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
This is a multi-part message in MIME format. --------------ABB395498BF5CF3F48035D92 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit This can be done in two ways. 1) Use 3.1i FPGA Editor and edit the RAM contents directly, instead of recompiling with a new ucf INIT value, or, 2) Use XDL and change it in the xdl version by searching for the RAM block and changing the INIT string... Dave Hawke eugenir wrote: > How about any utility for configurating Virtex BLOCK RAM AFTER implementation? For example, > to change program of internal microcontroller. --------------ABB395498BF5CF3F48035D92 Content-Type: text/x-vcard; charset=us-ascii; name="dhawke.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for David Hawke Content-Disposition: attachment; filename="dhawke.vcf" begin:vcard n:Hawke;David Hawke tel;cell:(+44) 778 875 5002 tel;work:(+44) 870 7350 517 x-mozilla-html:TRUE org:<br><img src="http://www.xilinx.com/images/smvirtex.gif" alt="Xilinx"> version:2.1 email;internet:dhawke@xilinx.com title:XILINX Field Applications Engineer adr;quoted-printable:;;Xilinx Northern Europe=0D=0ABenchmark House;203 Brooklands road;Weybridge;; x-mozilla-cpt:;2672 fn:David Hawke end:vcard --------------ABB395498BF5CF3F48035D92--Article: 23251
This is a multi-part message in MIME format. --------------B2E22B8607798811FDF91E9F Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit La SACET est un laboratoire privé de R&D en transmissions numériques sans fils implanté à Rennes (Bretagne). Dans le cadre d'un projet de developpement de modem radio haut débit pour le secteur de la défense, SACET recherche un ingénieur expérimenté en conception fpga et ayant un background telecom. Les personnes interessées peuvent me retourner un mail afin de recevoir des infos complémentaires. Merci Francois. --------------B2E22B8607798811FDF91E9F Content-Type: text/x-vcard; charset=us-ascii; name="francois.hamon.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for francois.hamon Content-Disposition: attachment; filename="francois.hamon.vcf" begin:vcard n:Hamon;François x-mozilla-html:FALSE org:SACET adr:;;;;;; version:2.1 email;internet:francois.hamon@sacet.com x-mozilla-cpt:;19264 fn:François Hamon end:vcard --------------B2E22B8607798811FDF91E9F--Article: 23252
Not the host side, but the card side. The one listed on the Xilinx site is out of business. Thanks.Article: 23253
Brian Drummond wrote in message <4tsmks4nrtgp1u0h1nj0qkgcv5srh2gnpv@4ax.com>... >On Fri, 16 Jun 2000 15:27:30 +0300, Utku Ozcan <ozcan@netas.com.tr> >wrote: > >>erika_uk@my-deja.com wrote: >> >>> why the version 3.1i has been already issued ??????? >> >>I think 3.1i is only being shipped to North America. >>It might take 1 quarter to come to overseas. >> >>Utku > >It reached Scotland about a week ago, I guess the 3.1i wavefront should >be somewhere around Austria by now > >;-) Which means that it'll hit Tucson by January. BTW: that's the Alliance 3.1i, not Foundation, right? -- a ----------------------------------------- Andy Peters Sr Electrical Engineer National Optical Astronomy Observatories 950 N Cherry Ave Tucson, AZ 85719 apeters (at) noao \dot\ edu "A sufficiently advanced technology is indistinguishable from magic" --Arthur C. ClarkeArticle: 23254
"Larry Edington" <larrye_Spam_Me_Not_@centurytel.net> wrote: >My method is tack 4 leads down with a fine tipped iron to hold the chip in >place, or use SMT glue to hold it in position. > >Put a small bead of solder paste around the complete part at the far end of >the leads ( away from the component body ) > >Use a hot air iron to melt the paste. Done deal. I have also done this with and think it is by far the quickest/neatest/easiest way. I pre-tin the pads, I don't think I could get anything like a consistent bead of solder paste on them and would be concerned about solder balls being stuck behind the legs (and being able to see them). Solder paste is quite good for tinning the pads its runny flux seems to be active at lower temperatures than rosin cored solder. Make sure there is some flux on the part legs. Place the part very carfully on the pads, it will tend to slip off becuase of the solder beads on the pads. Hold it down with firm pressure from tweezers or something in the middle of the part then run round the legs with hot air - a mini reflow system. I use a hot 'air' nozzle from a portable gas soldering iron. I think it works significantly better than hot air because the exhaust gasses contain less (little?) oxygen and tend not to oxidise the legs or solder. Cheers, Terry.Article: 23255
