod 50). Did the /50 IPL code actually place the CCW in memory at absolute zero? I know that the /65 did not, but the /40 did -- and there was an FE tip that explained how to rewind a tape from the console: you set up a ROS stop at such-and-such an address, set the IPL address switches to the address of the tape drive, pressed IPL, and when the ROS stop hung the microcode you then toggled the appropriate CCW opcode (0x07) into the byte at absolute zero and allowed the IPL microcode to continue. Admittedly, in most /40 shops it was much faster to get up, walk over to the tape drive, press RESET/REWIND/START, walk back to the console, and sit down -- but doing it all from the console was a nice hack. Joe Morris Joe Morris ###### From: jcmorris@mwunix.mitre.org (Joe Morris) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 20:28:44 GMT Organization: The MITRE Corporation Lines: 15 Message-ID: <753sds$6j0@top.mitre.org> References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> NNTP-Posting-Host: mwunix.mitre.org Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!logbridge.uoregon.edu!news.fas.harvard.edu!news.tufts.edu!blanket.mitre.org!news.mitre.org!mwunix!jcmorris jones@cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) writes: >IBM used a number of ROM technologies in the 360. The most interesting >was the capacative ROM used in the 360/40. This used punched cards as >the storage medium; specifically, mylar punched cards. [...] Um...CROS was used in the /50; the /40 used TROS. I've still got some old used TROS tapes (removed from a /40 when the floating-point architecture was chanted in 1968 (?), as well as a half-dozen or so blank tapes with a special punch tool intended to allow the CE to build a replacement TROS sheet in the field. TROS was also used in the 2314 disk controller. Joe Morris ###### From: spalding@iol.ie Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 20:40:35 GMT Organization: Deja News - The Leader in Internet Discussion Lines: 27 Message-ID: <753t43$ip7$1@nnrp1.dejanews.com> References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> <750oi6$ccj$2@hirame.wwa.com> NNTP-Posting-Host: 193.203.144.91 X-Article-Creation-Date: Mon Dec 14 20:40:35 1998 GMT X-Http-User-Agent: Mozilla/4.0 (compatible; MSIE 4.01; Windows 98) X-Http-Proxy: 1.1 x4.dejanews.com:80 (Squid/1.1.22) for client 193.203.144.91 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!netnews.com!newspeer1.nac.net!news.maxwell.syr.edu!nntp2.dejanews.com!nnrp1.dejanews.com!not-for-mail In article <750oi6$ccj$2@hirame.wwa.com>, jeverett@wwa.DEFEAT.UCE.BOTS.com (John Everett) wrote: > In article <3673CD55.2D800B7E@stoneweb.com>, carl.friend@stoneweb.com says... > > > > Core memory, itself, can be used as a ROM if you simply don't > >write to it. Core is non-volatile so whatever you put into it stays > >there while the power's off. > > True, in fact the Honeywell H-200's (circa mid-'60s) control memory was core > ROM. In 1963/4, just pre-360, I met up with an IBM 7740 which was a comms controller. It had about 100K bytes of memory and was normally attached to a host machine. It arrived from Poughkeepsie with all the diagnostics in core so that you could check it out without having the host interface working. Bytes weren't called bytes in that machine they were just 8-bit characters, and hex hadn't been heard of either so the content was described in octal, normally in 377 format but for certain 'bytes' in either an instruction or some control word 773 was used because that's how the functions of the bits in it broke down. -- Nick Spalding -----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 21:00:11 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 19 Message-ID: <3675635e.4102963@news.vip.net> References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.39 X-Trace: 913669071 A01OARAUVD427CCD1C usenet87.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!Supernews73!supernews.com!Supernews69!not-for-mail mwilson@interlog.com (Mel Wilson) wrote: [snip] > From hanging around with the Field Engineers, I seem to remember >hearing that some of the H-200s used core rope for some internal logic, >as well. (?) What is "core rope", please? Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### From: arargh@arargh.com (Arargh!) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 21:33:07 GMT Organization: Arargh!! Lines: 17 Message-ID: <367578e3.225001737@news.mcs.net> References: <3673B896.F2C58E2B@vnet.ibm.com> <367478A1.B588E1A2@acm.org> Reply-To: arargh@arargh.com NNTP-Posting-Host: jwright.pr.mcs.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!schenectady.netmonger.net!news.mcs.net!ddsw1!news.mcs.net!not-for-mail On Sun, 13 Dec 1998 20:32:01 -0600, "Joel C. Ewing" wrote: >On the 360/370/390 architecture level there is no read-only storage >involved. The architecture definition provides for an IPL (Initial >Program Load) sequence in which a standard I/O channel read from an >operator specified external device is initiated to read data into a >fixed location in memory followed a startup of a central processor on >the resulting data. The resulting bootstrap process eventually results >in the loading and initialization of the entire operating system. On the few S/370's that I saw, they IPML'd from an 8" floppy. >Historical source: Microprogramming: Principles and Practices, Samir S. >Husson, Prentice Hall 1970. > ###### From: jsavard@tenMAPSONeerf.edmonton.ab.ca (John Savard) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 21:50:33 GMT Organization: Videotron Communications Ltd. Lines: 19 Message-ID: <367587b9.20293582@news.prosurfr.com> References: <3673B896.F2C58E2B@vnet.ibm.com> NNTP-Posting-Host: c9169-002.prosurfr.com X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newshub.northeast.verio.net!cyclone.news.idirect.com!island.idirect.com!news1.tor.metronet.ca!news1.cal.metronet.ca!news1.van.metronet.ca!news.insinc.net!stimpy.cal.sfl.net!news.videotron.ab.ca!not-for-mail Andrei Heilper wrote, in part: >I remember that in the old days, in order to start a minicomputer you >had to set some switches on the panel. This sequence of bits >represented a very short loader (by the way, these were tricky >self-modifying programs, and it seems quite interesting to have a >collection of them). My question is if it was another solution for >larger computers? You could get a ROM for the PDP-8 which was a big printed circuit board with one diode soldered in for each bit set a certain way (no, I don't know if it was 1 or 0). And some versions of the IBM 360 used, for their microprogram ROM, a kind of punched card stuck between two sets of metal strips, one down, one across. John Savard http://www.freenet.edmonton.ab.ca/~jsavard/index.html ###### From: bill_h Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Mon, 14 Dec 1998 23:31:06 -0800 Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 7 Message-ID: <3676103A.2B68@sunsouthwest.com> References: <3673B896.F2C58E2B@vnet.ibm.com> Reply-To: bill_h@sunsouthwest.com NNTP-Posting-Host: 169.197.40.91 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: 913703933 W4NEQTDLS285BA9C5C usenet87.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Mailer: Mozilla 3.04Gold (Win16; I) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.direct.ca!Supernews73!supernews.com!Supernews69!not-for-mail Early ROM? Plug Boards. Full of either actual jumper wire or just pins that made connections in a matrix. And, you could swap them in and out quickly. ###### From: The Bakers Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 14 Dec 1998 23:50:56 GMT Organization: AT&T WorldNet Services Lines: 23 Message-ID: <3675A44B.D0694F63@no.spam> References: <3673B896.F2C58E2B@vnet.ibm.com> <3674B7AB.3543FBDC@tnglwood.demon.co.uk> Reply-To: donteventry@no.spam NNTP-Posting-Host: 208.250.113.192 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.5 [en] (Win95; I) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!dca1-hub1.news.digex.net!digex!cpk-news-hub1.bbnplanet.com!news.gtei.net!wn3feed!worldnet.att.net!135.173.83.225!attworldnet!newsadm Robert Billing wrote: (referring to the IBM 1130) > The sequence was started by a large blue button engraved IPL, somewhere > in the lower right of the panel. Not to nitpick [OK, I'll nitpick] but the button was labeled "Program Load". I believe it was indeed blue rectangular plastic --- same size as the "standard" IBM buttons found on a card reader, printer, etc. of that era....1" x 1.5" or so. The cold start sequence was "IMMediate STOP", "RESET", and then "PROGRAM LOAD". Don't forget that a couple of the sense switches located along the front of the console were used to select which disk drive got IPL'ed. Never saw an 1130 with more than one disk drive, so always made sure all the switches were off or the bootstrap would hang. Another great machine of its day (emotional attachment-wise, if not technologically speaking)......... ###### From: The Bakers Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 15 Dec 1998 00:15:52 GMT Organization: AT&T WorldNet Services Lines: 38 Message-ID: <3675AA23.F56BDEA6@no.spam> References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> <753fmd$g13@freenet-news.carleton.ca> Reply-To: donteventry@no.spam NNTP-Posting-Host: 208.250.113.252 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.5 [en] (Win95; I) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!wn4feed!worldnet.att.net!135.173.83.225!attworldnet!newsadm "Heinz W. Wiggeshoff" wrote: > Douglas W. Jones,201H MLH,3193350740,3193382879 (jones@cs.uiowa.edu) writes: > [snip] > > > > This was one of the first machines to be user microprogrammed. After > > all, card punches were commonplace around computers, and all you needed > > to begin microprogramming experiments was a stock of mylar cardstock. > > Which made the -40 an ideal machine to try an APL processor. > (And I'll bet that keypunching errors weren't cheap!) Some quick questions on this --- Did anyone ever actually microcode an APL processor and implement it ? Were the microcode changes just to add some machine instructions which were especially well suited for APL operation ? What other implementations were done in user microcode ? How much capacity was in the microcode store ? Could you add entirely new instructions, or were you limited to modifying existing ones ? If you added a new instruction, did you have to give up one of the original ones to make room ? Was the microcode "hand assembled" (for lack of a better term; you know what I mean) or was there software which let you microcode in a symbolic language ? I remember someone saying that IBM supplied the microcode listings for the original machine. Did they encourage user microcode ? Tolerate it ? Give you a hard time if you wrote your own ? Were there IBM manuals describing how to microcode the processor ? I would imagine that IBM made you put the original microcode back before the serviceman would work on the machine, no ? Are there any references in the literature to user microprogramming the S/360 or its cousins ? Thanks for any information on this interesting topic. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 06:47:08 +0000 Organization: Tanglewood Message-ID: <367605EB.9FEB8D8B@tnglwood.demon.co.uk> References: <3673B896.F2C58E2B@vnet.ibm.com> <3674B7AB.3543FBDC@tnglwood.demon.co.uk> <3675A44B.D0694F63@no.spam> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913706349 nnrp-11:9188 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 23 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!newsgate.cistron.nl!het.net!bullseye.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail The Bakers wrote: > Not to nitpick [OK, I'll nitpick] but the button was labeled "Program > Load". I believe it was indeed blue rectangular plastic --- same size as You could well be right. It's been such a long time... > the console were used to select which disk drive got IPL'ed. Never saw an > 1130 with more than one disk drive, so always made sure all the switches > were off or the bootstrap would hang. There was a little rack about the same size as a 2-drawer filing cabinet which contained IIRC 2 add-on cartridge drives. These drives had, by modern standards, a capacity of around half a floppy. The cartridges were circular, 17" in diameter, and about an inch thick. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: abaum@pa.dec.com (Allen J. Baum) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 09:46:32 -0800 Organization: Compaq Computer Lines: 6 Message-ID: References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> <753sds$6j0@top.mitre.org> <3676848b$2$wg$mr2ice@news.epix.net> NNTP-Posting-Host: althea.pa.dec.com X-Newsreader: MT-NewsWatcher 2.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.wli.net!su-news-hub1.bbnplanet.com!su-news-feed2.bbnplanet.com!news.gtei.net!news1.digital.com!pa.dec.com!src.dec.com!abaum The HP9100 used a ROM that was a PC board. It had 14 layers, and used the presence or absence of traces on inner layers to change the capacitance, which could be sensed. This was actually cloned by the Russians, who peeled back the board layer by layer & copied it. ###### Message-ID: <3676A394.5FEF7346@compuserve.com> Date: Tue, 15 Dec 1998 09:59:48 -0800 From: Dale DePriest X-Mailer: Mozilla 4.5 [en] (WinNT; I) X-Accept-Language: en MIME-Version: 1.0 Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <3676103A.2B68@sunsouthwest.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Newsgroups: alt.folklore.computers Lines: 15 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.idt.net!WCG!arl-news-svc-3.compuserve.com!news-master.compuserve.com!nntp-ntawwabp.compuserve.com bill_h wrote: > Early ROM? > > Plug Boards. Full of either actual jumper wire or just > pins that made connections in a matrix. > > And, you could swap them in and out quickly. However reprogramming them took a bit of time. The eniac was programmed this way. Actually plug boards preceded, not only transistors, but computers. The tabulating equipment was programmed this way. Dale DePriest ###### Message-ID: <36767A5F.415B@gazonk.del> Date: Tue, 15 Dec 1998 10:03:59 -0500 From: "Foobar T. Clown" Organization: Blurp X-Mailer: Mozilla 3.01 (WinNT; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <3674B7AB.3543FBDC@tnglwood.demon.co.uk> <3675A44B.D0694F63@no.spam> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 198.133.110.49 X-Trace: 15 Dec 1998 10:05:32 -0500, 198.133.110.49 Lines: 9 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!ix.netcom.com!zeus.nomos.com!198.133.110.49 The Bakers wrote: > > Never saw an 1130 with more than one disk drive I used an 1130 in the late 70's that had an after-market drive in addition to the internal one. The third party drive had twenty platters on a single spindle, and from a program's point of view, it appeared to be twenty separate 1Mbyte drives. I don't remember the details of how a program would specify which logical drive it wanted to use. ###### From: jtNOSPAM@epix.net (Julian Thomas) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 10:37:40 -0500 Organization: epix Internet Services Lines: 24 Message-ID: <367682a7$1$wg$mr2ice@news.epix.net> References: <750ude$8va@netaxs.com> <36752844$1$wg$mr2ice@news.epix.net> <753s5c$6e9@top.mitre.org> NNTP-Posting-Host: itha-125ppp35.epix.net X-Newsreader: MR/2 Internet Cruiser Edition for OS/2 v1.51 b51 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-xfer.newsread.com!netaxs.com!newsread.com!news-xfer.epix.net!news1.epix.net!epix-news In <753s5c$6e9@top.mitre.org>, on 12/14/98 at 08:24 PM, jcmorris@mwunix.mitre.org (Joe Morris) said: >Did the /50 IPL code actually place the CCW in memory at absolute zero? >I know that the /65 did not, but the /40 did -- and there was an FE tip >that explained how to rewind a tape from the console: you set up a ROS >stop at such-and-such an address, set the IPL address switches to the >address of the tape drive, pressed IPL, and when the ROS stop hung the >microcode you then toggled the appropriate CCW opcode (0x07) into the >byte at absolute zero and allowed the IPL microcode to continue. Yes, I personally laid down that first CCW in location 0. Don't remember ever using the tape rewinding trick on the 50, but it could well have worked with the right ROS address. -- Julian Thomas: jt at epix dot net http://home.epix.net/~jt Boardmember of POSSI.org - Phoenix OS/2 Society, Inc http://www.possi.org In the beautiful Finger Lakes Wine Country of New York State! -- -- Air conditioned environment - Do not open Windows. ###### From: jtNOSPAM@epix.net (Julian Thomas) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 10:40:24 -0500 Organization: epix Internet Services Lines: 37 Message-ID: <3676848b$2$wg$mr2ice@news.epix.net> References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> <753sds$6j0@top.mitre.org> NNTP-Posting-Host: itha-125ppp35.epix.net X-Newsreader: MR/2 Internet Cruiser Edition for OS/2 v1.51 b51 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news-xfer.newsread.com!netaxs.com!newsread.com!news-xfer.epix.net!news1.epix.net!epix-news In <753sds$6j0@top.mitre.org>, on 12/14/98 at 08:28 PM, jcmorris@mwunix.mitre.org (Joe Morris) said: >Um...CROS was used in the /50; the /40 used TROS. The 30 used CCROS (Card Capacitor ROS); the 40 TROS as Joe said. The 50 used CROS - which wasn't easily modified - the process for making the bit planes was fairly complex, and took several days on a special rush. For a minor patch, there was a method that involved "painting" bits on an intermediate glass master and shortcircuiting part of the process. I had one of the 50 bit planes in my "museum" at IBM until sometime earlier in the '90s. 176 words - each word 120 bits (I think - somewhere between 100 and 120). It used a balanced capacitor readout, which meant that a 0 looked conceptually like this: | O| | | |O | and a 1 was: |O | | | | O| These are actually gold traces on one word line running up the bit plane; the sense lines run across. -- Julian Thomas: jt at epix dot net http://home.epix.net/~jt Boardmember of POSSI.org - Phoenix OS/2 Society, Inc http://www.possi.org In the beautiful Finger Lakes Wine Country of New York State! -- -- 99 percent of lawyers give the rest a bad name. ###### From: jtNOSPAM@epix.net (Julian Thomas) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 10:50:44 -0500 Organization: epix Internet Services Lines: 63 Message-ID: <3676889d$3$wg$mr2ice@news.epix.net> References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> <753fmd$g13@freenet-news.carleton.ca> <3675AA23.F56BDEA6@no.spam> NNTP-Posting-Host: itha-125ppp35.epix.net X-Newsreader: MR/2 Internet Cruiser Edition for OS/2 v1.51 b51 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed-east.supernews.com!supernews.com!