Hi, I would like to know which is the smallest (read cheapest) board with a Spartan3 that can be programmed directly by a usb cable.
I just need a couple to test a design that must be operated remotely and in the need of a change in the fpga I would not bother the other party with much software/hardware issues.
Budget is very tight, there are no many requirements for the board itself, a bunch of I/O (less than 20) and I would need to add a serial LCD, a couple of quad encoders and some buttons.
Serial port and/or USB data exchange at plus.
Ideally, I use a Xylo-LM that has a very easy programming method, sadly way overbudget.
I have been using Opal Kelly 3050 boards. They can be configured via USB, have IP that lets you do I/O through USB and have a reasonable number of I/O pins. The 3050 use Spartan XC33S4000 parts. There are
3010, 3005 and 3001 that have smaller parts on them, but I have not used them. They are not cheap, the 3050's are about $750 US. The smaller boards go down to $175 US.
Their customer service is not stellar and their docs are kind of weak. Last time I contacted their customer service, I was told to RTFM until I told them that what I needed was not in the FM.
They have a cheaper version which is exactly like this but has a smaller FPGA. I don't have any experience with this board myself, but have a much older board from them (the XSA-100) which I like a lot.
Thanks to all for your answers, really good advices.
I need a very cheap board, so Papilio and/or Xula are best suited for the task (2 x 175$ is too much, I could buy another Xylo-LM).
I need to build a "high speed" HST IAMBIC keyer.
The device is very simple, 2 digital input for the paddles and some user interface to set the WPM speed, sensitivity and sidetone frequency. The output will be one digitally optocoupled and a little amplified loudspeaker. The device must be standalone in operation, without a computer so probably a serial 16x2 LCD will provide feedback.
It needs to be tested by a skilled "brass pounder", that does not know much of fpga and is 600km apart.
I don't want him to experience the wonderful parallel cable circus, every time he discover something to be changed, that's the reason for the USB programming.
I would like to have a twin system to remotely help him to update the bitfile, thus the need of a cheap board, since I have to buy two.
I just have the basic stuff working, still need to add the LCD and the encoders to let the user to set the WPM and so on and I doubt it will fit in the 50k part.
Xula is really low price, small and give me most of what I want (except maybe usb data transfer), and even DRAM.
Papilio, on the other side, has a serial port embedded into the USB but lacks the DRAM (that I don't need) and is much bigger (250k is really okay).
If Xula would be able to talk with a program using just the USB could be the best fit, otherwise Papilio.
This isn't an appropriate FPGA application, unless your keyer needs to work at 50,000,000 words per minute. Look into the Arduino family of microcontroller boards, maybe?
This seems like an ideal application for a microcontroller. NXP is selling some little ARM M0 demo boards for $30US and ST is selling some ARM M3 demo boards for $12US. The NXP uses GCC and Eclipse open source tool chains and the ST board uses an IAR demo version. Both have generous amounts of Flash memory for code and RAM. Both use USB for programming and debugging.
It might just be that an Arduino would be less expensive - maybe not - but this is a perfect project for someone who is into Ham Radio and wants to learn about FPGAs/CPLDS and/or HDLs.
The basic iambic keyer with dot and dash memories will fit in a Xilinx XC9536XL which is $1.18. I don't think you can build an Arduino for that price.
I actually used a Digilent Nexys2 board for my keyer. Then I added a PS/2 keyboard interface so I could type. Then I added a split-screen Color VGA display (128 x 64 characters). I use the display memory for receive and transmit buffers. Someday I'll add the signal processing for a tone decoder and then morse code decoder. Yes it is a little weird to do this with a 50 MHz clock but then again the clock is PLL'd to 108.33 MHz for the VGA pixel clock...
I use the Nexys2 board just because it is fast and easy (i.e., fun). Later I'll build a circuit card.
... it's better than a vending machine or a traffic light.
Yes, expensive compared to Xula and/or Papilio, I did know about them, nice product.
I was not able to locate one either, and I would like a reasonable support.
John, you are right, but I should have said that this is not a commercial product, just a hw shell for a "virtual keyer". Once the design would be proven ok, I could release under GPL / openhardware and everyone would implement it as they please.
I already have a K42 from
formatting link
that performs flawlessly at human speeds (up to 60-100 WPM) but this is done for different reasons:
1) HST competitions Look here:
formatting link
Probably your Arduino could have some problem with gals like these...