Hi Peter, Peter Alfke <palfke@earthlink.net> wrote in message news:394D9E6D.D940FF11@earthlink.net... Thanks for the answer. I have some extra questions: > Then there are "dirty" tricks like clock gating that do miracles, but > require a deeper understanding of clock delay and hold-time issues. I hear so often about clock gating and all the "tricks" but I don't have deeper knowledge about, so where I should look for some articles/books which could point me in right direction how to design clock gated circuits. > In general: > Any node that moves consumes power. Make sure that there is no > unnecessary movement. Node that moves ?? I don't understand what do you mean by that. ------------------------------------------- - Domagoj - - Domagoj@engineer.com - -------------------------------------------Article: 23256
John Larkin wrote: > > I will *not* install special viewers or plugins everytime somebody sends > me a CD... Windows is fragile enough without adding this level of chaos. > > John Not to mention the possibility of viruses. For that reason I refused to install the 2.1i documentation brower when it asked for extra Java permissions - just copied the PDF files from the CD. It didn't help that the so-called security certificate timed out at the end of August and we didn't get our 2.1i copy until near the end of Sept.Article: 23257
Since I am in desperate need of a certain Xilinx VirtexE chip and Xilinx cannot help me for at least 3 weeks (I had already been waiting for 3 weeks before they told me, I had to go for a lower speed grade), I try this forum as a last resort: I need only one (1) XCV1000E-7 BG560, but I need it yesterday. Notice it is speed grade 7 (well 8 is also OK!). If, for some reason you should have a box full in your drawer I will gladly help you lighten your burden. Money's (almost) no object. Thanks in advance Younès LeroulArticle: 23258
can anybody tell me "what is the differnce b/w a Programmable Logic Device and a Microcontroller. Which of them is powerful.Article: 23259
zee <zkshan1@yahoo.com> wrote in message news:ee6ccf0.-1@WebX.sUN8CHnE... > can anybody tell me "what is the differnce b/w a > Programmable Logic Device and a Microcontroller. > Which of them is powerful. Mr Andraka, I believe you know the answer to this one. AlunArticle: 23260
Domagoj wrote: > Node that moves ?? I don't understand what do you mean by that. > Mayb that's bit colloquial: make that: ...any node that changes its voltage... ( moves up and down in voltage) PeterArticle: 23261
Yall, I'm still trying to get the hang of instantiating Xilinx macros/primitives in my VHDL designs. I've gotting IBUF/OBUF/BUFG to work, but am now having trouble with ADSU16 on a Virtex part. Has anyone else gotten this to work? My code looks something like this: component adsu16 port ( a: in std_logic_vector(15 downto 0); add: in std_logic; b: in std_logic_vector(15 downto 0); ci: in std_logic; co: out std_logic; ofl: out std_logic; s: out std_logic_vector(15 downto 0) ); end component; Then, I instantiate it as usual: adsu16_1 : adsu16 port map ( a, b, ci, co, ofl, s); However, when I Implement the design, ngdbuild complains that "logical block adsu16_1 of type adsu16 is unexapnded." Since this is a standard Virtex macro, and the virtex library is part of my project, shouldn't this work? Thanks for the help! Bryan Jones bryan@rice.eduArticle: 23262
zee wrote: > > can anybody tell me "what is the differnce b/w a > Programmable Logic Device and a Microcontroller. > Which of them is powerful. That's easy - both together is the (most) powerful. Microcontroller fetch opcodes, and decode/execute them one at a time. So, they are well suited to human interface tasks, and many control applications, but sub-mS, and certainly sub uS responses are very hard to achieve. PLDs have hardwired logic, so can be in 'many places at once', and are most comfortable in the sub uS world, and are driving down to the low nS end of the scale. Between the mS and uS worlds, there is some overlap. So, they are quite complementary, and that is why many large/new FPGAs promote uC soft/hard cores, and products from Triscend and Atmel already exist that have 8 bit uC, with ProgLogic on one die. Our web site has more info, on using SPLDs as uC expansion, and that will give you some more idea of the 'task split' between the two. -jg -- ======= 80x51 Tools & IP Specialists ========= = http://www.DesignTools.co.nzArticle: 23263
On Mon, 19 Jun 2000 20:40:42 +0100, Rick Filipkiewicz <rick@algor.co.uk> wrote: | | |John Larkin wrote: | |> |> I will *not* install special viewers or plugins everytime somebody sends |> me a CD... Windows is fragile enough without adding this level of chaos. |> |> John | |Not to mention the possibility of viruses. For that reason I refused to |install the 2.1i documentation brower when it asked for extra Java |permissions - just copied the PDF files from the CD. It didn't help that the |so-called security certificate timed out at the end of August and we didn't |get our 2.1i copy until near the end of Sept. Rick, the idea of FPGA software 'expiring' is terrifying. What do we do three years from now when we have to upgrade or fix a design, the part is no longer supported, and the tools have expired? WHY would any part vendor design tools that expire? Makes no sense to me. JohnArticle: 23264
