Site Home Archive Home FAQ Home How to search the Archive How to Navigate the Archive
Compare FPGA features and resources
Threads starting:
Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
I download the activeHDL 4.1 and xilinx verilog library,and install it successfully. But it seems the xilinx lib don't bind the activeHDL. When i configure the device in flow setting, the console display warning message: Library VIRTEX required for the selected device has not been installed. (I can see the xilinx lib in "Verilog Library" tab in "Design setting" menu). Further,when I run timing simulation,the console display: ELBREAD: Warning: No design unit found (library simprim_edif ,name x_or2). ELBREAD:Warning: Component .UUT.count_reg_1_GSR_OR :x_or2 not bound. etc. Is the situation lib problem? So what's the step to bind the lib to design? Regards Frank Sent via Deja.com http://www.deja.com/Article: 28001
Hi, You could use HCT- logic since it's inputs high-state is about 2.4V. I read this at some manufactures datasheets so that should be ok. At least Philips (and maybe others) have single logic ports in one tiny package (picogate). If your clock inputs are allready TTL- compatible, there's no problem to drive those signals directly. I have used this method in my mp3 player (still proto) where the decoder works with 3.3V and the dac (CS4334) needs 5V supply. According to dac's datasheets it's inputs are TTL-compatible. Works fine! Hope this helps. Tommi In article <3A3F3339.AC12CB9E@sqf.hp.com>, nials@sqf.hp.com wrote: > I'm looking at a problem where we need to drive a > couple of 5V CMOS clock inputs from a SpartanII > 3v output. > > There are data lines being driven from the 3V output, > we can get away with the 'tristate and pull high > for logic high' trick, but I don't want to do this with > the clock signals. The active (rising) edges are > _very_ slow and the risk of double clocking etc > would be too high. > > Space is fairly tight so my immediate thought was > to use an 8 pin soic dual comparator with the -ve > input tied to 1.8V (power plane). > > My only concern with this is that I think I read a > while ago that comparators shouldn't be used for this > sort of application, I can't remember where I read this > so I can't check if I'm right. It might have been > because of the lask of hysteresis on the input, > but if the -ve input is set to a 'clean' part > of the waveform I don't think we should see > any problems. > > Can anyone think of any drawbacks of using a fast > comparator for this conversion? > > Nial. > Sent via Deja.com http://www.deja.com/Article: 28002
In article <3A3DD507.F7214EBF@wp.pl>, Daniel =?iso-8859-1?Q?Ha=F1czewski?= <danhan@wp.pl> wrote: > Dear all, I allways wondered, whether it would be possible to configure an FPGA PCI interface using PCIs JTAG pins. This would allow you to get rid of the expensive configuration ROM. However, the JTAG pins are optional for the PCI standard and such a device would only work in mainboards that support jtag. Also, I do not think that there is a common way to access the jtag functions of a mainboard. So you probably would need to obtain seperate drivers for each mainboard. Also, as you only configure your board after your operation system has been booted, you can not use much of the plug and play functionality. (No big deal :-) Kolja > I'm developing a PCI card with Xilinx's XC9572 on board and I have a > problem. I would like to use PCI bus to program XC9572 via JTAG > interface. This solution would allow me to program the card without > opening computer case and in the future remotely with a use of a modem. > Sent via Deja.com http://www.deja.com/Article: 28003
hoyte@ucsu.colorado.edu wrote: > I recently worked on a senior project where we designed a 16-bit RISC > microprocessor, and implemented the design in an FPGA. I'd like to be > able to do something similar on my own, and I'm trying to find a good > FPGA/board combination that is (relatively) affordable, and compatible > with the Xilinx student edition software. If anyone has any suggestions, > they would be greatly appreciated. > I'll propose using an XS40 Board. You can see more about those by checking the links at http://www.xess.com/ho04000.html. You can find information about Jan Gray's XSOC CPU and associated C compiler that work with the XS40-005XL Board at http://www.fpgacpu.org. You can also look at the list of FPGA boards at http://www.optimagic.com. The list summarizes various features and prices so you can make a quick determination of suitable manufacturers. -- || Dr. Dave Van den Bout XESS Corp. (919) 387-0076 || || devb@xess.com 2608 Sweetgum Dr. (800) 549-9377 || || http://www.xess.com Apex, NC 27502 USA FAX:(919) 387-1302 ||Article: 28004