2) Learn Verilog and testbenches I have a very basic knowledge of both (and sometimes you could see my naif questions in this newsgroup)
3) Try to build a "absolute precision" keyer, sometimes a pic or arduino seems not perfect. I want the maximum precision available in terms of uniformity of response and timing accuracy
4) wet my feet for building a standaline CW decoder, that seems so rare animal this days.
If I wanted a fast MCU, I would have just used a Isopod that would have done a very fast keyer without hassles in minutes using IsoMax/Forth.
I wanted it to be a cheap FPGA project, so a MCU is not an option anyway. I have a LPC2138 and I got sick trying to put RTOS on it, my FPGA board does have it side by side of a Spartan 3E 500 and would be a very powerful combination. I regularly lose myself in the quirk of Eclipse/ZC something plugin for debugging, makefiles and so on.
Bingo! Maybe I didn't say it explicitly but this is the reason, plus the fact that a FPGA is way more "realtime" than a MCU. I would like to talk with you offlist, please contact me at giuseppe.marullo @ nospam.iname.com (remove nospam.)
... he started out asking for a Spartan3 board, which is a bit much if he wants to build a keyer and be done with it. He didn't mention that he wanted to learn about FPGA development, but he *did* mention that he had a tight budget, so... horses for courses.
Yep, Nexys2s are great boards. Good for climbing the USB learning curve as well as coming up to speed on FPGAs. Very well documented, and very inexpensive considering what you get...
ie3.html. If you really want to cut costs this board has an option to drive the JTAG chain as a bit bash from the on-board FT232 chip that also forms a serial port connection. As yet we don't have formal support for this feature but we do of a customer that has used it successfully in another of products with the same feature.
This cheap JTAG connection will need the fit of resistor array, or solder bridges, and the work of getting the bit bash to work. Otherwise we do have our own ISE compatible USB programming cable.
Price for this board with parallel port cable GBP =A340 / USD $68 + VAT (if applys) or with USB programming cable option GBP =A390 / USD $153.
We also have a another not widely known product Prog4 that has a FT245 front end but can not support the bit-bash JTAG at a similar price point. It's main use is as a programming cable for our Altera based products but we also sell it as a development board. It has slightly less I/O but is physically smaller. Like the Polmaddie3 it has a XC3S50AN FPGA. Details of Prog4
Stephen, very aggressive price but I don't want the other ham to play with parallel cables. I don't even own a USB programmer cable myself.
My Xylo-LM can be programmed by the FX2 USB chip, I did purchase the el-cheapo parallel cable (I have a old digilent board with Sump LA analyzer on it) and another Xilinx/Altera compatible one from KNJN but it is a PITA on today's computers, parallel port is simply dead.
I have somewhat high end laptop (x7200) and there is no pcmcia or parallel port.
Antoher (usual) problem is the esoteric I/O connector, otherwise it could be a good fit.
Jim, I have so much to learn with Xilinx I can't afford to switch vendor...at least I know how to create a bitfile and do a very basic simulation with WebISE.
He is a fine brass pounder with his Vibroplex but I am pretty sure not much tolerant with the quirks of bitfile programming webise/impact and cables. I fear that at the first problem he would ship the board back to me, so it is really hit/miss situation.
formatting link
Gabor, they should suffice:
2 input for the paddles
3 x 2 input for 3 quad encoders (maximum)
2 x CW output (one is the Keyout for the radio and the other is the Tone)
1 x Serial LCD output
This board is 89EUR from Trenz (that has a lot of interesting stuff, sadly not cheap for my needs (micromodules) or lacks USB programming (OHO Elektronik), and there is a discount if purchased during AVNet presentation seminars.
Yes, I have enrolled in one (end of November in Milan). Maybe I could buy two at discounted rate and my problems will be solved while I progress in the development.
It has a device-locked SDK license and ChipScope (much more that I could expect). The only thing I need to understand is if I could use it to try also the Microblaze (NOT FOR THIS PROJECT!).
It is not that hard to flip across vendor's tools ?
The two Lattice choices I gave have different tool flows, and I'd suggest the Breakout 4256ZE : less overall logic, but the IspLEVER tools give better reports, and run faster from my brief tests. IspLever is a little more cumbersome on the install.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.