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!news-xfer.epix.net!news1.epix.net!epix-news Joe said most of it quite well. Here's a little more input: In <3675AA23.F56BDEA6@no.spam>, on 12/15/98 at 12:15 AM, The Bakers said: >Did anyone ever actually microcode an APL processor and implement it ? >Were the microcode changes just to add some machine instructions which >were especially well suited for APL operation ? One machine was built with some assists - this was for a timesharing company that offered APL. One of the interesting features of this was a set of floating point DECIMAL operations. (This was NOT user microcode!). >How much capacity was in the microcode store ? Could you add entirely >new instructions, or were you limited to modifying existing ones ? If >you added a new instruction, did you have to give up one of the original >ones to make room ? The mod 50 had 2816 words. There weren't a lot of spares, especially after the infamous "floating point fix". The most common way of adding new instructions was to use the diagnose instruction with an address pointing to the starting ROS address of the microcode doing the work; saved all sorts of hassles about op code assignments (altho there were still a goodly number of unused op codes in 360 days; it wasn't until 370 and later that they went to the B2xx 2 byte op codes for new features). >Was the microcode "hand assembled" (for lack of a better term; you know >what I mean) or was there software which let you microcode in a symbolic >language ? There was a "Control Automation System" that took as input the machine specification (definition of the microorders, fields, etc.) and (via keypunch transcription) a somewhat graphical form of the microprogram - quite similar in concept to the format for logic desigh that was input into the [logic] Design Automation system. It also included a simulator (the rules for simulation were part of the machine description) - the mod 50 and 60 (with the hard to change CROS) used this much more extensively than the 30 and 40 (with the ability to punch up microcode changes on the spot). I wrote pieces of this also, including the programs that took the ROS contents and turned them into instructions to the machinery that actually personalized the bit planes. A bit of kludgery there! >I remember someone saying that IBM supplied the microcode listings for >the original machine. Did they encourage user microcode ? Tolerate it ? >Give you a hard time if you wrote your own ? Were there IBM manuals >describing how to microcode the processor ? I would imagine that IBM >made you put the original microcode back before the serviceman would work >on the machine, no ? Like Joe said. Nobody outside IBM EVER did microcode for the 50. -- Julian Thomas: jt at epix dot net http://home.epix.net/~jt Boardmember of POSSI.org - Phoenix OS/2 Society, Inc http://www.possi.org In the beautiful Finger Lakes Wine Country of New York State! -- -- 42.7 percent of all statistics are made up on the spot. ###### From: spalding@iol.ie (Nick Spalding) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 12:02:08 GMT Organization: http://extra.newsguy.com Lines: 10 Message-ID: <36854e16.501822001@news.iol.ie> References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> <753fmd$g13@freenet-news.carleton.ca> Reply-To: spalding@iol.ie NNTP-Posting-Host: p-107.newsdawg.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!ix.netcom.com!pln-e!spln!extra.newsguy.com!newsp.newsguy.com!news3 Heinz W. Wiggeshoff wrote: > Which made the -40 an ideal machine to try an APL processor. > (And I'll bet that keypunching errors weren't cheap!) The -40 was the one with the transformer ROS for which you had to punch each hole individually with a hand punch. The -30 used the mylar cards. -- Nick Spalding ###### From: jcmorris@mwunix.mitre.org (Joe Morris) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 15 Dec 1998 14:02:26 GMT Organization: The MITRE Corporation Lines: 40 Message-ID: <755q5i$sqk@top.mitre.org> References: <3673B896.F2C58E2B@vnet.ibm.com> <367478A1.B588E1A2@acm.org> <367578e3.225001737@news.mcs.net> NNTP-Posting-Host: mwunix.mitre.org Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!204.238.120.130!news-feeds.jump.net!world!blanket.mitre.org!news.mitre.org!mwunix!jcmorris arargh@arargh.com (Arargh!) writes: >On Sun, 13 Dec 1998 20:32:01 -0600, "Joel C. Ewing" >wrote: >>On the 360/370/390 architecture level there is no read-only storage >>involved. The architecture definition provides for an IPL (Initial >>Program Load) sequence in which a standard I/O channel read from an >>operator specified external device is initiated to read data into a >>fixed location in memory followed a startup of a central processor on >>the resulting data. The resulting bootstrap process eventually results >>in the loading and initialization of the entire operating system. >On the few S/370's that I saw, they IPML'd from an 8" floppy. That's IMPL, not IPML. Initial Micro Program Load. I've long suspected that someone in IBM's upper management levels decreed that any new product had to use whatever components the company hadn't been able to sell in the past; examples include: - the 8" floppy drives which appeared not only in the CPUs but also in the 3880 disk controllers and the 327x display controllers - the idiotic MCST (Magnetic Card Selectric Typewriter) card readers that were used in the 6670 printers - the FBA hard disks ("Fixed Block Architecture") that became the supervisor storage for the internal control OS within the 308x systems. (FBA disks used fixed record sizes, effectively the same concept as is used in PC disks today.) IIRC the disk used in the 3081 was the Piccolo drive. >>Historical source: Microprogramming: Principles and Practices, Samir S. >>Husson, Prentice Hall 1970. Does anyone know where Samir is these days? I've not heard of him for many years. Joe Morris ###### From: jcmorris@mwunix.mitre.org (Joe Morris) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 15 Dec 1998 14:31:41 GMT Organization: The MITRE Corporation Lines: 79 Message-ID: <755rsd$5l@top.mitre.org> References: <750ude$8va@netaxs.com> <753cjf$nog$1@flood.weeg.uiowa.edu> <753fmd$g13@freenet-news.carleton.ca> <3675AA23.F56BDEA6@no.spam> NNTP-Posting-Host: mwunix.mitre.org Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!18.24.4.11!newsswitch.lcs.mit.edu!news-f.std.com.MISMATCH!world!blanket.mitre.org!news.mitre.org!mwunix!jcmorris The Bakers writes: >"Heinz W. Wiggeshoff" wrote: >> Douglas W. Jones,201H MLH,3193350740,3193382879 (jones@cs.uiowa.edu) writes: >> [snip] >> Which made the -40 an ideal machine to try an APL processor. >> (And I'll bet that keypunching errors weren't cheap!) >How much capacity was in the microcode store ? Could you add entirely new >instructions, or were you limited to modifying existing ones ? If you >added a new instruction, did you have to give up one of the original ones >to make room ? I don't recall the size of the address space in the /40 ROS. There was probably some slack space in any configuration since the buckets that held the ROS tapes had a fixed size, but for a given set of options the CPU was built with only enough buckets to hold the required number of tapes. Adding floating point arithmetic, as an example (and IIRC) required the installation of additional buckets. >Was the microcode "hand assembled" (for lack of a better term; you know what I >mean) or was there software which let you microcode in a symbolic language ? >I remember someone saying that IBM supplied the microcode listings for the >original machine. True; back in Ye Olde Days (and through the days of the S/370 line) one of the standard items in the CE cart was a binder with the complete microcode assembly listings. The later listings were clearly created by a symbolic assembler (with some intriguing and otherwise undocumented features) but ROS listings for the older systems such as the /40 were what appeared to be roughly formatted binary dumps, along with narratives describing the control flow. They were quite useful in figuring out some hacks; for example, I using them I built a program that was able to read the bit switches on the operator's console for several different S/360 models. (The program was distributed through one of the SHARE projects, with an hilarious mistake that attributed it to "Joe Morris, of the University of Tennessee in Lexington, Kentucky".) > Did they encourage user microcode ? Gawd, NO. While there were certainly places that played games with the microcode, maintenance of either the hardware