Author Topic: Spartan 3E board  (Read 2922 times)

0 Members and 1 Guest are viewing this topic.

Offline avogadro

  • Contributor
  • Posts: 32
  • Country: hr
Spartan 3E board
« on: November 21, 2018, 07:32:03 pm »
Is 60 euro a good price for used Spartan 3E board? Are there any better newer boards for better price? Any recommendations?
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 6700
  • Country: us
    • Personal site
Re: Spartan 3E board
« Reply #1 on: November 21, 2018, 07:59:31 pm »
That entirely depends on what kind of a board it is and what you want to do with it.
Alex
 

Online NorthGuy

  • Super Contributor
  • ***
  • Posts: 2120
  • Country: ca
Re: Spartan 3E board
« Reply #2 on: November 21, 2018, 08:57:58 pm »
Is 60 euro a good price for used Spartan 3E board? Are there any better newer boards for better price? Any recommendations?

You can get Spartan-7 board for less, so I wouldn't go with Spartan-3E. Of course, what is on the board is also important and will dramatically affect the price.
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 7526
  • Country: us
Re: Spartan 3E board
« Reply #3 on: November 21, 2018, 09:58:37 pm »
A link to the board would help.  Some 3E Starter Boards (Digilent) have 7 segment displays, switches, LEDs and buttons.  In my view, these add a LOT of value.  At least one large IO header is a nice feature.

The second important item is WHICH Spartan 3E chip is being used.  Of course the less capable chips are going to cost less.

This board has a lot of 'gadgets' and ABSOLUTELY I would buy it for 60 Euros considering what it cost some years back:

https://store.digilentinc.com/nexys-2-spartan-3e-fpga-trainer-board-retired-see-nexys-4-ddr/

Today, the Artix 7 is the flavor of the month and the Arty-7 is fairly popular and more capable than the Nexys 2

https://store.digilentinc.com/arty-a7-artix-7-fpga-development-board-for-makers-and-hobbyists/

But it doesn't have any gadgets and while I have a couple, they aren't my favorite boards.  Instead I prefer the Nexys 4 DDR because it has a ton of gadgets:

https://store.digilentinc.com/nexys-a7-fpga-trainer-board-recommended-for-ece-curriculum/

Boards are priced all over the place and, in general, I would buy most any board priced at around 60 Euros if it had a decent number of gadgets.  But it's the gadgets I want.

Think about what the board has and what you might need.  I find it easier to debug projects if I have gadgets.
 

Offline legacy

  • Super Contributor
  • ***
  • !
  • Posts: 4415
  • Country: ch
Re: Spartan 3E board
« Reply #4 on: November 21, 2018, 10:25:35 pm »
Is 60 euro a good price for used Spartan 3E board? Are there any better newer boards for better price? Any recommendations?

it depends ... which board?

The Digilent Xilinx S3-500 and S3-1500 are GOOD boards, extremely GOOD, useful, well manufactured, with a lot of gadgets, and well documented.

There is a dude in the UK that has several for sale. Second hand, of course.

I bought my first board brand new several years ago, I have just got a second board from him. This time a 1500 board, since it offers more LE.
 

Offline legacy

  • Super Contributor
  • ***
  • !
  • Posts: 4415
  • Country: ch
Re: Spartan 3E board
« Reply #5 on: November 21, 2018, 10:27:39 pm »
3E Starter Boards (Digilent)

There is also a mini-3E version, it comes with a S3-200, small factor, no gadgets (except a few LEDs), but the price is something like 10 euro/board.

It seems nobody wants them  :-DD
 
The following users thanked this post: avogadro

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 7526
  • Country: us
Re: Spartan 3E board
« Reply #6 on: November 21, 2018, 10:44:22 pm »
3E Starter Boards (Digilent)

There is also a mini-3E version, it comes with a S3-200, small factor, no gadgets (except a few LEDs), but the price is something like 10 euro/board.

It seems nobody wants them  :-DD

I went with S3-1000 boards back when I was playing with Digilent's Sparten 3 Starter Board (prior to 3E boards).  It had 3 50-pin headers and was very friendly in terms of plugging in attachments.  None of the PMod stuff!

More LEs is always better.  More BlockRAM is even better!

It's also important to try to decode what a logic element is.  Sometimes it is a 4 input MUX with flop, sometimes it is a 6 input MUX with flop, sometimes it includes a carry-chain and, in general, different vendor specifications aren't directly comparable.  Marketing gets into the picture and new buzz words flow!
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 12335
  • Country: us
Re: Spartan 3E board
« Reply #7 on: November 21, 2018, 11:55:21 pm »
Yes block RAM, for my purposes my projects are almost always bound by block RAM, I would love it if FPGAs had ~4x as much block RAM for a given amount of logic resources as they do.

