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
Look for the "VELAB" beta release at Xilinx; try http://www.xilinx.com. It contains a number of parameterized VHDL entities, including SERPAR and PARSER, which do the kind of conversion you need, for the XC6200. Alan Reetinder P. S. Sidhu wrote: > I'm doing bit-serial computation using the XC6216 FPGA and require > parallel to serial and serial to parallel convertors. I would > appreciate any info on how to implement them on FPGAs > (esp. XC6200). Thanks in advance. > > Reetinder SidhuArticle: 7876
where can i find a design example (vhdl) for DCT (discrete cosinus transform). where can i find library or how can i get it <ieee.std_logic_unsigned.all>.Article: 7877
Hello everyone, I would be very pleased if someone could send me interesting links to FPGA design sites, information besides the vendors of course. Thanks in advance, RobbieArticle: 7878
Hi, About the delay - I'll take it into account. My problem was not the power supply - I tied the RD/WR signals from the CPU to the Flex8000 device - and although those pins were supposed to be tri-stated in the configuration process they were actully pulled to ground so that crashed my CPU cause it couldn't read the RAM/ROM data as it should. About CS/nCS - I tried CS to VCC and nCS is coming from my 386EX chip-select unit. About RDYnBUSY -- I am developing an embedded PCI card, not ISA. Even though I thought about making the RDYnBUSY a waitstate generator with a little glue logic and my 386EX READY# pin - I decided against it because the software is a simpler solution. About TTF - I included it as a array of bytes. What did you do with the RBF? Are you reading from some mass-storage device? I only have Flashes. -- --== Ido Kleinman ==-- kleinn@erez.cc.biu.ac.il Ilija Hadzic wrote in message <62j2dr$42$1@netnews.upenn.edu>... >Ido Kleinman (kleinn@erez.cc.biu.ac.il) wrote: >: 1. I create a high-low-high transition in the nCONFIG line >: 2. wait for nSTATUS to go high > >Make sure that the nCONFIG pulse is long enough (ca. 2 microseconds). >Also before you start writing put some delay. Altera says that the delay >should be about 5 microseconds and I have seen nSTATUS go up earlier >(BTW, in my design I don't even test it, I just wait after the nCONFIG >pulse and it works for me). > >: 3. I poll the RDYnBUSY line for ready state (high) >: 4. I do a I/O write cycle to the FLEX address and write the next byte from >: the array. >: 5. I check if nSTATUS gone low = frame error. >: 6. repeat steps 2-5 for the exact number of bytes needed to be loaded >: (sizeof (arraybytes)) >: 7. wait for conf_done. > >This scheme works fine for me. BTW, what do you do with CS and nCS >signals. I keep them high and low respecitvely (i.e. active) during the >entire configuration. > >Also instead of polling RDYnBSY signal in software, you can use it to drive >the IOCHRDY signal on the ISA bus and let the bus controller add wait >states as necessary. > >Also make sure that you are reading the file properly. I prefer using RBF >instead of TTF and read it as the binary file. > >IlijaArticle: 7879
feydo wrote: > > Hi All, > I having a problem get Workview Office to do what I want. I' m trying > to make an adder tree which will add together five 16-bit numbers. I' ve > used Xilinx's Core Generator program to make the adders. The first adder > adds A[15:0] and B[15:0] resulting in S1[16:0]. The second adder adds > C[15:0] and D[15:0] resulting in S2[16:0]. The third adder adds S1[16:0] > and S2[16:0] resulting in S3[17:0]. Now comes the problem ... I want to > add S3[17:0] and E[15:0]. The adder generated by CoreGen expects two 18 > bit inputs, but the one input is only 16 bits wide. Obviously the upper > two bits should be zero to make it 18 bits wide. Does anyone know how to > make WVOffice do this? I have tried simply connecting the 16 wide bus E to > the 18 wide input, but WVO lines up the 16 wide bus to the upper 16 bits of > the 18 wide input, leaving the lower two bits zero (or unconnected?). Any > suggestions will be greatly appreciated. > > Thanks, > mark You need to make the bus have more than one segment. The bus segment that attaches to the 16 bit adder output must have a label on it that matches the width of the adder output (in your case bus[15:0]). The segment that connects to the next adder's input must also be labelled with an appropriate label (in this case bus[17:0]). The two extra bits need to be driven by something that explicitly calls out the bits. This can