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
Hi! Thanks for the reply. Actually I am describing the circuit from a high level environment that I developed. It is a lot easier for me to use a single LFSR. Using multiple LFSRs needs extra development of the my tools to generate the proper layout ( I am generating placement information myself from a high level programming environment which generates EDIF netlist). That's why I wanted to use high fanout signals. May be I'll do an extra effort! - use multiple LFSRs-.Article: 15076
On Thu, 4 Mar 1999 00:13:25 +0100, mesenich@t-online.de (Gerhard Mesenich) wrote: >Hi >I intend to start to start with isp CPLD designs. > >My designs are usually quite simple. They are mainly intended for simple >state machines, address decoding, input/output buffering of analog >converters and input capture of digital signals. They will mainly be used >with 16bit DSP's. Packages should ideally be PLCC 44/84 (easy handling on >experimental boards). I am already familiar with small PAL's (20v8 etc.). > >I have looked already at several chips. These are: > >Xilinx - XC 95xx >Altera - MAX 7000 >Lattice - ispl.. > I have used the Lattice devices in the 2000 series: ispLSI2032, 2096, and 2128. They work well and the ispDS+ software is easy to use. Well, at least the ABEL compiler is. I haven't tried schematic capture or VHDL with it. I also have the Xilinx Foundation dev system but have not used it for an actual product yet, so I can't give you a fair comparison. >My applications are low volume/ research type. I do design work only very >infrequently, so the design software should be as simple as possible and >somewhat foolproof. A simple homebrewed parallel port cable should hopefully >do the programming. > The Synario development system comes with a parallel download cable. So does the Foundation dev system from Xilinx. >Are there any suitable free integrated software packages? > Not that I know of, but the basic dev software has gotten pretty cheap. I got the Xilinx Foundation Base 1.4 for $99 U.S. It does not include VHDL - that is an extra cost add-on. >Which family/ software package is best suited to my need? What are the >approx. prices and how is availability? How well does schematic capture >work - is it useful? > Have not used the Lattice schematic capture. The Xilinx schematic editor is crude and annoying, but it works. Best regards, Ivan Baggett BagotronixArticle: 15077
Has any one seen (or done) a multi-port parallel to SPI interface? Thanks, Rinzai J. BellArticle: 15078
Hello there, This message is intended for mexicans working with FPGAs or VLSI architectures for their graduate research. Please forward if you know somebody. Thank you. Miguel ----- Hola: Estoy trabajando en el INAOE-Puebla con tecnologias FPGA (Field Programable Gate Arrays) y Computo Reconfigurable Dinamicamente, para el diseno de arquitecturas digitales orientadas a la vision por computadora. Estamos construyendo un sitio sobre diseno e investigacion en FPGAs y arquitecturas VLSI en Mexico y queremos contactar a gente estudiando posgrados en dicha area (en Mexico o en el extranjero) para montar un directorio. Por favor, la gente involucrada podria mandarme los siguientes datos? ----------------------------- Nombre: Programa y grado de estudio: Fecha estimada de titulacion: Universidad y Departamento: Correo Electronico: URL: Titulo de la Investigacion: Fabricantes de FPGAs que utiliza o mencionar si se trabaja en otros aspectos relacionados con FPGAs/VLSI (como CAD, paradigmas de computo reconfigurable, nuevas arquitecturas para FPGA, etc.): Areas de interes con los FPGA: ------------------------------- Gracias y saludos desde Puebla Miguel Arias Estrada ----------------------------------------------------------- Miguel Arias-Estrada | Researcher/professor ----------------------------------------------------------- INAOE - Electronics Dept. | Tel: +52 (22) 47-2011 #1410 Apdo. Postal 51 y 216 | Fax: +52 (22) 47-0517 72000 Puebla, Pue. | ariasm@inaoep.mx MEXICO | http://cseg.inaoep.mx/~ariasm -----------------------------------------------------------Article: 15079
I have 16 Xilinx 4010 FPGA in a PCB board, they used same SPROM, (same design). Can these FPGA share same SPROM when power on configuration? Could any one can show me how to do that? Thanks! J.P.Liao QtSArticle: 15080
Working with the new Xilinx Virtex family has made me curious about all the different I/O standards that it support. The Virtex supports the following standards: LVTTL LVCMOS2 PCI GTL GTL+ HSTL Class I HSTL Class III HSTL Class IV SSTL3 Class I and II SSTL2 Class I and II CTT AGP The electrical properties of these standards are described in the Virtex datasheet and XAPP133 "Using the Virtex SelectIO". What I would like to do now is gathering some information on the history and typical use of these standards. Searching the Internet gave me very little information on the newer standards, so I was hoping that you out there could help me to locate this information. Newer standards not mentioned above are also of interest. The result of this survey will be published at http://freecore.com Regards, Rune BaeverrudArticle: 15081
