From: "kris" Newsgroups: comp.arch.fpga Subject: accurate power measurements Date: Thu, 8 May 2003 17:03:40 -0700 Organization: University of California, Los Angeles Lines: 10 Message-ID: NNTP-Posting-Host: ambiorix.ee.ucla.edu X-Trace: gladiola.noc.ucla.edu 1052438622 22249 128.97.92.134 (9 May 2003 00:03:42 GMT) X-Complaints-To: abuse@ucla.edu NNTP-Posting-Date: Fri, 9 May 2003 00:03:42 +0000 (UTC) X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeed.icl.net!newsfeed.fjserv.net!logbridge.uoregon.edu!canoe.uoregon.edu!newsfeed.news.ucla.edu!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:28441 I want to measure the instantaneous power consumption of the internal core of an FPGA. Does anybody know of any testboards where one can do this? Like do there exist testboards where I can directly supply this voltage myself and measure the current through the lines. Or maybe there are boards where there are or can be put small resitances in the the supply lines just before the FPGA. Thanks! ###### Message-ID: <3EBAF293.10EC6D64@andraka.com> From: Ray Andraka Organization: Andraka Consulting Group, Inc X-Mailer: Mozilla 4.77 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: accurate power measurements References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 27 Date: Fri, 09 May 2003 00:09:29 GMT NNTP-Posting-Host: 68.15.41.165 X-Complaints-To: abuse@cox.net X-Trace: news1.east.cox.net 1052438969 68.15.41.165 (Thu, 08 May 2003 20:09:29 EDT) NNTP-Posting-Date: Thu, 08 May 2003 20:09:29 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeed.icl.net!newsfeed.fjserv.net!colt.net!diablo.theplanet.net!newspeer1-gui.server.ntli.net!ntli.net!east.cox.net!cox.net!p01!news1.east.cox.net.POSTED!53ab2750!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:28426 The ISI Osiris board has a monitor port in the regulator circuit that could be used to monitor current I think. You'd have to write an application to use it though. kris wrote: > I want to measure the instantaneous power consumption of the internal core > of an FPGA. Does anybody know of any testboards where one can do this? Like > do there exist testboards where I can directly supply this voltage myself > and measure the current through the lines. Or maybe there are boards where > there are or can be put small resitances in the the supply lines just before > the FPGA. > > Thanks! -- --Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email ray@andraka.com http://www.andraka.com "They that give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -Benjamin Franklin, 1759 ###### From: hmurray@suespammers.org (Hal Murray) Newsgroups: comp.arch.fpga Subject: Re: accurate power measurements Date: Fri, 09 May 2003 01:34:18 -0000 Message-ID: X-Newsreader: xrn 9.02 Sender: murray@glypnod (Hal Murray) References: X-Complaints-To: abuse@supernews.com Lines: 37 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!sn-xit-03!sn-xit-01!sn-post-01!supernews.com!corp.supernews.com!glypnod!hmurray Xref: chonsp.franklin.ch comp.arch.fpga:28443 >I want to measure the instantaneous power consumption of the internal core >of an FPGA. Does anybody know of any testboards where one can do this? Like >do there exist testboards where I can directly supply this voltage myself >and measure the current through the lines. Or maybe there are boards where >there are or can be put small resitances in the the supply lines just before >the FPGA. It's a hard problem. Maybe even very hard. What do you mean by "instantaneous"? Microsecond? Nanosecond? Do you want to measure an existing design, or do you want to write code and try various things? How hard are you willing to work and/or how much are you willing to pay? What sort of tools/equipment do you have? The small resistors you suggested are not common. Where would you put them? Between the FPGA and the bypass caps? Then you get crappy bypassing. If you put them just outisde the caps, then you can't measure the high speed currents because the bypass caps hide it from you. For a low cost project, I'd suggest setting things up so you don't need to measure the "instantaneous" power. Then you can measure it in the obvious way, say with a meter between the power supply and your board. (Doesn't work if the regulators are on the board.) But now you have to write some code to do things that are interesting to measure, and do it for long enough so that you can measure it. -- The suespammers.org mail server is located in California. So are all my other mailboxes. Please do not send unsolicited bulk e-mail or unsolicited commercial e-mail to my suespammers.org address or any of my other addresses. These are my opinions, not necessarily my employer's. I hate spam. ###### From: "kris" Newsgroups: comp.arch.fpga Subject: Re: accurate power measurements Date: Thu, 8 May 2003 20:09:51 -0700 Organization: University of California, Los Angeles Lines: 27 Message-ID: References: NNTP-Posting-Host: ambiorix.ee.ucla.edu X-Trace: gladiola.noc.ucla.edu 1052449792 29699 128.97.92.134 (9 May 2003 03:09:52 GMT) X-Complaints-To: abuse@ucla.edu NNTP-Posting-Date: Fri, 9 May 2003 03:09:52 +0000 (UTC) X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.news.ucla.edu!