either be a macro with a bus output or individual signals tied to the bus. Either case must have appropriate labels. You may need to sign extend to make your adder tree work right, in which case the extra bits come from a sign extension macro consisting of a one bit input driving an individual BUF for each bit you wish to sign extend. In your case you'd have a macro with two BUFs inside. The inputs would be a single signal (call it A internally) and a bus output with two outputs, Y0 and Y1. The input to the macro gets connected to the bus from the adder and must be labelled bus15. THe output gets a bus connection to the same bus between the adders and the segment connecting this macro gets labelled bus[17:16]. -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: 7880
ATMEL's new 40K family are pin for pin compatible with XILINX FPGAs. They currently have the 40K20 chip available but will soon have others available. They are also working on getting some very reasonably priced tools out for these. Their 6K family of FPGAs never quite caught on, but the potential for the these new FPGAs is high. With the XILINX pin compatibility coupled with the architecture features such as: * diagonal routing lines * a less course CLB structure where all flip flops can be accessed without using LUT resources * Internally distributed RAM with the RAM not using LUT resources. * Claims of a faster Multiplier structure (i haven't quite figured this out yet) * LOWER COST SILICON We are contemplating doing some low cost ATMEL kits including tech notes and development boards. ATMEL has kindly dropped off a copy of the software, and I will probably be using EXEMPLAR as the synthesis front end for now. I have heard rumors to the effect that there will be some low cost front end VHDL tools offered with the router at very reasonable prices. Stay tuned to our website and our newsletter for future developments along these lines. http://www.associatedpro.com/aps --Wade D. Peterson wrote: > Does anybody know if anybody makes pin compatible replacement parts > for XILINX XC4005, XC4006 or XC4008 devices? > > Thanks a bunch. > > Wade > > ----------------------------------------------------- > Wade D. Peterson | TEL: 612.722.3815 > Consultant to Industry | FAX: 612.722.5841 > 3525 E. 27th St. No. 301 |---------- EMAIL ---------- > Minneapolis, MN 55406 | peter299@maroon.tc.umn.edu > ---------------- Committed to Quality --------------- -- __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ Richard Schwarz, President EDA & Engineering Tools Associated Professional Systems (APS) http://www.associatedpro.com 3003 Latrobe Court richard@associatedpro.com Abingdon, Maryland 21009 Phone: 410.569.5897 Fax:410.661.2760 __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/ __/Article: 7881
Does anybody have a good idea, how to design an All Digital DLL or PLL on a 0.8 CMOS Process. The PLL or DLL should have a hold range from 30kHz to 50kHz, and veri low jitter (less than 20ps !). If you have some good ideas or you know a goog site concerning this problem. Send me a mail please ------------------------------------------------------------------ Michael Ammann +41 41 741 17 91 Keltenweg 20 http://www.stud.ee.ethz.ch/~mammann/ CH-6312 Steinhausen mammann@stud.ee.ethz.ch ------------------------------------------------------------------Article: 7882
Jacob W Janovetz wrote: > > Hello... > > I have a 'should it be done' question. We have a PC/104 embedded > system (ISA bus) and want to put a Xilinx part on a board with > the PC/104 connector. I'm wondering if it is ok to place the > FPGA directly on the bus (ie without the traditional 74245 > bidirectional buffers). When the Xilinx is programming or is > unprogrammed, the pins are in a weak pullup state which should be > ok for the ISA bus. > > I've tried this in a prototype situation and everything is > working fine. I can reprogram the Xilinx in-circuit and it doesn't > screw up the PC/104 system. I'm just wondering if this is a > common practice and if it will be reliable between PC/104 systems > and such. > > As an unrelated question, during an I/O write, data is latched > (by the FPGA) on the rising edge of IOW*. Does this mean a signal > like this (IOW*) should be on a clock pin of the FPGA to allow a > BUFG to be attached to it? > > Cheers, > Jake > > -- > janovetz@uiuc.edu | Once you have flown, you will walk the earth with > University of Illinois | your eyes turned skyward, for there you have been, > | there you long to return. -- da