Hi: Welcome at club of crashing-minds. It's no easy what you want to do. The assignment must be do from the "Floorplan editor". Then you can assing pin groups to LABS. It's convenient what you divide your logic core in sections, and you must intend to distribute it in the LABS. For example: if you have a 8 bit input data bus, you must assign this inputs to LAB A (or B or C ...), but in one LAB. Then you don't collapse the LABS interconnections. Luck. Only experience can help you. One note: You must begin to assign a pin group, the you must compile the design and you must probe that it's all right. Then you must assign another pin group and so on. Bye!!.Article: 15082
Hi! Thanks for the reply. Actually I am describing the circuit from a high level environment that I developed. It is a lot easier for me to use a single LFSR. Using multiple LFSRs needs extra development of the my tools to generate the proper layout ( I am generating placement information myself from a high level programming environment which generates EDIF netlist). That's why I wanted to use high fanout signals. May be I'll do an extra effort! - use multiple LFSRs-.Article: 15083
You can try the Xilinx CPLD starter kit for $100. Available from insight electronics, includes almost everything you need. Gerhard Mesenich wrote in message <7bkfp4$9ht$1@news05.btx.dtag.de>... >Hi >I intend to start to start with isp CPLD designs. > > >with 16bit DSP's. Packages should ideally be PLCC 44/84 (easy handling on >experimental boards). I am already familiar with small PAL's (20v8 etc.). > >I have looked already at several chips. These are: > >Xilinx - XC 95xx >Altera - MAX 7000 >Lattice - ispl.. > > >Are there any suitable free integrated software packages? > >Which family/ software package is best suited to my need? What are the >approx. prices and how is availability? How well does schematic capture >work - is it useful? > In a couple of days you will get the hang of it. The thing really works! UdayArticle: 15084
"J.P.Liao" wrote: > I have 16 Xilinx 4010 FPGA in a PCB board, they used same SPROM, (same > design). > Can these FPGA share same SPROM when power on configuration? > Could any one can show me how to do that? > > Thanks! > J.P.Liao > QtS You should be able to configure all 16 devices from a common sprom set. I can think of two approaches. The first is a simultaneous serial load and the second uses a daisy chained approach. If all 16 FPGAs are the same design I think you could use a single, globally connected sprom. I would make 1 FPGA a serial master and the other 15 serial slaves. Connect all the programming clocks together. Connect the sprom data out pin to all FPGA DIN pins. Connect the proper control signals from the master FPGA to the sprom and other slave FPGAs. This includes LDC* (to sprom) and INIT* (to sprom and other FPGAs) (see figure 52 in the Xilinx 1998 data book). Once the configuration process starts the master FPGA provides a common programming clock and all FPGA receive the programming data from the single sprom. I haven't attempt this approach and with 16 loads you may need to worry about signal fan out. Another, lower risk approach uses the Master/Slave Serial Mode. This is detailed in the Xilinx data book (for the 4000 series see Figure 52, page 4-61 in the 1998 Data Book). Basically the FPGA configuration data path is serially daisy chained. The size of the sprom(s) will be 16 * 178096 bits for X4010E or 16 * 283376 bits for the XC4010. If you can't find a sprom big enough you can also daisy chain sproms. The Xilinx tool set has a sprom formatter where the individual FPGA images are combined and assigned to the sprom. A single programming file is the end result. I hope this helps. Regards, Paul T. Shultz Chesapeake Sciences Corporation <paul@csciences.com>Article: 15085
In article <7bo632$ku4$1@romeo.dax.net>, "Rune Baeverrud" <fpga@iname.com> wrote: > Working with the new Xilinx Virtex family has made me curious about all the > different I/O standards that it support. The Virtex supports the following > standards: Another feature that should be possible: Why not use the Vref and the flexible SelectIO inputs for implementation of voltage comparators for analog signal with in voltage range. This should make it possible building simple A/D converters(using D/A and PWM outputs) without needs of external IC's. Any opinions? Or is this impossible? Regards, Tryggve Mathiesen -----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your OwnArticle: 15086
If you are already generating layout, why can't you just repeat the LFSR layout with an offset? Khaled benkrid wrote: > Hi! > > Thanks for the reply. Actually I am describing the circuit from a high level > environment that I developed. > It is a lot easier for me to use a single LFSR. Using multiple LFSRs needs > extra development of the my tools to generate the proper layout ( I am > generating placement information myself from a high level programming > environment which generates EDIF netlist). That's why I wanted to use high > fanout signals. May be I'll do an extra effort! - use multiple LFSRs-. -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email randraka@ids.net http://users.ids.net/~randrakaArticle: 15087