Hi all, I'm starting a new design using VIRTEX XCV600. I want to know if the methodology is the same than with a device like XCV 50 ? BFC Sent via Deja.com http://www.deja.com/Article: 28005
In article <3A377B68.26B507F1@algor.co.uk>, Rick Filipkiewicz <rick@algor.co.uk> wrote: > > > Mark Russell wrote: > > > How do I set a hold time constraint in virtex? > > > > I have a design where there are many input busses each with their own > > clock(too many to use dedicated clock resources). > > > > I can set up a setup constraint using OFFSET IN BEFORE, but how do I set > > a hold time requirement? > > > > Sent via Deja.com > > http://www.deja.com/ > > Basically you can't. Hold time is supposed to be taken care of, if a global > clock is used, by the design guarantees: > > (1) The global pin-pin su/hld characteristics of the IOB FFs. 1.5/-0.4 if > the DLL > is used and 2.1/0 if no DLL & input delay is set. > > (2) For any pair of FFs Tco(min) + routing(min) > Tgclkskew. > > Unfortunately this leaves open the case where you have an input from a pin > feeding an internal (not IOB) FF. Assuming 0 hold time at the pin, you need > to know whether the > > (3) No DLL: routing delay of the input is always > global clock routing, or > > (4) DLL: the routing delay > DLL jitter + phase delay. > > It would seem that (1), (2), (4) cover it *But* in the case of PCI the clock > frequency is allowed to vary or even stop so you can't use a DLL in systems > where this may happen. The Xilinx tools don't give a lot of help in > analysing case (3) esp. since the min delays for device families don't > usually get published for up to a year after device introduction - if ever. > The best I have been able to do is set an impossible MAXDELAY for the clock net. This obviously fails but gives a short clock track which helps the hold time. To improve the hold time further I would like to turn on the IOB delay. How do you turn it ON? My understanding is it defaults to on for registered inputs and may be turned off using NODELAY. What about non-registered inputs where the default seems to be off? Sent via Deja.com http://www.deja.com/Article: 28006
In article <3A3EBCB4.93ADDE5B@xilinx.com>, peter.alfke@xilinx.com says... <SNIP> > As I said, the gain of the slave latch makes that (almost) impossible. > But I also question the importance of the question. > For data, it is irrelevant whether it oscillates or not. Not if the signal has different delays to the following FF's (assuming synchronous design) and therefore is interpreted as '0' at some FF's and '1' at others. > And anybody using a metastable-prone signal as a clock should have his/her > head examined. Agreed. Søren A.Møller PS. I recommend 'The Commandments of Digital Design': The Commandments of Digital Design. =================================== I) Thou shalt have only one Master clock and thou shalt not build false idolatorous clocks from astables. II) Thou shalt not gate the clock, for that causeth false clock edges and skews. III) Thou shalt make all circuits synchronous unless thou canst convince he who pays thy salary, or assigns thy marks, that for reasons such as speed, pulse capture, or paper publishing, synchronous circuits cannot serve thy purpose. IV) Thou shalt not associate with undesirables such as ripple counters and one shots, but cultivate friendships with Johnson counters and enabled flip-flops. V) Thou shalt have a master reset for all filp-flops sp that the test engineer will love you, and you simulations will not remain undefined for time eternal. VI) Thou shalt not let analog ground mix with digital ground, for nothing but grief will come of such union. VII) He shall not be held blameless, who leavith CMOS inputs open. VIII) Asynchronous reset was not conceived for tasks such as returning a count to zero. Verily I say thee, that six circuitscreated after that manner will clear properly and bring honour to thy name, but the seventh shall fail and carry thee down in shame and disgrace. IX) Raw asynchronous inputs are unclean, and must be cleansed by passing through a single D flip-flop before they are allowed access to thy pure and chaste variables. X) Ye who completely understand the reasons for the commandments, then ye also know what liberties can be taken with them. Ye who would break them in ignorance, beware. ----- Dr. John Knight or http://www.cmc.ca/Design_Flows/commandments_digdesign.pdf It seems that the 10th commandment is missing.Article: 28007