Vinci > PP-ASEL | http://www.cen.uiuc.edu/~janovetz/index.html Shouldn't cause any problems in the unprogrammed or programming states, the 4k's are used as PCI controllers so should be fine as ISA controllers. Ideally you should use a clock pin and a BUFG to avoid skew problems in the xilinx. In our RC1000-II ISA board I used a registered transceiver in the data path between ISA and the 4010. The 4010 is programmed using our C-based programming language Handel-C and the whole fpga is clocked by a single clock which is asynchronous to the ISA bus clock. So the registered transceiver decouples the ISA bus from the program running in the 4010 so there are no synchronization problems. We can licence the RC1000-II design if you're interested. Please take a look at our web site for info on the RC1000-II and Handel-C. Charles Charles Sweeney, Engineering Director, Embedded Solutions Ltd Tel/fax +44 1235 510456 <http://www.embedded-solutions.ltd.uk/> Email CharlesSweeney@compuserve.com or csweeney@embedded-solutions.ltd.uk 6 Main Road, East Hagbourne, Didcot, Oxfordshire. OX11 9LJ. UK.Article: 7883
Hey, because ALTERA isn't interested in our problem I try to find help here. We are trying to configure the following three ALTERA-device-chains via Bitblaster & MAX+plusII 7.2: 1) 10k20 -> 740 -> 10k10 2) 10k10 -> 10k20 -> 740 While configuring we remarked the following: - The I/O Pins are driven low and high for a short time - We tried to configure different combination with different success as you can see in the table below. This means that configuration is successfull but the devices are not initialized. + means: with programming file or conf. successful and initialization - means: no file or configuration successful but no intialization 10k20 740 10k10 initialization - + - + - - + - + - - + + + - + - + + - + - + - + + + - 10k10 10k20 740 initialization - - + + - + - - - + + + + - - + + - + + + + - + + + + + In this chain we have also following phenomena: When configuring: - + - it isn't working. When directly after that configuring + - + then all three devices are initialized correctly. We connected the pins as recommended in ALTERA specification for configuring 10k devices. Has anyone an idea how to solve the problem? Erwin ******************************************************* Erwin Ruoff Institut fuer Experimentelle Kernphysik University of Karlsruhe Postfach 3640 76021 Karlsruhe Germany Phone: +49 7247 82 4254 FAX: +49 7247 82 3414 mailto:Erwin.Ruoff@phys.uni-karlsruhe.de ******************************************************Article: 7884
>Hello everyone, > >I would be very pleased if someone could send me interesting links to FPGA >design sites, information besides the vendors of course. Thanks in advance, Two places to stard: 1) My 'work links' page at http://members.aol/com/bdipert/wrklinks.htm 2) Steven Knapp's Programmable Logic Jump Station site (which is excellent) at http://www.optimagic.com/ Regards, Brian Dipert Technical Editor EDN Magazine: The Design Magazine Of The Electronics Industry 1864 52nd Street Sacramento, CA 95819 (916) 454-5242 (916) 454-5101 (fax) edndipert@worldnet.att.net Visit me at <http://members.aol.com/bdipert>Article: 7885
Hello, I wanted to translate my old design ( XC5202 ) to Altera Flex8282A. When I completed and compile it, it can't fit into Flex8282ATC100. Later, I find that the "Internal tri-state emulation" of the Altera use up many resource (over 40 LE's). Is there any function(s) of the compiler to solve this problem ? Xilinx's internal tri-state do not consume any logic cell! Roger Yau Easson Precision Ltd.Article: 7886
In article <01bce150$55136920$3f15440c@lincoln-labs>, "John McGibbon" <john_mcgibbon@memecdesign.com.stop.spams.please> writes: >Memec Design Services has a module for Xilinx FPGA's. It has been used and >is available. I'm not sure of cost and features but a datasheet is >available on their website: > >http://www.memecdesign.com/xf-twsi.htm > >Austin Franklin <darkroo4m@ix.netcom.com> wrote in article >> Is there anyone who has done an I2C controller in an FPGA? >> > > > Mentor Graphics offers an I2C core as well. You can view the datasheet at: http://www.mentorg.com/inventra/softcore/catalog/mi2c_pf.pdf -- ...Mike No Guts, No Glory. +--------------------------------------------------------------+ | Michael P. Walsh mike@rtp-nc.mentorg.com | | mike_walsh@mentorg.com | | Technical Account Manager - Ericsson | | Mentor Graphics Corporation V:(919) 484-2505 | | 2525 Meridian Parkway, Suite 260 F:(919) 544-0701 | | Research Triangle Park, North Carolina 27713 | +--------------------------------------------------------------+Article: 7887