I guess I was thinking that using the CE vs. might save power rather than asking if it used more power. But it sounds like the two methods are equivalent. From my past experiences, if you are doing HDL design, it can be hard to control whether the CE is used or not. Or in some cases like with FPGA Express which is the packaged VHDL compiler in the Foundation package, the CE can be overused. I found that if you put any type of conditional logic (if statements) inside the "if (risingedge(xxx))...endif" statement it would turn those other if statements in logic feeding the CE input. So it was not always easy to control when the CE got used or not. So if there is little difference in power consumption, I will structure all of my code to not use the CE unless I see a clear advantage in speed or LUT usage. Peter Alfke said in an email... > CE-inputs in Xilinx FPGAs really do nothing to the clock. They control > a multiplexer in front of the D input, so that the flip-flop loads its > existing value. Thus there is no advantage in implementing CE in a LUT > (and CE saves you either one or two LUT inputs ). > > XC4000, Spartan, and Virtex global clocks ar distributed along a > horizontal "spine" with "ribs" going up and down in each column. The > software is smart enough to drive a rib only if there is a clock > destination in this column. Vertically aligning register etc banks can > thus save power, and the carry structure automatically enforces that > kind of alignment for counters and accumulators. > > Peter Alfke, Xilinx Applications > ================================================ > Rickman wrote: > > > On this issue, I have never figured out about using the CE vs. gating > > the input with the LUT. Is there a difference in power consumption in > > the two cases? That is does a FF use any more power if it is enabled, > > but the input is the same as the current output so that it does not > > change state? -- Rick Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design Arius 4 King Ave Frederick, MD 21701-3110 301-682-7772 Voice 301-682-7666 FAX Internet URL http://www.arius.comArticle: 23265
John Larkin wrote: > On Mon, 19 Jun 2000 20:40:42 +0100, Rick Filipkiewicz <rick@algor.co.uk> > wrote: > |John Larkin wrote: > |> I will *not* install special viewers or plugins everytime somebody sends > |> me a CD... Windows is fragile enough without adding this level of chaos. > |> > |> John > | > |Not to mention the possibility of viruses. For that reason I refused to > |install the 2.1i documentation brower when it asked for extra Java > |permissions - just copied the PDF files from the CD. It didn't help that the > |so-called security certificate timed out at the end of August and we didn't > |get our 2.1i copy until near the end of Sept. > > Rick, > > the idea of FPGA software 'expiring' is terrifying. What do we do three > years from now when we have to upgrade or fix a design, the part is no > longer supported, and the tools have expired? WHY would any part vendor > design tools that expire? Makes no sense to me. > > John I don't think the software "expired" in either sense of the word. The software certificate is a cerification that the software is genuine and has not been tampered with. For example, it can't contain a virus. The ceritificates have an expiration date to allow for the possibility that they could be cracked after some amount of time. So they are periodically changed, just like a password should be. You do change your passwords every month, don't you??? On the other hand, the new license from Xilinx says you only rent the tools for a year. Then you must pay up or your license is dead. But I don't know that the tools stop working. This may be a moot point. If you really want to get support on software that is three years old, you are in trouble. I believe that all of the FPGA vendors are pretty reluctant to support old tools and many don't provide support for really old (>3 or so years) parts. I don't want to step on any toes here, so you can ask your FAE or salesperson about this. But from what I have read in this group, it is a true statement that you will be hard pressed to use the new 3.1 toolset to work XC4000 designs, for example. And of course you will get nothing but blank stares if you ask for support on the F5.x toolset (or whatever the last pre-Mx.x release was). -- Rick Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design Arius 4 King Ave Frederick, MD 21701-3110 301-682-7772 Voice 301-682-7666 FAX Internet URL http://www.arius.comArticle: 23266
Rickman wrote: > On the other hand, the new license from Xilinx says you only rent the > tools for a year. Then you must pay up or your license is dead. But I > don't know that the tools stop working. They do NOT stop working. Peter Alfke >Article: 23267