The Spartan 3E is an older but still capable part, whether the board is a good deal mostly depends on what else is on it.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 5495
  • Country: fr
Re: Spartan 3E board
« Reply #8 on: November 22, 2018, 01:26:16 am »
The Numato Mimas V2 at around $50 with a Spartan V6 is most likely a way better deal, but I don't know what's on your board except a Spartan 3E.

Heck, if you don't need anything more than I/Os, the Numato Mimas is like $35.
 

Offline KE5FX

  • Super Contributor
  • ***
  • Posts: 1243
  • Country: us
    • KE5FX.COM
Re: Spartan 3E board
« Reply #9 on: November 22, 2018, 04:22:34 am »
Don't mess with Spartan3E, it's obsolete along with the software needed to develop for it.  Go for an Artix board.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 12335
  • Country: us
Re: Spartan 3E board
« Reply #10 on: November 22, 2018, 04:30:40 am »
That depends on what you're trying to do with it. I looked into them a while back and they are more than twice the price of the parts I've been using and the additional features offered by the Artix are utterly useless to me, I'd be just as well off tossing a $20 bill in the fireplace for each board I buy. As an example, Galaxian the arcade game uses ~29% of a Spartan3 500, so what benefit would there be to using a much larger more expensive part? It seems some people cant get past the fact that not everyone is doing FPGA development for a living, designing high end products using the latest technology. I would go so far as to say the vast majority of people asking about FPGAs on here are doing nothing of the sort.

It all comes down to what you want to do with an FPGA, and as I previously stated, the other parts on the dev board are generally more important than the FPGA itself. You can get a minimal dev board with little more than the FPGA and config ROM for under $20 shipped, but if you need a bunch of peripherals then the cost starts to add up.



 

Offline KE5FX

  • Super Contributor
  • ***
  • Posts: 1243
  • Country: us
    • KE5FX.COM
Re: Spartan 3E board
« Reply #11 on: November 22, 2018, 05:05:10 am »
That depends on what you're trying to do with it. I looked into them a while back and they are more than twice the price of the parts I've been using and the additional features offered by the Artix are utterly useless to me, I'd be just as well off tossing a $20 bill in the fireplace for each board I buy. As an example, Galaxian the arcade game uses ~29% of a Spartan3 500, so what benefit would there be to using a much larger more expensive part? It seems some people cant get past the fact that not everyone is doing FPGA development for a living, designing high end products using the latest technology. I would go so far as to say the vast majority of people asking about FPGAs on here are doing nothing of the sort.

It all comes down to what you want to do with an FPGA, and as I previously stated, the other parts on the dev board are generally more important than the FPGA itself. You can get a minimal dev board with little more than the FPGA and config ROM for under $20 shipped, but if you need a bunch of peripherals then the cost starts to add up.

This amounts to being stingy with a small amount of money and wasteful with a large amount of time.  The hardware isn't what's important in this business, because the software is what you spend your time interacting with.  ISE is an ancient pile of junk compared to current-generation Vivado.  For a novice, there's just no sense in messing with the former.

If you haven't used both parts, you really have no idea how much nicer the 7-series parts are to work with than Spartan3.  I have, and all I can think of is how much I wish someone had given me the same advice (along with the time machine required to take advantage of it).
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 12335
  • Country: us
Re: Spartan 3E board
« Reply #12 on: November 22, 2018, 05:23:13 am »
Nicer in what sense? If the older parts are overkill for what I'm doing, precisely what advantage is there in changing? How is using a part that is massively overkill better than using one that is only twice the size it needs to be? I'm at a loss as to how it will save me time to re-learn something I already know how to do, that requires parts that cost more than twice as much, when the parts I'm already using are already much more powerful than I need?

I'd really like to see some data here, let's use the Galaxian example. I could easily port it to an Artix board, I could afford to pay twice as much for the boards if there was a reason to do so, but using this example, in what way will the Artix board improve Galaxian? Will it compile faster than the ~30 seconds it takes for the Spartan 3? Will it save the world? What's the advantage that you are so passionate about? Keeping in mind that retro projects like this are the *only* thing I have any interest in doing with FPGAs. It's not being stingy, it's being sensible by using a part appropriately matched to the project. If a $14 board has twice the resources I need, it is irrational to argue that a $50 board with 20 times the resources amounts to a savings. Personally it just sounds like elitism to me, the same sort of attitude of people who look down their nose at anyone who buys anything less than top of the line gear regardless of their needs.