In article <34544C3D.6771@stud.ee.ethz.ch> Ammann Michael <mammann@stud.ee.ethz.ch> writes: > > Does anybody have a good idea, how to design an All > Digital DLL or PLL on a 0.8 CMOS Process. The PLL or > DLL should have a hold range from 30kHz to 50kHz, > and veri low jitter (less than 20ps !). If you have > some good ideas or you know a goog site concerning this > problem. > The man who literally WROTE THE BOOK on Phase Locked Loops (_Phaselock_Techniques_, published by John Wiley and Sons) is a consulting engineer; perhaps you might like to hire him to do some of the work. His name is Floyd M. Gardner and his consultancy business has an advertisement in the classified ads of _IEEE_Spectrum_ magazine. He appears to be based in Palo Alto, California, USA. -Mark JohnsonArticle: 7888
Roger Yau <rogeryau@net.polyu.edu.hk> wrote: >Hello, > >I wanted to translate my old design ( XC5202 ) to Altera Flex8282A. >When I completed and compile it, it can't fit into Flex8282ATC100. >Later, I find that the "Internal tri-state emulation" of the Altera >use up many resource (over 40 LE's). Is there any function(s) of the >compiler to solve this problem ? Xilinx's internal tri-state do not >consume any logic cell! > >Roger Yau >Easson Precision Ltd. I generally use muxes. The Altera parts don't have tri-state capability except at the I/O pins. Regards, Aaron Quantz \^ ^/ )@ @( +---------------------------oOO--(_)------------------------------------+ + Mgr Software Development, Turret Control Systems + + HR Textron | Phone: (805) 253-5471 + + 25200 W. Rye Canyon Rd. | Fax: (805) 253-5962 + + Valencia, CA USA 91355-1265 | Email: aquantz@ibm.net + + Visit the Textron web site: http://www.textron.com + +-----------------------------------Oooo--oOO---------------------------+ oooO ( ) ( ) ) / \ ( (_/ \_)Article: 7889
Hi Robbie, I'll give my own company a plug. Go to... www.memecdesign.com We are Xilinx design experts, so we are naturally biased to Xilinx, but don't let that scare you. It is always best to use whatever vendor and tool best serves your needs. take care, tom On 26 Oct 1997 19:28:50 GMT, "Thielemans - Van Heghe" <robbren@skynet.be> wrote: >Hello everyone, > >I would be very pleased if someone could send me interesting links to FPGA >design sites, information besides the vendors of course. Thanks in advance, > >RobbieArticle: 7890
Does anybody know if anybody makes pin compatible replacement parts for XILINX XC4005, XC4006 or XC4008 devices? Thanks a bunch. Wade ----------------------------------------------------- Wade D. Peterson | TEL: 612.722.3815 Consultant to Industry | FAX: 612.722.5841 3525 E. 27th St. No. 301 |---------- EMAIL ---------- Minneapolis, MN 55406 | peter299@maroon.tc.umn.edu ---------------- Committed to Quality ---------------Article: 7891
Hello! I'm implementing several counters in a Lattice ispLSI1016. For larger counters in a syncronous design, you need lots of ProductTerms. -> Split in many MacroCells. Grrrr. With an asyncronous counter on the other hand, you need 2 Macrocells (FlipFlop + Carry) per bit. Aaaaargh. What could I do to save resources??? Thanks ArminArticle: 7892
Robert E. Engle Jr. wrote: > > i have a sizable library of generic modules i have written for the > synario system, that works with the versin lattice gives away. The explanation goes without saying, I suppose: The .SYM and .ABL files of the desired function or functions must be included as part of the project. That means not only loading the symbol into your schematic, but importing the .ABL associated with it as a source module in your ISP Synario project. Otherwise you get that little "?" thing in the sources window. -TBBArticle: 7893
You can create an internediate carry signal that you then feed to all the higher-order bits, and thus you still have a fully synchronous counter. Or you build a synchronous counter of maximum efficient length, then use its output as the clock for a new more significant counter. It all depends how highly you value synchronous operation, and whether you are loadable or even up/down. Peter Alfke, Xilinx.Article: 7894
Xilinx makes the pin-compatible functional supersets XC4000E ( 5 V) and XC4000XL (3.3 V ) and the XC5200 family ( 5 V ) which has compatible pin-outs, in case you are allergic to the original XC4000 ( why?). Nobody makes pin- and functional compatible substitutes, and it is unlikely that there will be much second-sourcing in the future. At most, there might be devices that happen to have their supply pins in the same location, whatever significance that has. Peter Alfke, Xilinx ApplicationsArticle: 7895