or software would be a nightmare if the CEs or PSRs could not be certain that the microcode logic was reliable (or at least a known quantity). I recall one IBM sales pitch in the late 1970s which included a vugraph attacking the idea of user microcode: it had a drawing of a cartoon revolver (labeled "User Microcode") with the barrel mounted backwards, aimed at whoever was holding the grip in the usual manner. This was a direct slap at some of the clone manufacturers who explicitly provided for user-written microcode -- and if you looked carefully at the gun you could find the legend "Made in Japan," a reference to the source of several of the clone systems. (Hitachi? I can't recall which vendor had the most blatant advertising of its user-written microcode capability.) As a hacker (in the original sense of the word!) I liked the idea of being to write my own microcode, but as the sysprog in a production mainframe computer center it scared me because of its potential for all sorts of mischief. > Tolerate it ? Give you >a hard time if you wrote your own ? See above. > Were there IBM manuals describing how to >microcode the processor ? There weren't any SRLs on the subject that I know of, but the standard set of CE documentation was quite complete in its description of both the microengine and the logic of the microcode -- at least until New World hit IBM in 1969. Joe Morris ###### From: jones@cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 15 Dec 1998 16:47:19 GMT Organization: The University of Iowa Lines: 34 Message-ID: <7563qn$rqq$1@flood.weeg.uiowa.edu> References: <3676889d$3$wg$mr2ice@news.epix.net> NNTP-Posting-Host: pyrite.cs.uiowa.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newspeer.monmouth.com!feeder1.wwnet.net!streamer1.cleveland.iagnet.net!NewsNG.Chicago.Qual.Net!news.uiowa.edu!not-for-mail >>Did anyone ever actually microcode an APL processor and implement it ? >>Were the microcode changes just to add some machine instructions which >>were especially well suited for APL operation ? Outside IBM, there were, of course, a number of other machines with microcode that users could twiddle. The HP2100MX (Hewlett Packard) was built in the mid 1970's, and was user microprogrammable. The default instruction set was, of course, the HP2100 instruction set, upward compatable from the HP2115 mini designed in the mid to late 1960's. I just recently sent my copy of the microcode for this machine, along with a battered CPU and other documentation, to a collector in St. Louis who is trying to restore a machine. The Burroughs B1700 and B1800 (Burroughs) were also user microprogrammable; on these machines, it was expected that compiler writers would design a target instruction set for each programming language and write microcode to interpret that instruction set. The programmer's manual set for this machine included documentation of the microarchitecture, and the available support tools included a microassembler. We had these machines at Iowa when I arrived on the faculty back in 1980. And, of course, machines like the VAX 11/780 and my IBM RT PC load microcode as part of their bootstrap sequence, but no users of these machines are expected to fiddle with the microcode. All of the above machines use modern ROM and RAM technology to hold microcode, so they're a bit off topic for this thread. Doug Jones jones@cs.uiowa.edu ###### From: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 15 Dec 1998 17:20:25 GMT Organization: The National Capital FreeNet Lines: 24 Message-ID: <7565op$sum@freenet-news.carleton.ca> References: <3676889d$3$wg$mr2ice@news.epix.net> <7563qn$rqq$1@flood.weeg.uiowa.edu> Reply-To: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) NNTP-Posting-Host: freenet6.carleton.ca X-Given-Sender: ab528@freenet6.carleton.ca (Heinz W. Wiggeshoff) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newshub.northeast.verio.net!newsserver.jvnc.net!xcski.com!freenet-news.carleton.ca!FreeNet.Carleton.CA!ab528 Douglas W. Jones,201H MLH,3193350740,3193382879 (jones@cs.uiowa.edu) writes: >>>Did anyone ever actually microcode an APL processor and implement it ? >>>Were the microcode changes just to add some machine instructions which >>>were especially well suited for APL operation ? > > Outside IBM, there were, of course, a number of other machines with > microcode that users could twiddle. I posted the comment about APL on a model 40 thinking that I would find the CACM (?) item next day. What I have dredged up so far is A Microprogrammed APL Implementation Rodnay Zaks Sybex 2020 Milvia Berkeley, Ca. 94704 ISBN 89588-005-9 c/r 1978 (purchased from the long gone Heathkit store in Ottawa for $27.20) In the preface, he writes "The program can be transcoded into any 'standard' instruction-set such as an 8080, 6800, or other, with reasonable ease." No, I never did build that 6809 APL machine. B-( ###### From: jsavard@tenMAPSONeerf.edmonton.ab.ca (John Savard) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 18:33:06 GMT Organization: Videotron Communications Ltd. Lines: 17 Message-ID: <3676ab04.5538369@news.prosurfr.com> References: <3673B896.F2C58E2B@vnet.ibm.com> <367587b9.20293582@news.prosurfr.com> <3675B940.C470D74F@jkmicro.com> NNTP-Posting-Host: c9169-002.prosurfr.com X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.direct.ca!news.insinc.net!stimpy.cal.sfl.net!news.videotron.ab.ca!not-for-mail Jim Stewart wrote, in part: >Didn't the early ESS central office switches have a really perverse >memory composed of aluminium plates with ferrite chips embedded in >them? I seem to recall that the plates could be written in an external >device then installed in the switch where they were read-only. I remember that what they did was have two computers connected to the same memory, and the one that ran the program to handle phone calls couldn't write on the memory. This meant that the system was immune to any kind of a bug that might lead to memory being overwritten, and this was done so that the ESS would have very high reliability. John Savard http://www.freenet.edmonton.ab.ca/~jsavard/index.html ###### From: "T. Cook" Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 15 Dec 1998 21:17:05 -0800 Organization: Splat Kings Lines: 20 Message-ID: <36774251.10C802C2@nettaxi.com> References: <3673B896.F2C58E2B@vnet.ibm.com> Reply-To: kukuman@N0SPaM.nettaxi.com NNTP-Posting-Host: 198.68.31.232 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.5 [en] (Win98; I) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.he.net!news.or.nw.verio.net!198.68.31.232!tsn-ppp194.rattles.com Andrei Heilper wrote: > > I remember that in the old days, in order to start a minicomputer you > had to set some switches on the panel. This sequence of bits > represented a very short loader (by the way, these were tricky > self-modifying programs, and it seems quite interesting to have a > collection of them). My question is if it was another solution for > larger computers? > > Andrei Heilper Vacuum tubes, I guess... -- Kukuman http://splatkings.cjb.net/ -------------------------- "And the mame rathe outgrabe." - Jabberwocky ###### From: mwilson@interlog.com (Mel Wilson) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 16 Dec 1998 11:00:26 -0500 Organization: Interlog Internet Services -Voice (416) 975-2655 -Data 515-1414 Lines: 19 Message-ID: References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> NNTP-Posting-Host: shell1.interlog.com NNTP-Posting-Time: 16 Dec 1998 16:00:28 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.berkeley.edu!peerfeed.ncal.verio.net!newsfeed.interlog.com!news.interlog.com!not-for-mail In article <3675635e.4102963@news.vip.net>, genew@vip.net (Gene Wirchenko) wrote: >mwilson@interlog.com (Mel Wilson) wrote: >[snip] >> From hanging around with the Field Engineers, I seem to remember >>hearing that some of the H-200s used core rope for some internal logic, >>as well. (?) > What is "core rope", please? I think that was a Read-Only-Memory made by weaving drive wires, one per word, through a series of ferrite rings, one for each bit in a word. Strobing a drive wire would induce a 1 or 0 in the sense coil on a ring depending on whether the drive wire went through or outside the ring. I picked up the "core rope" term from another post. I begin to recall the FEs calling it a "snake". Regards. Mel. ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> X-Disclaimer: Everything I write is false. Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 16 Dec 1998 12:33:35 -0800 Message-ID: Lines: 27 X-Newsreader: Gnus v5.5/Emacs 20.2 NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 16 Dec 1998 12:28:06 -0800, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!news-peer.gip.net!news.gsl.net!gip.net!enews.sgi.com!news.sgi.com!news.spies.com!ruckus.brouhaha.com Gene asked: > What is "core rope", please? Mel Wilson replied: > I think that was a Read-Only-Memory made by weaving drive wires, one > per word, through a series of ferrite rings, one for each bit in a word. > Strobing a drive wire would induce a 1 or 0 in the sense coil on a ring > depending on whether the drive wire went through or outside the ring. > > I picked up the "core rope" term from another post. I begin to > recall the FEs calling it a "snake". There is apparently no definitive name for this. In addition to "core rope memory", it was also known as "wire braid memory" and (by IBM) TROS (Transformer Read-Only Store). If memory serves, the PDP-9 used 64 words (by what width, I don't know) of core rope memory for its control store. The HP 9100 desktop calculator uses a capactive printed-circuit board memory for its 512x64 microcode store, but it also has some core rope memory for its 64x29 control logic (nanocode). And of course, the Apollo Guidance Computer used core rope memory. Anyone have a spare AGC for sale? :-) ###### From: jtNOSPAM@epix.net (Julian Thomas) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Wed, 16 Dec 1998 18:53:58 -0500 Organization: epix Internet Services Lines: 20 Message-ID: <3678484b$1$wg$mr2ice@news.epix.net> References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> NNTP-Posting-Host: itha-125ppp27.epix.net X-Newsreader: MR/2 Internet Cruiser Edition for OS/2 v1.51 b51 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news-xfer.newsread.com!netaxs.com!newsread.com!news-xfer.epix.net!news1.epix.net!epix-news In , on 12/16/98 at 12:33 PM, Eric Smith said: >> I picked up the "core rope" term from another post. I begin to >> recall the FEs calling it a "snake". >There is apparently no definitive name for this. In addition to "core >rope memory", it was also known as "wire braid memory" and (by IBM) TROS >(Transformer Read-Only Store). "Snake" was the term used at Honeywell when the device was first used. -- Julian Thomas: jt at epix dot net http://home.epix.net/~jt Boardmember of POSSI.org - Phoenix OS/2 Society, Inc http://www.possi.org In the beautiful Finger Lakes Wine Country of New York State! -- -- Walk through doors, don't crawl through Windows. ###### From: jcmorris@mwunix.mitre.org (Joe Morris) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 16 Dec 1998 22:37:27 GMT Organization: The MITRE Corporation Lines: 58 Message-ID: <759cn7$akl@top.mitre.org> References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> NNTP-Posting-Host: mwunix.mitre.org Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet!rill.news.pipex.net!pipex!ams.news.uu.net!uunet!in2.uu.net!world!blanket.mitre.org!news.mitre.org!mwunix!jcmorris Eric Smith writes: >There is apparently no definitive name for this. In addition to >"core rope memory", it was also known as "wire braid memory" and >(by IBM) TROS (Transformer Read-Only Store). TROS used the concept, but the implementation consisted of Mylar strips with square holes (through which the transformer cores passed), and with printed-circuit traces that ran around the card with branches on both sides of each hole. There were two traces on each strip, corresponding to two ROS words. Let's see what ASCII art can do to show the strip in the vicinity of a pair of transformer holes: ---------+-------X-X-----+----------- | | | ---- | | |**| | | ---- | +---------------+ +-------X-X-----+ | ---- | | |**| | | ---- | | | --------+---------------+----------- (Repeat this about 70 times and bring the two ends of each word line out to a connector if you want to actually build a single TROS tape.) This shows the two holes in the strip; the two prongs of a single U-shaped transformer core would go through them, meeting end pieces in the buckets where the sense wire would be located. The two word leads are manufactured so that each one splits and goes both inside the transformer core (where they are close to each other) and outside as well. The strip is coded by punching holes in either the inside or outside trace at *each* core location; here I've shown the punches with the character "X" (every strip was punched twice at each hole to ensure a true break.) As shown, the upper word line would have a binary 1 at this point (the outer trace is broken, so when the word line is pulsed current will flow through the transformer core and cause a pulse in the attached sense wire); in the lower word line it's the inner trace that's broken so when it is pulsed nothing happens on the sense wire and a binary 0 is recognized. So...while other implementations of this mechanism may have used flexible "core ropes" to control which bits were ones and zeros, the IBM TROS design used the strips which were *far* easier to work with. Do you *really* want to have to figure out which wire of a core rope is misthreaded? Or to have to install a complex update to the design if it requires that you rethread a zillion wires? (I hope that I don't have to explain the last comment, especially to readers here who ever had to work on wire-wrapped backplanes in the days before planar boards replaced all but a small amount of real wire connections.) Joe Morris ###### From: spalding@iol.ie (Nick Spalding) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Thu, 17 Dec 1998 11:40:37 GMT Organization: Ireland On-Line Lines: 11 Message-ID: <367cecae.53454484@news.iol.ie> References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> <759cn7$akl@top.mitre.org> Reply-To: spalding@iol.ie NNTP-Posting-Host: dialup-0102.dublin.iol.ie Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!news-lond.gip.net!news.gsl.net!gip.net!iol!not-for-mail Joe Morris wrote: > (I hope that I don't have to explain the last comment, especially to > readers here who ever had to work on wire-wrapped backplanes in the > days before planar boards replaced all but a small amount of real wire > connections.) Why was the wire you had to remove for some damned EC always right at the bottom of the mat? -- Nick Spalding ###### From: "Donald Tees" Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Thu, 17 Dec 1998 13:54:23 -0500 Organization: IGS - Information Gateway Services Lines: 16 Message-ID: <75bjd8$q5s$1@news.igs.net> References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> <759cn7$akl@top.mitre.org> <367cecae.53454484@news.iol.ie> <75b92o$1l9@top.mitre.org> NNTP-Posting-Host: ttye10.kw.igs.net X-Trace: news.igs.net 913920232 26812 206.248.37.144 (17 Dec 1998 18:43:52 GMT) X-Complaints-To: abuse@igs.net NNTP-Posting-Date: 17 Dec 1998 18:43:52 GMT X-Newsreader: Microsoft Outlook Express 4.72.2106.4 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!130.185.14.36!torn!nntp.igs.net!news.igs.net!not-for-mail Joe Morris wrote in message <75b92o$1l9@top.mitre.org>... >What you apparently don't understand is that this was *intentional*. The >design team that chose which wire needed to be replaced was actually >under contract to a psychological analysis firm; the intent was to >determine just how much invective could be used by the CE before the >customer's management filed a complaint. Gee, I thought it was just Murphy's law #762451. I learn something new every day. ;<) ###### From: jcmorris@mwunix.mitre.org (Joe Morris) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 17 Dec 1998 15:47:36 GMT Organization: The MITRE Corporation Lines: 21 Message-ID: <75b92o$1l9@top.mitre.org> References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> <759cn7$akl@top.mitre.org> <367cecae.53454484@news.iol.ie> NNTP-Posting-Host: mwunix.mitre.org Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!18.24.4.11!newsswitch.lcs.mit.edu!news-f.std.com.MISMATCH!world!blanket.mitre.org!news.mitre.org!mwunix!jcmorris [Mailed and posted] spalding@iol.ie (Nick Spalding) writes: >Joe Morris wrote: >> (I hope that I don't have to explain the last comment, especially to >> readers here who ever had to work on wire-wrapped backplanes in the >> days before planar boards replaced all but a small amount of real wire >> connections.) >Why was the wire you had to remove for some damned EC always right at >the bottom of the mat? What you apparently don't understand is that this was *intentional*. The design team that chose which wire needed to be replaced was actually under contract to a psychological analysis firm; the intent was to determine just how much invective could be used by the CE before the customer's management filed a complaint. Joe Morris ###### From: jtkare@ibm.net (Jordin Kare) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Thu, 17 Dec 1998 20:01:52 -0800 Organization: Sirius Connections Lines: 35 Message-ID: References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> NNTP-Posting-Host: ppp-asok01--035.sirius.net Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!news2.best.com!news1.best.com!newshub.sirius.com!newsfiler.sirius.com!ppp-asok01--035.sirius.net!user In article , Eric Smith wrote: > > What is "core rope", please? > > I think that was a Read-Only-Memory made by weaving drive wires, one > > per word, through a series of ferrite rings, ... > > I picked up the "core rope" term from another post. I begin to > > recall the FEs calling it a "snake". > > If memory serves, the PDP-9 used 64 words (by what width, I don't know) of > core rope memory for its control store. Core rope memory turned up in several other places in DEC hardware as well. The one I remember was the large graphics CRT display used with