Tomppa wrote: > You could use HCT- logic since it's inputs high-state > is about 2.4V. I read this at some manufactures datasheets > so that should be ok. At least Philips (and maybe others) > have single logic ports in one tiny package (picogate). Vinhigh for these two signals is 3.5V, and we're getting a 'glitch' at ~2.4V that would cause double clocking if that was the threshold. I also don't think you can get any 74XX series devices in an 8 pin package. Nial.Article: 28008
Announcing the new BED-SPARTAN2+ FPGA Prototyping Kit ***Hot*** - 200,000 gates!!! (Xilinx Spartan II device) - free Xilinx Webpack software CD included - introductory price US$120!!! Great for some serious prototyping, or for education. See http://www.burched.com.au/bedspartan2.html for full specs and secure online shop. Low cost plug-on modules also available - great companion modules for the BED-SPARTAN2+ kit: - BED-SRAM (2 MBit) - BED-FPGA-CPU-IO (for computer architecture experimenters) - BED-7SEG-DISPLAYS - BED-DIP-SWITCH See http://www.burched.com.au/products.html We currently have stock of all kits, but stocks are decreasing rapidly (even before this announcement). Reserve your kit for the new year by placing your order now! Our secure online shop is open 365 days of the year, and is accepting orders. Please note, however, that Burch Electronic Designs is closed until 2nd January 2001 - we will start processing orders again on that day, in the order that they were received over the break. International orders are very welcome. Best regards, and seasons greetings, Tony Burch www.BurchED.com.auArticle: 28009
Hi Eric, You may wish to consider the new, low cost BED-SPARTAN2+ FPGA Protoyping Kit from Burch Electronic Designs http://www.burched.com.au (Also see announcement posted on this group "New 200K gate, low cost FPGA proto kit") There are also some very neat Plug-On Modules that may be appropriate for you project: - BED-SRAM (2 MBit) - BED-FPGA-CPU-IO (for computer architecture experimenters) - BED-7SEG-DISPLAYS - BED-DIP-SWITCH See http://www.burched.com.au/products.html The Xilinx Webpack software CD comes in the box with the kit. Great free software, no license required! Supports Spartan II. You can also download this software for free from the Xilinx website. 200K gates is alot of gates :) Quite suitable for some serious computer architecture investigation, and for implementing your own RISC CPUs. As an aside, here's a link with some great work from the highly esteemed Jan Gray: http://www.fpgacpu.org Well worth a visit to this site. Good luck with your project! Best regards Tony Burch www.BurchED.com.au <hoyte@ucsu.colorado.edu> wrote in message news:91mdlq$pi6$1@nnrp1.deja.com... > I recently worked on a senior project where we designed a 16-bit RISC > microprocessor, and implemented the design in an FPGA. I'd like to be > able to do something similar on my own, and I'm trying to find a good > FPGA/board combination that is (relatively) affordable, and compatible > with the Xilinx student edition software. If anyone has any suggestions, > they would be greatly appreciated. > > Thanks, > Eric Hoyt > > > Sent via Deja.com > http://www.deja.com/Article: 28010
I have 2 questions about Xilinx pins used in high-frequency designs.<br> <br> First:<br> There are DFFs on the OE in VirtexE IOBs. These, however, are not used by the PAR, even if they are needed. When my colleague called Xilinx support, they said the only way was to use them was to use FPGA Editor and place them manually. On a 32-bit bus, however, this is hardly practical (esp. since you have to do it on each implementation).<br> <br> Second:<br> In another high-frequency design, I found it was necessary to put a FAST 12 mA DRIVE on an output pin. However, in the lab, I saw it created a 1V overshoot on the LVTTL pin, which would damage external components. I would like to reduce the drive, but PAR does not allow me to go below 12 mA: when I do attempt it, it tells me it cannot respect constraints. However, it says this assuming a 35 pF load, when in fact I a have a much lighter 5 pF load. Considering this load, I know I can go well below 12 mA, however PAR stops and refuses to go on despite violations. Is there any work-around?Article: 28011