In article <7bok8u$hjl$1@nnrp1.dejanews.com>, <tryggvem@my-dejanews.com> wrote: >In article <7bo632$ku4$1@romeo.dax.net>, > "Rune Baeverrud" <fpga@iname.com> wrote: >> Working with the new Xilinx Virtex family has made me curious about all the >> different I/O standards that it support. The Virtex supports the following >> standards: > >Another feature that should be possible: >Why not use the Vref and the flexible SelectIO inputs for implementation >of voltage comparators for analog signal with in voltage range. > >This should make it possible building simple A/D converters(using D/A and PWM >outputs) without needs of external IC's. > >Any opinions? Or is this impossible? > Unfortunately this is not possible within the Virtex device. The diff-amp/voltage comparator input buffers do not have rail-to-rail capability and so they are not capable of discriminating across the full 0-3.3V range. EdArticle: 15088
On Thu, 04 Mar 1999 19:26:51 GMT, Todd Kline <todd@wgate.com> wrote: >Actually, I'd love to use the ARC core. It is hands down my first choice >for a soft core. Unfortunately, my boss balks at the up-front license fee >of soft cores, even the "relatively" low fee for the ARC. We are a >start-up, so cash flow management is extremely important. Todd, your email address bounced. Please mail me on stuart 'at' saros 'dot' co 'dot' uk Thanks Stuart For Email remove "NOSPAM" from the addressArticle: 15089
Put one device in the serial master mode ( This device will give the serial clock ) put the others in the serial slave mode and connect all datalines in together ( do not daisy chain the data ). J.P.Liao wrote: > I have 16 Xilinx 4010 FPGA in a PCB board, they used same SPROM, (same > design). > Can these FPGA share same SPROM when power on configuration? > Could any one can show me how to do that? > > Thanks! > J.P.Liao > QtSArticle: 15090
In article <7bkfp4$9ht$1@news05.btx.dtag.de>, Gerhard Mesenich <mesenich@t-online.de> writes >Hi >I intend to start to start with isp CPLD designs. [snip] >Thanks a lot for any help, opinion and recommendation. > The programmable logic jump station at www.optimagic.com is a good place to start hunting. -- Kindest Regards | gerry@devantech | We manufacture Pic programmers, 8031, Gerald Coe | .demon.co.uk | 68302, 64180, 80C188EB cpu modules. http://www.devantech.demon.co.uk | Full custom uP control systems designed.Article: 15091
J.P.Liao wrote: > I have 16 Xilinx 4010 FPGA in a PCB board, they used same > SPROM, (same > design). > Can these FPGA share same SPROM when power on > configuration? > Could any one can show me how to do that? > Very simple: One of your XC4010s gets its mode pins connected as Master Serial, all the other ones as Sleve Serial. Then you interconnect all CCLKpins and connect them to the SPROM CLK, and you interconnect all DIN pins and connect them to the SPROM Data pin. You also interconnect all INIT pins and use them as the SPROM active Low RESET. Considering the high fan-out, it may be wise to stay with the slower CCLK rate of 1 MHz. Good luck, this is easy. Peter Alfke, Xilinx ApplicationsArticle: 15092
Andy Peters wrote: > > I'm not sure you can have a chip with more than one top-level module! > > Why would you want to do that???? > > -andy I believe he said in earlier postings that he was instantiating these modules in a seperate schematic diagram as black boxes. Then that is the top level from the device point of view. The trick is to get the synthesis tool to understand that the VHDL is no the top level... sort of like "macro mode" in Exemplar Leonardo. I don't know how that is done in FPGA Express, and apparently neither does Sergio. That is the nature of the question now. -- Brian C. Boorman Harris RF Communications Rochester, NY 14610 XYZ.bboorman@harris.com <Remove the XYZ. for valid address>Article: 15093
Since I also scrapped the gui in favor of an automatic batch file, I am beginning to wonder how many others have done the same thing. Since I know Xilinx monitors this news group, I would be interested in generating some user ffedback to them about the gui interface and it's lack of flexibility for the power users. I personally found the attempt at version management pretty bad and the gui hard to use because of it (maybe they could provide a method to TURN IT OFF!). I also found it a pain to find the options I wanted to change buried in the various option dialogs, and there are just too many buttons to press to get from edif file to Intel Hex file. I like to start the batch file running (which I built in error checking), walk away, and come back in half-hour to see if it done. My time is too valuable to sit there and babysit a GUI just to press buttons. Jeff Hunsinger wrote: <snip> > .....I quickly > gave up on it and just run the whole thing from the command line with a > make file. > -- Brian C. Boorman Harris RF Communications Rochester, NY 14610 XYZ.bboorman@harris.com <Remove the XYZ. for valid address>Article: 15094