the PDP-7 (something like a 14" diameter circular CRT, displaying vector graphics) which had core rope ROM defining the character shapes for the builtin character generator. Also, I still have a keyboard from an airline reservation terminal or similar, probably dating from the early '70s, that used a crude form of "core rope" as the keyboard encoder. Seven individual toroids were threaded with jumpers to define a seven-bit output when a key was pressed, although the keyboard also used a matrixing scheme (as I recall) that allowed keys with similar codes to share some of the jumpers, reducing the number of jumpers through each toroid to a manageable number. The rest of the circuitry was TTL gates/flipflops. The output was ASCII, with a number of strange function codes. I actually traced the entire thing at one point in order to modify the ROM so that some of the "function" keys would generate more useful codes like escape. That keyboard also used very fancy key switches: reed switches actuated by magnets attached to the keys. The whole thing was about 2 inches deep from keytop to circuit board... Jordin (Write Only Memory) Kare ###### From: bbreynolds@aol.com (BBReynolds) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Lines: 25 NNTP-Posting-Host: ladder01.news.aol.com X-Admin: news@aol.com Date: 18 Dec 1998 02:02:10 GMT References: <753sds$6j0@top.mitre.org> Organization: AOL http://www.aol.com X-Newsreader: AOL Offline Reader Message-ID: <19981217210210.12961.00001523@ngol04.aol.com> Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!portc02.blue.aol.com!audrey01.news.aol.com!not-for-mail In article <753sds$6j0@top.mitre.org>, jcmorris@mwunix.mitre.org (Joe Morris) writes: >TROS was also used in the 2314 disk controller. > > And, I believe, in the 2841 controller for the 2311 drives. Had a 2841 attached to a channel on a 1800. Full access failure: the 1800 (GSD) FE's came in and everything pointed to the 2841, except the 2841 reported no errors from its diagnostics. The 2841 (DPD) FE's forced the GSD FE's to move every card from one 1800 channel to the other (system also had a 2848/2260 combo); still no access. DPD FE finally showed up, reran all the 2841 diagnostics, and determined that one bit was not being read when accessed (he had to put a scope on each single bit of the array until he found the faulty one); he physically patched that bit, and the 2841 started running immediately. -- Bruce B. Reynolds, Systems Consultant: Founder of Trailing Edge Technologies, Glenside PA and Niles IL: Sweeping Up Behind Data Processing Dinosaurs ###### From: jhall@hanover.corp.adobe.com (Jerry Hall) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 18 Dec 1998 22:03:22 -0800 Organization: Adobe Systems, Incorporated Lines: 51 Message-ID: <75ffja$r4h@hanover.corp.adobe.com> References: <3676889d$3$wg$mr2ice@news.epix.net> <7563qn$rqq$1@flood.weeg.uiowa.edu> NNTP-Posting-Host: hanover.corp.adobe.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newspeer.monmouth.com!newsxfer.itd.umich.edu!chronicle.adobe.com!enquirer.corp.adobe.com!not-for-mail In article <7563qn$rqq$1@flood.weeg.uiowa.edu>, Douglas W. Jones,201H MLH,3193350740,3193382879 wrote: >>>Did anyone ever actually microcode an APL processor and implement it ? >>>Were the microcode changes just to add some machine instructions which >>>were especially well suited for APL operation ? > >Outside IBM, there were, of course, a number of other machines with >microcode that users could twiddle. > >The HP2100MX (Hewlett Packard) was built in the mid 1970's, and was >user microprogrammable. The default instruction set was, of course, >the HP2100 instruction set, upward compatable from the HP2115 mini >designed in the mid to late 1960's. I just recently sent my copy of >the microcode for this machine, along with a battered CPU and other >documentation, to a collector in St. Louis who is trying to restore >a machine. > >The Burroughs B1700 and B1800 (Burroughs) were also user >microprogrammable; on these machines, it was expected that compiler >writers would design a target instruction set for each programming >language and write microcode to interpret that instruction set. The >programmer's manual set for this machine included documentation of >the microarchitecture, and the available support tools included a >microassembler. We had these machines at Iowa when I arrived on the >faculty back in 1980. Another minicomputer that allowed user microcode was the Data General Eclipse. I implemented a Pascal system at the University of Lancaster in about 1979 where the P-machine was almost completely implemented in about 600 57 bit words user microcode on an Eclipse S/130, IIRC. The only part that I deferred to the standard instruction set was floating point handling which seemed too daunting to me at the time. Programs complied by my system would include the microcoded P-code interpreter and code to automatically load it when executed. I performed timing tests on a variety of programs and compared them to the same programs written in Algol-W and Fortran. I seem to recall being about 3-4 faster than Fortran on the computationally intensive programs and may have had the fastest HHL running on that machine at the time. I always got the feeling that there were very few sites running with the Writable Control Store (WCS) option on the Eclipse but there at least one other site in the UK at that time: International Computers Limited (ICL)! They took a copy of my code and were using it to develop Pascal programs on the Eclipse for computer design or some such. This introduction lead to work on several microcoded bit-slice graphics controllers in the 80s when that kind of thing still made economic sense. ###### From: phil@panix.com (Phil Gustafson) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: 20 Dec 1998 17:29:18 -0500 Organization: Once-famed Parquet Floor Lines: 11 Message-ID: <75jtnu$p2b@panix2.panix.com> References: <3673B896.F2C58E2B@vnet.ibm.com> <367485B4.367384D3@no.spam> <3675635e.4102963@news.vip.net> NNTP-Posting-Host: panix2.nfs100.access.net X-Newsposter: trn 4.0-test55 (26 Feb 97) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!howland.erols.net!panix!news.panix.com!not-for-mail When I was a kid (50's) I had a few electrical quiz games. The "game board" was a sheet of aluminum foil glued to a piece of cardboard. The foil was cut up by a series of wiggly lines. The quizzes were printed on thick paper overlays with holes at the answer points. If you touched a wired stylus to a bit of foil at one of the right spots, a lamp lit and a buzzer sounded. I'd call that a ROM. Phil ###### Message-ID: <367DD15A.314352D1@flash.net> From: Richard Lamb Reply-To: lamb01@flash.net Organization: Earth X-Sender: "Richard Lamb" (Unverified) X-Mailer: Mozilla 4.04 [en]C-FLASHNET (Win95; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <750ude$8va@netaxs.com> <3674100e.220383965@news.monmouth.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 41 Date: Mon, 21 Dec 1998 04:33:39 GMT NNTP-Posting-Host: 209.30.91.147 X-Complaints-To: abuse@flash.net X-Trace: news.flash.net 914214819 209.30.91.147 (Sun, 20 Dec 1998 22:33:39 CDT) NNTP-Posting-Date: Sun, 20 Dec 1998 22:33:39 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!nntp.flash.net!news.flash.net!not-for-mail I tought on flight simulators for the Navy until 1979. The A4 platform used a pair of Sigma V's which were loaded drom paper tape. THere uas a "ROM" loaded made from a diode array that would load the tape into memory. In 1979, we got a 5" floppy for it! Wooooooooooo... The T2 (Buckeye) platform was run by PDP-11/34's. They had 5 MB hard drives. Really up town... Richard Wilbur Streett wrote: > > I'm amazed that people don't know this one.. > > The first time that I got to see a computer, in 1967, my father showed me > the different strips of paper tape that they had setup to initialize the > computer with different hardware configurations. They would load the paper > tape into the system default reader and when they powered on the system it > would load and execute whatever was on the paper tape by default. The > paper tape would end up on the floor, and they would pick it up and tape it > to the wall for the next time that they would startup the system. > Eventually they figured out the dirt on the floor was messing up the tape > reader, and they figured out that loading the paper tape for every load > meant that they had to walk across the room and would take more time. They > then looped the paper tape by taping the beginning to the end, and the loop > sat on the tape reader since the tape had pretty much be replaced by the > magnetic tape and cards for I/O. They also figured out that reconfiguring > the system required quite a bit of manual intervention, so they would > configure the system with all the appropriate devices and leave them all > connected even if they were only doing a card sort. Eventually, the loop > of paper tape, which looked like a loop the the back of a pair of military > boots, was identified with the process of putting on the boots in the > morning, or "picking yourself up by your own bootstraps".. or more simply > "booting up" > > The Papertape was Read Only Memory or ROM, and we still "boot" computers up > when we start them. > > Wilbur ###### Message-ID: <367F1B4E.D4B6AB2E@jkmicro.com> From: Jim Stewart Organization: http://www.jkmicro.com X-Mailer: Mozilla 4.04 [en] (Win95; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> <752057$pnc$1@news.wlg.netlink.net.nz> <367D8761.34B06FD7@stoneweb.