In article <3A3F61AA.13A69394@sqf.hp.com>, nials@sqf.hp.com says... > Tomppa wrote: > > > You could use HCT- logic since it's inputs high-state > > is about 2.4V. I read this at some manufactures datasheets > > so that should be ok. At least Philips (and maybe others) > > have single logic ports in one tiny package (picogate). > > Vinhigh for these two signals is 3.5V, and we're getting > a 'glitch' at ~2.4V that would cause double clocking if that > was the threshold. > > I also don't think you can get any 74XX series devices in an > 8 pin package. No, but you can get them in SOT23-5 and SC70-5 from e.g. TI: http://www-s.ti.com/cgi-bin/sc/family3.cgi?family=SINGLE-GATES or Fairchild (they have some in US-8): http://www.fairchildsemi.com/products/logic/tinylogic/ or Philips: http://www.philipslogic.com/products/picogate/ http://www.philipslogic.com/products/picogate/overview/ or Toshiba or On-semi and probably others. Søren A.MøllerArticle: 28012
Hi: Assuming you always plan to program the FPGAs via an on-board microprocessor (such as a PowerPC), and you have many FPGAs (all different Xilinx) devices, why would you want to use a JTAG chain ? Wouldn't it make sense to use serial parallel mode and create seperate image files so that you can download any FPGA image you want at any time ? It also allows you to upgrade (in the field) any image without having to program/update them all. Comments ? Also, has anyone done any analysis of the image sizes and time to program a JTAG chain with many FPGAs or to not use JTAG at all and to program each image individually, one at a time ? Thanks, SteveArticle: 28013
Hi: Assuming you always plan to program the FPGAs via an on-board microprocessor (such as a PowerPC), and you have many FPGAs (all different Xilinx) devices, why would you want to use a JTAG chain ? Wouldn't it make sense to use serial parallel mode and create seperate image files so that you can download any FPGA image you want at any time ? It also allows you to upgrade (in the field) any image without having to program/update them all. Comments ? Also, has anyone done any analysis of the image sizes and time to program a JTAG chain with many FPGAs or to not use JTAG at all and to program each image individually, one at a time ? Thanks, Steve email: saz@sonusnet.comArticle: 28014
Hi: Assuming you always plan to program the FPGAs via an on-board microprocessor (such as a PowerPC), and you have many FPGAs (all different Xilinx) devices, why would you want to use a JTAG chain ? Wouldn't it make sense to use serial parallel mode and create seperate image files so that you can download any FPGA image you want at any time ? It also allows you to upgrade (in the field) any image without having to program/update them all. Comments ? Also, has anyone done any analysis of the image sizes and time to program a JTAG chain with many FPGAs or to not use JTAG at all and to program each image individually, one at a time ? Thanks, Steve email: saz@sonusnet.comArticle: 28015
Perhaps I was to abit to strong in my last reply but: > Well, how far can we drive the definition of "true dual port RAM". > Bob Dittmar thinks it should include an arbiter that resolves the > case when both ports access the same location "kind of > simultaneously". I never indicated that I thought it should include an arbitor. > Virtex BlockRAM has two totally independent ports accessing a common > storage array. > Both might access the same location. > If one writes and the other one reads, it depends on the relative > clock timing whether the reader gets the old or the new data. Note > that read is a synchronous, clocked operation. Each port has its own > clock. I indicated that I thought the above statement didnot reflect what is published. What is published is that the read port is invalid - not the new or the previous data value is what is read. > If both ports write, the one with the later clock wins. Definitely no > damage. > This operation is clear and unambiguous. Solving the contention in > dedicated hardware would not be in the user's best interest. > I agreee completely. I only posted response to original post because I am doing design with common clock where 1 side constantly reads only and the other side constantly writes. I had several designeers tell me that the Xilinx DPRAM could handle this without arbitration. I decided to verify it myself and found that not to be true (or so I believe) so I added arbitartion. Yet my initial arch did not have it. I would of found the problem in the lab at the expense of other developers time. So I was passing this along Regards, Bob Dittmar Sent via Deja.com http://www.deja.com/Article: 28016