Steve Dewey wrote:.... > A good example is the 10 bit counter you need. Typically you might have to wire > that up using 3 4-bit counters, if restricted to TTL 74 series style functions. > The alternative is just to use a LPM_COUNTER WITH (LPM_WIDTH=9). If your > requirment changes, then just change the LPM_WIDTH value. If you choose to > design in Altera Hardware Design Language (AHDL) you can define a parameter > early in the design file, and declare your counter and any other datapath > elements in terms of that parameter. This parameter can be passed down to > whatever level of hierarcy nesting you like. > > I know of no other HDLs that provide this facility - allowing that VHDL & > Verilog are synthesis languages rather than simple hardware description > languages. Please correct me if I'm wrong. > VHDL will let you do this by using generics. You can explicitly construct hardware just as you can in AHDL. The synthesis engines don't have the nice functions like LOG2 in them, so constructing something whose number of stages varies with the width like an adder tree is easier in AHDL (although if you add the IEEE.math_real library to the synthesizer that is fixed). The biggest drawback to AHDL is that you can't use it for another vendor's part. -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email randraka@ids.net http://users.ids.net/~randrakaArticle: 15095
Is there a common, single solution set for hw and sw to program FPGAs and serial ROMs and CPLDs, from a PC parallel port to a JTAG port? Xilinx has the Parallel Cable (with software modules that are proprietary to Xilinx)... Altera has the JAM language and Byte Blaster (and JAM is officially non-propietary, but is there JAM support for non-Altera devices?)... Altera's ByteBlaster is very similar to Xilinx' Parallel Cable, except the pinouts are different. Hence the SW to drive them is different. Is there a single, common solution for programming and/or testing A's and X's devices that are both JTAG supporters ? Thanx in advance, -- Bob Elkind, eteam@aracnet.comArticle: 15096
Hi all, About a year and a half ago I was considering implementing a PCI interface in an FPGA. At that time I decided that the design was going to be very time consuming and posed compatibility risks. Altera's and Xilinx's cores seemed to require quite a bit of tweaking and they were expensive. As I embark down this path again, I thought I might illicit the response of this newsgroup to the current state of PCI cores and development tools supplied by Altera, Xilinx, and others. Hopefully, this topic was not just recently discussed in a thread that I missed. If so, I would appreciate being directed to an archive. From my initial survey it appears that Xilinx has put the most effort into supplying PCI cores of various flavors that are supposedly easy to plunk into a design. Is this observation correct ? Does anyone who has used either or both of the Altera and Xilinx cores have any strong opinions? Generally, are how much tweaking with placement and timing constraints is required for the current crop of cores? Thanks in advance. Bob BaumanArticle: 15097
You can get the low voltage standards right from JEDEC at: http://www.jedec.org/download/freestd/jesd8-xx/ Rune Baeverrud wrote: > > Working with the new Xilinx Virtex family has made me curious about all the > different I/O standards that it support. The Virtex supports the following > standards: > > LVTTL > LVCMOS2 > PCI > GTL > GTL+ > HSTL Class I > HSTL Class III > HSTL Class IV > SSTL3 Class I and II > SSTL2 Class I and II > CTT > AGP > > The electrical properties of these standards are described in the Virtex > datasheet and XAPP133 "Using the Virtex SelectIO". > > What I would like to do now is gathering some information on the history and > typical use of these standards. Searching the Internet gave me very little > information on the newer standards, so I was hoping that you out there could > help me to locate this information. Newer standards not mentioned above are > also of interest. > > The result of this survey will be published at http://freecore.com > > Regards, > Rune Baeverrud -- ---------------------------------------------------------------- Web Page: file:////chamfs/share/applications/web/top.html ---------------------------------------------------------------- Brad Taylor Chameleon Systems Phone: 1-408-730-3300 ext 108 Fax: 1-408-730-3303 Email: <Brad Taylor> blt@cmln.com WWW: www.cmln.com Location: 1195 W. Fremont Ave Sunnyvale, CA 94087-3825 ----------------------------------------------------------------Article: 15098
On Tue, 02 Mar 1999 18:13:37 +0000, Eduardo Augusto Bezerra <E.A.Bezerra@sussex.ac.uk> wrote: > >Hello > >I want to use VHDL in order to have a technology independent >implementation, but the problem is that VHDL is not synthesis >tool independent. Is it correct? Is it possible to have only >one VHDL description which can be used for several synthesis >tools to infer memory to different FPGAs? The problem here is that simulation and synthesis tools work differently. A synthesiser doesn't attempt to "run" (ie. simulate) your code - it looks for templates for things that it knows about. With the common tools, both Spectrum and Synplify understand RAM templates, but Express doesn't. Express's code will probably work, but it will be inefficient, as you've found out. I wouldn't personally recommend coding for RAM, at least until everyone agrees on a common template. Todd's suggestion would be better. EvanArticle: 15099
Hi, How can I specify the PAL package in PALASM 1.5? I'm trying to compile for PALCE22V10, PLCC package. Best Regards, Wilton Padrao
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