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 33 NNTP-Posting-Host: 168.150.253.139 X-Trace: news15.ispnews.com 914213140 168.150.253.139 (Sun, 20 Dec 1998 23:05:40 EDT) NNTP-Posting-Date: Sun, 20 Dec 1998 23:05:40 EDT Date: Mon, 21 Dec 1998 20:08:46 -0800 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!News.Amsterdam.UnisourceCS!news.freedom2surf.net!ayres.ftech.net!news.ftech.net!baron.netcom.net.uk!netcom.net.uk!news5.cableinet.net!diablo.theplanet.net!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!hub1.ispnews.com!news15.ispnews.com.POSTED!not-for-mail Carl R. Friend wrote: > Don Stokes wrote: > > > > Carl R. Friend wrote: > > > Core memory, itself, can be used as a ROM if you simply don't > > >write to it. Core is non-volatile so whatever you put into it stays > > >there while the power's off. > > > > This isn't actually a good idea, because to read from core memory > > you in fact have to erase it. A core read sequence goes: > > [edit] > > This is true, but in that case you hope the P/S capacitors have > enough punch to write that word back into the memory before they give > out (hence 2 levels of power-fail indication: e.g. ACLO/DCLO for a > pdp11). > > In all honesty, I really should have mentioned that core is a > destructive-read mechanism which requires that data be re-written > to a location following a read. 20/20 hindsight, yes, I know. Sorry. > Yours is still a valid point. The BART system in the Bay area had a system of animated signs at each station that were driven by PDP 8/e or 8/f computers. The signs themselves had no provisions for loading the program and they were updated or repaired by swapping the core memory units with ones that had been preprogrammed at the shop. Not much different from changing an eprom. ###### Date: 21 Dec 98 11:30:20 -0800 From: "Charlie Gibbs" Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> <752057$pnc$1@news.wlg.netlink.net.nz> <367D8761.34B06FD7@stoneweb.com> <367F1B4E.D4B6AB2E@jkmicro.com> Message-ID: <2222.659T1010T6904519@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 49 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 21 Dec 1998 16:46:22 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.123 In article <367F1B4E.D4B6AB2E@jkmicro.com> jstewart@jkmicro.com (Jim Stewart) writes: >Carl R. Friend wrote: > >> Don Stokes wrote: >> > >> > Carl R. Friend wrote: >> > > Core memory, itself, can be used as a ROM if you simply don't >> > >write to it. Core is non-volatile so whatever you put into it stays >> > >there while the power's off. >> > >> > This isn't actually a good idea, because to read from core memory >> > you in fact have to erase it. A core read sequence goes: >> > [edit] >> >> This is true, but in that case you hope the P/S capacitors have >> enough punch to write that word back into the memory before they give >> out (hence 2 levels of power-fail indication: e.g. ACLO/DCLO for a >> pdp11). >> >> In all honesty, I really should have mentioned that core is a >> destructive-read mechanism which requires that data be re-written >> to a location following a read. 20/20 hindsight, yes, I know. Sorry. > >Yours is still a valid point. The BART system in the Bay area had a >system of animated signs at each station that were driven by PDP 8/e >or 8/f computers. The signs themselves had no provisions for loading >the program and they were updated or repaired by swapping the core >memory units with ones that had been preprogrammed at the shop. Not >much different from changing an eprom. On the other hand, you could try another technology. The plated-wire memories used in the Univac 9300, 9400, and probably other systems of the time (late '60s and early '70s) had a nondestructive read. But speaking of reprogramming remote units, I think of a friend who built a number of IMSAIs that were used as communications front-end processors. The way he'd get new programs out into the field was to write them into a RAM-4A board (a 4K S-100 RAM board - hey, it took a boardful of 2102s to hold 4K), attach a 6-volt battery to the power pads provided on the board, then power down the machine and remove the board. As long as the battery remained connected, it would hold the program while he went to the remote machine, powered it down, and dropped the board in. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 22 Dec 1998 06:25:23 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 34 Message-ID: <367f1aae.40233279@news.vip.net> References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> <752057$pnc$1@news.wlg.netlink.net.nz> <367D8761.34B06FD7@stoneweb.com> <367F1B4E.D4B6AB2E@jkmicro.com> <2222.659T1010T6904519@sky.bus.com> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.13 X-Trace: 914307731 A01OARAUVD4D CCD1C usenet87.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!Supernews73!supernews.com!Supernews69!not-for-mail "Charlie Gibbs" wrote: [snipped previous] >On the other hand, you could try another technology. The plated-wire >memories used in the Univac 9300, 9400, and probably other systems of >the time (late '60s and early '70s) had a nondestructive read. > >But speaking of reprogramming remote units, I think of a friend who >built a number of IMSAIs that were used as communications front-end >processors. The way he'd get new programs out into the field was >to write them into a RAM-4A board (a 4K S-100 RAM board - hey, it >took a boardful of 2102s to hold 4K), attach a 6-volt battery to >the power pads provided on the board, then power down the machine >and remove the board. As long as the battery remained connected, >it would hold the program while he went to the remote machine, >powered it down, and dropped the board in. Nice! I have a few questions. 1) A SIX volt battery? Was it really 6V or was it several batteries combined? Either way, more details please. 2) Didn't the memory operate on 5V or a) Was the extra volt within tolerances? b) Did the memory card have a voltage regulator? Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> X-Newsreader: NN version 6.5.1 (NOV) From: kbn@shell1.iglou.com (Ken Nelson) NNTP-Posting-Host: 192.107.41.17 X-NNTP-Posting-Host: 192.107.41.17 Message-ID: <367f976d.0@news.iglou.com> Date: 22 Dec 1998 07:58:21 -0500 X-Trace: 22 Dec 1998 07:58:21 -0500, 192.107.41.17 Lines: 22 X-NNTP-Posting-Host: news-incoming.iglou.com Organization: IgLou Internet Services, Inc. Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!netnews.com!dca1-hub1.news.digex.net!dca1-feed3.news.digex.net!digex!news.iglou.com!shell1!kbn The RCA Spectra computers used a line of E cores, one for each bit. The E cores had wire running through them; a wire for each word. The wires would pass through either the first opening(zero) or the second(one). To read each word, a current was passed through the wire. There were sense wires wrapped around the center of the E core. A pulse was produced when the pulse was passed through the read word wire. The pulse was either positive or negitive depending upon whether the read wire went through the first opening or the second. The program was put in by placing the wires through the first or second opening and required no holding voltage. EEEEE EEEEE EEEEE * E **** E E E * E E EEEE * EEEE EEEE E ** E ***** E*************word wire, sense wires wrapped around center E E E EEEEE EEEEE EEEEE regards ken ###### From: william.hamblen@nashville.com (William Hamblen) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Date: Tue, 22 Dec 1998 06:31:22 -0600 Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 27 Message-ID: References: <3673B896.F2C58E2B@vnet.ibm.com> <3673CD55.2D800B7E@stoneweb.com> <752057$pnc$1@news.wlg.netlink.net.nz> <367D8761.34B06FD7@stoneweb.com> <367F1B4E.D4B6AB2E@jkmicro.com> <2222.659T1010T6904519@sky.bus.com> <367f1aae.40233279@news.vip.net> Reply-To: william.hamblen@nashville.com NNTP-Posting-Host: 207.65.139.169 X-Trace: 914407320 4TJCV727N8BA9CF41C usenet57.