The answer depends a lot on your goal. If you want devices that are directly pin-compatible and function-compatible, then you probably want to look at the following devices: XC4005 --> XC4005E --> XC4005XL (3.3 volt with 5 volt tolerant I/O) XC4006 --> XC4006E XC4008 --> XC4008E The Xilinx XC5200 family also has devices that are pin compatible to the XC4000 family but with a different internal architecture. Most designs are easily transferred, except for those that contain RAM. Supposedly, the new Atmel's new AT40K family is pin compatible with the XC4000 family, but also with a different internal architecture. ----------------------------------------------------------------------- Steven K. Knapp 'Great designs happen "OptiMagic"-ally' OptiMagic, Inc. E-mail: sknapp@optimagic.com Web: http://www.optimagic.com Phone: 1-408-454-1811 Wade D. Peterson wrote in message <632ja0$4g3@epx.cis.umn.edu>... |Does anybody know if anybody makes pin compatible replacement parts |for XILINX XC4005, XC4006 or XC4008 devices? | |Thanks a bunch.Article: 7896
I am looking at a host of DSP applications (some high fidelity audio, some telephony quality ... multiple channels), and am interested in implementing them in FPGAs. I am interested understanding from designers who have implemented their own DSP functions about the challenges and pitfalls of implementing DSP functions in FPGAs. (I am also interested in the relative merits of the 'core/megafunction' offerings from Xilinx and Altera). -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to UsenetArticle: 7897
Please see the paper: J. Dunning et al., ``An All-Digital Phase-Locked Loop with 50-Cycle Lock Time Suitable for High-Performance Microprocessor'' in IEEE J. of Solid-State Circuits, April 1995, pp. 412-422. Best regards, yueqiang On Mon, 27 Oct 1997, Ammann Michael wrote: > Does anybody have a good idea, how to design an All Digital DLL or > PLL on a 0.8 CMOS Process. The PLL or DLL should have a hold range > from 30kHz to 50kHz, and veri low jitter (less than 20ps !). > If you have some good ideas or you know a goog site concerning this > problem. > > Send me a mail please > > ------------------------------------------------------------------ > Michael Ammann +41 41 741 17 91 > Keltenweg 20 http://www.stud.ee.ethz.ch/~mammann/ > CH-6312 Steinhausen mammann@stud.ee.ethz.ch > ------------------------------------------------------------------ > > > -- Best regards, yueqiangArticle: 7898
Hi all, We've hit a snag here at work, trying to program the Altera EPC1 configuration EPROM using the Chipmaster 6000 from Logical Devices Inc. Logical Devices says they have verified the 6000 with the EPC1, but we can't seem to program any with our unit (it always fails on verify pass). Has anyone else experienced problems with the EPC1 part? We've heard rumours (from another EPROM programmer maker) that Altera has somehow changed the programming requirements of the EPC1 *without* changing the part number! If so, all I can say is "$%@#!$!". We would have bought Altera's own programming unit, but the delivery lead time was many weeks. Sigh.. Any advice appreciated, -Russ MageeArticle: 7899
Teilnehmer Informatik I wrote: > > Hello! > > I'm implementing several counters in a Lattice ispLSI1016. > For larger counters in a syncronous design, you need lots of ProductTerms. > -> Split in many MacroCells. Grrrr. > With an asyncronous counter on the other hand, you need 2 Macrocells > (FlipFlop + Carry) per bit. Aaaaargh. > What could I do to save resources??? > > Thanks > Armin I don't have the Lattice data handy, but the secret to large counters is either TOGGLE flip flops. ( can build Up.Dn.Load ) or D FF, with a CLOCK Enable product term If you have neither, then you are stuck with growing width terms... - jg -- ======= Manufacturers of Serious Design Tools for uC and PLD ========= = Optimising Modula-2 Structured Text compilers for ALL 80X51 variants = Reusable object modules, for i2c, SPI and SPL bus interfaces = Safe, Readable & Fast code - Step up from Assembler and C = Emulators / Programmers for ATMEL 89C1051, 2051, 89C51 89S8252 89C55 = *NEW* Bondout ICE for 89C51/89C52/89C55 = for more info, Email : DesignTools@xtra.co.nz Subject : c51Tools
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