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:28439 With instantaneous current I mean sampled at e.g. 250-500MHz, (4-2 ns) which would result, depending on the clock speed at 10 to 20 samples per clock cycle. I'm not interested in the mean or average power consumption. In our lab, we have some oscilloscopes available that can handle this and higher sampling frequencies. The boards that we have available now (insight virtex 2, xc2v1000 and the avnet virtex 2 development kit) both have the voltage regulators on the board itself. Can you recommend a board where I can provide vcc_int from outside the board? (I would prefer with virtex 2 devices as the design is completely finished). The hw-afx boards from xilinx seem to offer this possibility. But I cannot find the datasheets in order to confirm this and in order to see if the decoupling capacitances are already on the board or not. But by using such a board this current doesn't correspond with the current that flows into the fpga. Do only the decoupling capacitances have their influence or also the line inductances of the traces on the board? Thanks! ###### From: hmurray@suespammers.org (Hal Murray) Newsgroups: comp.arch.fpga Subject: Re: accurate power measurements Date: Fri, 09 May 2003 04:44:47 -0000 Message-ID: X-Newsreader: xrn 9.02 Sender: murray@glypnod (Hal Murray) References: X-Complaints-To: abuse@supernews.com Lines: 68 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!newsfeed.stueberl.de!newspeer1-gui.server.ntli.net!ntli.net!sn-xit-02!sn-xit-06!sn-post-02!sn-post-01!supernews.com!corp.supernews.com!glypnod!hmurray Xref: chonsp.franklin.ch comp.arch.fpga:28445 Have you looked at XAPP623 and similar documents? Are you familiar with Johnson and Graham's Black Magic book? (If you want to do this sort of measurement, you need to sleep with it under your pillow.) >With instantaneous current I mean sampled at e.g. 250-500MHz, (4-2 ns) which >would result, depending on the clock speed at 10 to 20 samples per clock >cycle. I'm not interested in the mean or average power consumption. I think that's going to be hard to measure at that time scale. >In our lab, we have some oscilloscopes available that can handle this and >higher sampling frequencies. > >The boards that we have available now (insight virtex 2, xc2v1000 and the >avnet virtex 2 development kit) both have the voltage regulators on the >board itself. Can you recommend a board where I can provide vcc_int from >outside the board? (I would prefer with virtex 2 devices as the design is >completely finished). The hw-afx boards from xilinx seem to offer this >possibility. But I cannot find the datasheets in order to confirm this and >in order to see if the decoupling capacitances are already on the board or >not. What good is a board without bypass caps? (Why would anybody build or ship one?) You can unsolder the regulator chip and provide your own core power, perhaps including a small sense resistor. I'd probably put a big cap with low ESR on the non-FPGA side of the sense resistor to make sure the wires over to the power supply don't confuse things. This gets you closer to the chip. Or perhaps you can just lift the leg on the regulator chip and insert a sense resistor between leg and board. Can you single step your design? (Or patch it so you can?) You might get some info by watching the voltage as you kick the clock. You would get a bigger signal if you let it run for 10 cycles. You can also make the signal bigger by removing some of the bypass caps. If you can measure a step change on a single clock, then you can compute the charge that got pulled out of the bypass caps and convert that back to current at a given clock rate. That's assuming the regulator chip doesn't have time to change the amount of current it is providing. That's probably valid, but check the step response graphs in the data sheet. >But by using such a board this current doesn't correspond with the current >that flows into the fpga. Do only the decoupling capacitances have their >influence or also the line inductances of the traces on the board? Generally, the whole idea of the power supply system is to avoid inductances in the power/ground signals so the bypass caps will do their job. -- The suespammers.org mail server is located in California. So are all my other mailboxes. Please do not send unsolicited bulk e-mail or unsolicited commercial e-mail to my suespammers.org address or any of my other addresses. These are my opinions, not necessarily my employer's. I hate spam.