GB wrote: > I don't recall the App Note number, but there is an XAPP### > on the Xilinx web site which includes source code for > configuring an FPGA over the parallel port (in DOS). The > app note title is something like "Xilinx FPGA Configuration > using 8051 microcontroller". <snip> Xilinx In-System Programming Using an Embedded Microcontroller http://support.xilinx.com/apps/xapp.htm#xapp058 There is example code available that can be touched up and compiled for windows to drive the parallel port. However, the interface is serial (JTAG) and assumes you will save your data in xilinx binary svf (xsvf). RobertArticle: 23268
Hello, To your problem there might be two basic solutions: - A PLL To lock to a single tone buried in noise you have to use a PLL with an analog phase comparator (analog multiplier, analog switch or a mixer). The problem with a carrier buried in a populated spectrum is you can't pass it through limiters to square it, the useful signal would disappear (capture effect). Thence you must ensure the whole input signal (wanted and unwanted) is kept in the linear zone of the phase comparator. In that type of PLL to get a lock without special circuits you must have the command range of vco within loop banwidth (around 100Hz would be fine in your case). The obvious choice for vco would be a VCXO (voltage controlled crystal oscillator) the command range and stability are just fine. You can do a Colpitts vcxo with a varicap a (parallel resonant) crystal one inverting gate and some capacitors. - A crystal filter First thing to compute in a bandpass filter is the loaded Q=dF/Fo. Here Q=1.5 E6/500 = 6000. There is a rule of thumb to know if your filter could be synthetized, Qc of components > 10*Q with a butterworth or Qc of components > 10*Q with an elliptical type... Inductors are ruled out. You must use quartz crystals. For mention crystal elliptic filters does exist but those can only be designed by crystal manufacturers. Let us try to keep simple. You can use a single crystal between input and output, only you need to cancel the parallel capacitance of the crystal with a capacitor of the same value between load and a source of opposite phase (ct transformer). You just measure the crystal // cap with a multimeter. if one crystal is not enough you can add a second cell back to back. The frequency of the crystals must specified for serial resonance. Also you must add a wide band LC filter to cut out the spurious responses of the crystals. Those are only guidelines you may find many more falltraps I think. On PLLs you can find useful ANs at Motorola NS Philips etc... You might sought advice in amateur radio circles, I've seen a ladder crystal filter in a ARRL book I think. Look also at crystal manufacturers sites KVG, Tele Quarz etc... d. Nestor wrote: > > Thanks for everyone's replies. > > Here are some details of what I wish to do using the narrowband BPF: > > 1) I would like to synchronize to a tone of 1.5625MHz that is > generated from another node on a network. I do not have access to the > transmitting node, and the tone will always be present on the line. > 2) Because of noise on the channel, I need to perform very narrowband > filtering so that I may be able to distinguish and lock to the tone > only with a Phase-Locked Loop (PLL). > 3) I am thinking of at least a 35dB out-of-band rejection, and a > passband (3dB frequencies) in say +/-150ppm % of the center of > 1.5625MHz (or ~+/-250Hz relative to the center). > > Because I am fairly new to analog design I am not sure what would be > the best approach: elliptic filters, crystal/resonator filters? > > Also, if a single component solution does not exist, do you know of a > good reference or application note which explains how to build my own > narrowband BPF? I have a good reference for elliptic filters, but not > for the other kinds. > > Thanks in advance for any further suggestions. > > NestorArticle: 23269
dave_admin@my-deja.com wrote: > Great site , man . As I understand, ADD_WIDTH (2**ADD_WIDTH) defines the > length of the FIFO, and WIDTH defines width. Are all the architectures > working ? In FIFO_v7 you used " ADD_WIDTH : integer := 4;" in component > DPMEM. In all other architectures you used "ADD_WIDTH : integer := 4" . Why > the change ? Yes the ADD_WIDTH defines the length of the fifo and WIDTH the width (no. of bits" regarding the last change it is not a problem I used it to synthesize the code on a small FPGA and forgot to reupdate the code. You can also get the latest update from the OpenCores CVS you will find test bench and test vectors that you can modify to test the core > Is the code synthesizable ? The last architecture should be synthesizable FIFO_V7 the others may not. As I remeber all versions are working but with different size and speed but I prefer to use only the last arch. and if you are able to get the last files from the cvs please