Nial Stewart wrote: > I'm looking at a problem where we need to drive a > couple of 5V CMOS clock inputs from a SpartanII > 3v output. > > There are data lines being driven from the 3V output, > we can get away with the 'tristate and pull high > for logic high' trick, but I don't want to do this with > the clock signals. The active (rising) edges are > _very_ slow and the risk of double clocking etc > would be too high. > > Space is fairly tight so my immediate thought was > to use an 8 pin soic dual comparator with the -ve > input tied to 1.8V (power plane). > > My only concern with this is that I think I read a > while ago that comparators shouldn't be used for this > sort of application, I can't remember where I read this > so I can't check if I'm right. It might have been > because of the lask of hysteresis on the input, > but if the -ve input is set to a 'clean' part > of the waveform I don't think we should see > any problems. > > Can anyone think of any drawbacks of using a fast > comparator for this conversion? Yes- the comparator has an even greater chance of producing double clocking when driven directly off the line because of an abrupt input impedance change on switchover which interacts with the line impedance. The time-tested technique for clean clock reception is the Schmitt trigger with RC low pass filter at the input. Make the RC time constant be something on the order of, or a bit longer than, the Tpd for the Schmitt trigger. Usually a 100 ohm- 100pF will do just fine. This will be suitable for an actual CW clock as well as an asynchronous clock used to latch parallel data off a line which I take to be the application here. > > > Nial.Article: 28017
Hi, According to ds001_1.pdf, some Spartan2 -5 speed grades are available in the industrial temperature range -- this should mean that the worst-case timings are guaranteed at die temperatures -40 to 100°C instead of 0 to 85°C. What does the "Temperature Prorating" then mean in the WebPack Timing Analyzer? The Timing Analyzer allows Temperature Prorating down to -40°C. Can commercial grade parts safely be used down at these temperatures, and can I expect other effects than faster timings and larger power-up currents? Thanks a lot, Karl OlsenArticle: 28018
In article <3A3EBCB4.93ADDE5B@xilinx.com>, peter.alfke@xilinx.com wrote: > > > Greg Neff wrote: > (snip) > > > IOW, they are both functioning as high gain inverting > > amplifiers, with VOUT=VIN. In this case, I would not be surprised to > > see some brief oscillation before the inverter pair snaps back to > > stable digital operating area states. > > No, I don't see how such a primitive structure can sustain an oscillation. > Once it leaves its metastable balanced state, there is no way to return > back to or through it. ( As is possible in multi-stage structures popular > with TTL technology) > I'm thinking that the phase delay of noise amplified through the inverters may permit a very brief low amplitude oscillation around Vth before the latch stabilizes. Depending on the structure of the flip- flop, this oscillation could be amplified by the next stage. This is hypothetically speaking of course, I'm not saying that Xilinx FPGAs do this. (snip) > But I also question the importance of the question. Maybe not important from a practical perspective, but interesting none the less. I just wanted a little clarification on your statement in your reply to the OP. > For data, it is irrelevant whether it oscillates or not. Irrelevant only if the data has one destination, hence the need for two- stage synchronizers. > And anybody using a metastable-prone signal as a clock should have his/her > head examined. Absolutely right! I was not, in any way shape or form, implying that this should be done. -- Greg Neff VP Engineering *Microsym* Computers Inc. greg@guesswhichwordgoeshere.com Sent via Deja.com http://www.deja.com/Article: 28019