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.corridex.com!Supernews73!supernews.com!Supernews69!william.hamblen Gene Wirchenko (genew@vip.net) wrote: : Nice! I have a few questions. : 1) A SIX volt battery? Was it really 6V or was it several batteries : combined? Either way, more details please. : 2) Didn't the memory operate on 5V or a) Was the extra volt within : tolerances? b) Did the memory card have a voltage regulator? S100 machines supplied unregulated 8 V to the circuit boards and each board had, usually, 7805 voltage regulators to supply 5 V to the ICs. A 6 V battery wouldn't work because the regulator needs at least 7 V, but a 9 V battery would. Those 4K RAM boards drew nearly an amp, but an IMSAI could supply 28 A of 8 V. A loaded IMSAI was toasty warm. The reason for on board regulation was when Roberts, Mims, et al. designed the Altair they wanted to save money on the power supply. The original Altair power supply was pretty puny. Most people who built Altairs eventually put in bigger transformers. Ed Roberts once called people who built and sold Altair bus microcomputers parasites, and one company that made a replacement power supply for Altairs called itself Parasitic Engineering. IMSAI, Digital Group, TDL, Cromemco, Processor Technology, Technical Design Labs, Cromemco, Morrow, Polymorphic Systems, Godbout and a bunch of others made S100 boards and systems 20 years ago. All gone now, I think. ###### Date: 22 Dec 98 12:16:19 -0800 From: "Charlie Gibbs" Subject: Re: How was a ROM constructed before the transistor era? References: <3673CD55.2D800B7E@stoneweb.com> <752057$pnc$1@news.wlg.netlink.net.nz> <367D8761.34B06FD7@stoneweb.com> <367F1B4E.D4B6AB2E@jkmicro.com> <2222.659T1010T6904519@sky.bus.com> <367f1aae.40233279@news.vip.net> Message-ID: <2063.660T990T7363863@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 45 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 22 Dec 1998 13:00:35 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.119 In article <367f1aae.40233279@news.vip.net> genew@vip.net (Gene Wirchenko) writes: >"Charlie Gibbs" wrote: >>But speaking of reprogramming remote units, I think of a friend who >>built a number of IMSAIs that were used as communications front-end >>processors. The way he'd get new programs out into the field was >>to write them into a RAM-4A board (a 4K S-100 RAM board - hey, it >>took a boardful of 2102s to hold 4K), attach a 6-volt battery to >>the power pads provided on the board, then power down the machine >>and remove the board. As long as the battery remained connected, >>it would hold the program while he went to the remote machine, >>powered it down, and dropped the board in. > > Nice! I have a few questions. > > 1) A SIX volt battery? Was it really 6V or was it several batteries >combined? Either way, more details please. It was one of those square 6-volt latern batteries. There was just enough room inside the IMSAI for it, so it could be used for battery backup as well (I think that was the original purpose of the external power pads). > 2) Didn't the memory operate on 5V or a) Was the extra volt within >tolerances? b) Did the memory card have a voltage regulator? I think the board had three regulators, and in the IMSAI they all got fairly warm. The IMSAI's S-100 bus provides 8 volts, which gets regulated down to 5 (it also has +/- 16-volt lines, which get cut to +/- 12 volts for the processor and serial ports). So a 6-volt supply was enough to keep the board warm and happy. Ironically, when I replaced the RAM-4As in my own IMSAI with a newer board (2K x 8 static RAMs to give 64K on a single board, wow), I found that its one regulator stayed cool even after I built it up to 62K (had to leave room for the boot ROM). I built it from a kit with 16K, and every week I'd stop by the Heath store on my way home from work and shell out $16 or so for another 2K, so it slowly grew in capacity. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### Message-ID: <36801C0D.6AE70814@jkmicro.com> From: Jim Stewart Organization: http://www.jkmicro.com X-Mailer: Mozilla 4.04 [en] (Win95; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? References: <3673CD55.2D800B7E@stoneweb.com> <752057$pnc$1@news.wlg.netlink.net.nz> <367D8761.34B06FD7@stoneweb.com> <367F1B4E.D4B6AB2E@jkmicro.com> <2222.659T1010T6904519@sky.bus.com> <367f1aae.40233279@news.vip.net> <2063.660T990T7363863@sky.bus.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 50 NNTP-Posting-Host: 168.150.253.147 X-Trace: news13.ispnews.com 914365543 168.150.253.147 (Tue, 22 Dec 1998 17:25:43 EDT) NNTP-Posting-Date: Tue, 22 Dec 1998 17:25:43 EDT Date: Tue, 22 Dec 1998 14:24:13 -0800 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newsfeed.cwix.com!209.44.33.119!hub1.ispnews.com!news13.ispnews.com.POSTED!not-for-mail Charlie Gibbs wrote: > > In article <367f1aae.40233279@news.vip.net> genew@vip.net > (Gene Wirchenko) writes: > > >"Charlie Gibbs" wrote: > > >>But speaking of reprogramming remote units, I think of a friend who > >>built a number of IMSAIs that were used as communications front-end > >>processors. The way he'd get new programs out into the field was > >>to write them into a RAM-4A board (a 4K S-100 RAM board - hey, it > >>took a boardful of 2102s to hold 4K), attach a 6-volt battery to > >>the power pads provided on the board, then power down the machine > >>and remove the board. As long as the battery remained connected, > >>it would hold the program while he went to the remote machine, > >>powered it down, and dropped the board in. > > > > Nice! I have a few questions. > > > > 1) A SIX volt battery? Was it really 6V or was it several batteries > >combined? Either way, more details please. > > It was one of those square 6-volt latern batteries. There was > just enough room inside the IMSAI for it, so it could be used > for battery backup as well (I think that was the original purpose > of the external power pads). > > > 2) Didn't the memory operate on 5V or a) Was the extra volt within > >tolerances? b) Did the memory card have a voltage regulator? > > I think the board had three regulators, and in the IMSAI they all > got fairly warm. The IMSAI's S-100 bus provides 8 volts, which gets > regulated down to 5 (it also has +/- 16-volt lines, which get cut to > +/- 12 volts for the processor and serial ports). So a 6-volt supply > was enough to keep the board warm and happy. As I recall, the 5 volt regulators had to have a minimum 1.5 volt delta between the input and output. They were typically LM7805 parts. This means with 6 volts in, the board would get 4.5 volts. Not enough to happily run, but maybe enough to maintain the contents of the memory. I don't doubt he got away with it, but it's not a solution that most engineers would approve of.... > Ironically, when I replaced the RAM-4As in my own IMSAI with a > newer board (2K x 8 static RAMs to give 64K on a single board, > wow), I found that its one regulator stayed cool even after I > built it up to 62K (had to leave room for the boot ROM). I built > it from a kit with 16K, and every week I'd stop by the Heath store > on my way home from work and shell out $16 or so for another 2K, > so it slowly grew in capacity. ###### From: dweber6940@aol.com (DWeber6940) Newsgroups: alt.folklore.computers Subject: Re: How was a ROM constructed before the transistor era? Lines: 19 NNTP-Posting-Host: ladder03.news.aol.com X-Admin: news@aol.com Date: 23 Dec 1998 02:38:16 GMT Organization: AOL http://www.aol.com References: <3673CD55.2D800B7E@stoneweb.com> Message-ID: <19981222213816.24755.00000227@ng59.aol.com> Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!portc02.blue.aol.com!audrey03.news.aol.com!not-for-mail Now I cannot assure the veracity of this, but my father, who worked on the development of the original ESS 101's in the late 50's and early 60's told it to me once. Bell Labs had the ROM problem big. These switches were to be designed with minimal down time(code upgrades counted as down time). They needed to assure that the code was stable and not overwritten and they need to be able to update it quickly. What they had for main store was core. The answer they came up with was small magnets in an array under the core plane. These forced the core to the correct alignment and a new code load involved just swapping the magnet plane. Remember these were the original ESS's and the market was only the largest of offices. Very few were expected to be built. Douglas Weber Douglas Weber ###### From: "Foxy" Newsgroups: alt.folklore.computers References: <3673B896.F2C58E2B@vnet.ibm.com> <367587b9.20293582@news.prosurfr.com> <3675B940.C470D74F@jkmicro.com> Subject: Re: How was a ROM constructed before the transistor era? Lines: 22 X-Newsreader: Microsoft Outlook Express 4.72.3110.1 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Message-ID: NNTP-Posting-Host: 204.49.63.58 X-Trace: news12.ispnews.com 914462329 204.49.63.58 (Wed, 23 Dec 1998 20:18:49 EDT) NNTP-Posting-Date: Wed, 23 Dec 1998 20:18:49 EDT Organization: ISPNews http://ispnews.com Date: Wed, 23 Dec 1998 13:32:12 -0600 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!hub1.ispnews.com!news12.ispnews.com.POSTED!not-for-mail Jim Stewart wrote in message <3675B940.C470D74F@jkmicro.com>... >John Savard wrote: > > > >Didn't the early ESS central office switches have a really perverse >memory composed of aluminium plates with ferrite chips embedded in >them? I seem to recall that the plates could be written in an external >device then installed in the switch where they were read-only. They were mylar cards about the same size as a standard IBM punched card with little shiney squares of something magnetizable on them. My understanding is (in the 1A ESS, at least) they couls be either read only or read/write depending on where they were installed. Then there was the old "piggyback twister" core replacement. Anyone remember that?