feel free to modify the code and test it and let me know with your changes and test results. Also if you implement it on HW Regards, Jamil Khatib OpenCores org http://www.opencores.org > > regards, > Dave. > > In article <39487175.46DFC43C@yahoo.com>, > khatib@ieee.org wrote: > > Try to visit my memory cores page at > > http://www.geocities.com/SiliconValley/Pines/6639/ip/memory_cores.html > > You can also check http://www.opencores.org and http://www.free-ip.com > > You can find Free Open design cores > > > > Let me know if you are going to use my FIFO core > > > > Regards > > Jamil Khatib > > OpenCores Organization > > > > dave_admin@my-deja.com wrote: > > > > > Hi, > > > > > > Can somebody give me an example of FIFO design with width >1 ? > > > If it is parameterizable, even better. > > > VHDL is preferable, but Verilog is fine too. > > > > > > regards, > > > Dave. > > > > > > Sent via Deja.com http://www.deja.com/ > > > Before you buy. > > > > > > Sent via Deja.com http://www.deja.com/ > Before you buy.Article: 23270
This is a multi-part message in MIME format. --------------A57626EBAB50B39199787FC1 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, When I try to access the XC4000X library in VHDL by this syntax :"library xc4000x;", I get following error message: `Library logical name XC4000X is not mapped to a host directory.` although the library is clearly added and specified in the project. Any ideas? Thanx in advance. Steven. --------------A57626EBAB50B39199787FC1 Content-Type: text/x-vcard; charset=us-ascii; name="sanders.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Steven Sanders Content-Disposition: attachment; filename="sanders.vcf" begin:vcard n:Sanders;Steven tel;cell:0476 582 612 tel;work:057 230 140 x-mozilla-html:FALSE org:imec ;DTS adr:;;Flanders Language Valley 44;Ieper;;8900;Belgium version:2.1 email;internet:sanders@imec.be fn:Steven Sanders end:vcard --------------A57626EBAB50B39199787FC1--Article: 23271
What are you doing that requires a -7, or is it just to make up for less than optimal design??? Younes Leroul wrote: > Since I am in desperate need of a certain Xilinx VirtexE chip and Xilinx > cannot help me for at least 3 weeks (I had already been waiting for 3 weeks > before they told me, I had to go for a lower speed grade), I try this forum > as a last resort: > I need only one (1) XCV1000E-7 BG560, but I need it yesterday. Notice it is > speed grade 7 (well 8 is also OK!). If, for some reason you should have a > box full in your drawer I will gladly help you lighten your burden. Money's > (almost) no object. > Thanks in advance > > Younès Leroul -- -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 or http://www.fpga-guru.comArticle: 23272
Hi, I was wondering if anyone had heard about some free cores for flotaing point operators such as addition, multiplication, divission in VHDL or Edif targeted over Xilinx Virtex series ? Thanks StevenArticle: 23273
FYI, I got a couple of helpful responses from Dr. Dave Van den Bout (devb@xess.com) and Stan Ramsden (sramsden@hoflink.com). With their thoughts, and a little more fooling around, I fixed the problem by doing the following. These instructions are for the Foundation Express 2.1i tools: 1. Open the Virtex library by double-clicking on it (I'm using a Virtex part), and copy the desired macro (adsu16) to my local project's library. 2. Right-click on the desired macro in the local library, then select Export Netlist | EDIF. 3. Now, re-run the Implement process, and everything works! Thanks for the help... Bryan Jones bryan@rice.eduArticle: 23274
Technically a CPLD problem, but I thought to post it here anyway. I am having a problem with a XC9572XL @ -40 degrees. It seems that when I cycle power, the JTAG controller becomes confused and takes over the chip. The only way I can reset the JTAG controller is to power down and up at room temp. Once the JTAG controller is confused, if I warm the part to room, it still misbehaves (TAP controller is in control). I do have a switching power supply on the board, so the +5v starts coming up about 6 ms before the 3.3v, but if I apply a cap between the 3.3 and +5 so the 3.3 starts coming up with the +5v, it still doesn't fix it. I tried shorting the TCK directly to gnd and through a 1.1K pulldown, and I tied the TMS to +5 through 4.7K and to 3.3 through a wire. I am really starting the suspect a ground bounce problem, or the TAP controller in the chip itself. Anyone else have similar experiences the the 9500XL family at cold? If so, how did you address it? Also, anyone know how Xilinx tests the industrial rated parts at cold? Any chance the JTAG controller is not tested? If I access the JTAG port in the right way at cold, the parts seems to come back to life, so maybe if the testing uses the JTAG port, then it wouldn't demonstrate the fault we are experiencing?
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