In article <ee6f135.-1@WebX.sUN8CHnE>, "Pascal C." <> wrote: (snip) > Second:<br> > In another high-frequency design, I found it was necessary to put a FAST 12 mA DRIVE on an output pin. However, in the lab, I saw it created a 1V overshoot on the LVTTL pin, which would damage external components. I would like to reduce the drive, but PAR does not allow me to go below 12 mA: when I do attempt it, it tells me it cannot respect constraints. However, it says this assuming a 35 pF load, when in fact I a have a much lighter 5 pF load. Considering this load, I know I can go well below 12 mA, however PAR stops and refuses to go on despite violations. Is there any work-around? > First of all, make sure that all FAST outputs are properly terminated. With your 5pf load you should expect rise and fall times in the order of 0.2ns, so proper termination is *required*. Second, how did you measure the overshoot? If you are using a typical scope probe with a 4" ground lead then you are not seeing the real signal. You need a low capacitance active probe, with a very short (<1/2") ground lead. Also, with 0.2ns edge rates, you need a scope with at least 5GHz bandwidth to make reasonable measurements. -- Greg Neff VP Engineering *Microsym* Computers Inc. greg@guesswhichwordgoeshere.com Sent via Deja.com http://www.deja.com/Article: 28020
"There are data lines being driven from the 3V output, we can get away with the 'tristate and pull high for logic high' trick, but I don't want to do this with the clock signals. The active (rising) edges are _very_ slow and the risk of double clocking etc would be too high." Something worth considering would be a modified version of this method. I have had great success driving a clock line by configuring the logic so that it drives the output pad high until it is a high level and then tristate and let the pullup work the rest. This allows the output to be driven all of the way up to 3V, creating some much faster slew rates. Xilinx clains that this can decrease the rise time from 0.4 to 3.0V from 20ns to 3ns. I have seen this to be true on a spartan-2. See: http://www.xilinx.com/products/virtex/techtopic/5volt.htm http://www.xilinx.com/products/virtex/techtopic/vtt002.pdf Jason Daughenbaugh http://www.aedinc.netArticle: 28021
In article <3A3F3339.AC12CB9E@sqf.hp.com>, nials@sqf.hp.com wrote: > I'm looking at a problem where we need to drive a > couple of 5V CMOS clock inputs from a SpartanII > 3v output. > (snip) I like to have a reel of these on hand: http://www.fairchildsemi.com/pf/NC/NC7ST86.html They can be used as inverters or buffers, depending on how you strap the other input. Other manufacturers (such as Toshiba) make similar parts. -- Greg Neff VP Engineering *Microsym* Computers Inc. greg@guesswhichwordgoeshere.com Sent via Deja.com http://www.deja.com/Article: 28022
I don't think any terminations were put on those pins. The measure was taken with an active probe, at 5 GS/s sampling.Article: 28023
"Søren A.Møller" wrote: , peter.alfke@xilinx.com says... > <SNIP> > > As I said, the gain of the slave latch makes that (almost) impossible. > > But I also question the importance of the question. > > For data, it is irrelevant whether it oscillates or not. > > Not if the signal has different delays to the following FF's (assuming > synchronous design) and therefore is interpreted as '0' at some FF's and > '1' at others. > I should have said: Oscillation or unknown delay can lead to exactly the same bad result. So, oscillation is no worse than an unpredictably long delay. I think we agree. Nice list of commandments. Especially #10! Peter AlfkeArticle: 28024
I am sure that the reborn Fairchild ( I am an old Fairchilder of 30 years ago!) has circuits with an input threshold of 2.4 V and available in really tiny packages. Peter Alfke Nial Stewart wrote: > Tomppa wrote: > > > You could use HCT- logic since it's inputs high-state > > is about 2.4V. I read this at some manufactures datasheets > > so that should be ok. At least Philips (and maybe others) > > have single logic ports in one tiny package (picogate). > > Vinhigh for these two signals is 3.5V, and we're getting > a 'glitch' at ~2.4V that would cause double clocking if that > was the threshold. > > I also don't think you can get any 74XX series devices in an > 8 pin package. > > Nial.
Site Home Archive Home FAQ Home How to search the Archive How to Navigate the Archive
Compare FPGA features and resources
Threads starting:
Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z