 

Offline KE5FX

  • Super Contributor
  • ***
  • Posts: 1243
  • Country: us
    • KE5FX.COM
Re: Spartan 3E board
« Reply #13 on: November 22, 2018, 07:29:34 am »
Nicer in what sense? If the older parts are overkill for what I'm doing, precisely what advantage is there in changing? How is using a part that is massively overkill better than using one that is only twice the size it needs to be? I'm at a loss as to how it will save me time to re-learn something I already know how to do, that requires parts that cost more than twice as much, when the parts I'm already using are already much more powerful than I need?

I'd really like to see some data here, let's use the Galaxian example. I could easily port it to an Artix board, I could afford to pay twice as much for the boards if there was a reason to do so, but using this example, in what way will the Artix board improve Galaxian? Will it compile faster than the ~30 seconds it takes for the Spartan 3? Will it save the world? What's the advantage that you are so passionate about? Keeping in mind that retro projects like this are the *only* thing I have any interest in doing with FPGAs. It's not being stingy, it's being sensible by using a part appropriately matched to the project. If a $14 board has twice the resources I need, it is irrational to argue that a $50 board with 20 times the resources amounts to a savings. Personally it just sounds like elitism to me, the same sort of attitude of people who look down their nose at anyone who buys anything less than top of the line gear regardless of their needs.

The question was, "Is 60 Euros a good price for a used Spartan3E board?"  The answer is, "No."

You may have other questions, but I was only attempting to answer the one that was actually asked.   

The second one, "Are there better, newer boards for a better price?" can be answered with "Yes and no."  About 90 Euros will buy an Arty or similar board.  The extra 30 Euros is well worth spending simply because the software is better, and time spent learning it is more likely to pay dividends later.

If the poster were looking to go into production, they wouldn't be asking about a single used Spartan3E board, would they?  The assumption is that they are looking to get started with FPGA work and need a trainer/dev board of some kind.  For that purpose, it's not wise to start with an obsolete/NRND part that uses software that is no longer officially supported.
 

Offline legacy

  • Super Contributor
  • ***
  • !
  • Posts: 4415
  • Country: ch
Re: Spartan 3E board
« Reply #14 on: November 22, 2018, 01:17:59 pm »
LOL, in a university laboratory, they are still using boards based on XC3000, which is very very limited if compared to modern chip.

I LOVE when the not-being allowed to use Logic/Arithmetic-SHIFTers and Barrel-SHIFTers is a mandatory constraint, so they have to implement everything while with Sp3 and Sp6 it would be just one line of VHDL  :D

But they also have XC4000 ... which I speculate are still in use because the XC4000XL I/O structures have been designed to tolerate a constant input voltage of 5.5V using only the single 3.3V power supply connected to the device. The XC4000XL is able to directly drive 5V Inputs with TTL Thresholds, while the XC4000XL outputs drive rail to rail and it completely meets all TTL Standard Specifications.
 

Online asmi

  • Super Contributor
  • ***
  • Posts: 1188
  • Country: ca
Re: Spartan 3E board
« Reply #15 on: November 22, 2018, 02:40:49 pm »
Yes block RAM, for my purposes my projects are almost always bound by block RAM, I would love it if FPGAs had ~4x as much block RAM for a given amount of logic resources as they do.
Then you definitely need to get 7 series FPGA as they have A LOT of BRAM compared to previous generations. And the ratio of BRAM-to-CLB is increasing as you go from S7->A7->K7->V7. And it's FAST (7 series BRAM's Fmax is more than double of that of S6).

Online asmi

  • Super Contributor
  • ***
  • Posts: 1188
  • Country: ca
Re: Spartan 3E board
« Reply #16 on: November 22, 2018, 02:50:10 pm »
Nicer in what sense? If the older parts are overkill for what I'm doing, precisely what advantage is there in changing? How is using a part that is massively overkill better than using one that is only twice the size it needs to be? I'm at a loss as to how it will save me time to re-learn something I already know how to do, that requires parts that cost more than twice as much, when the parts I'm already using are already much more powerful than I need?
I think he's talking about novice's perspective, and I'm 100% agree with him - usability-wise Vivado is eons ahead of ISE. For example, in Vivado I can design a system soft-core CPU, DDR3 memory controller and any amount of any standard peripherals I desire without even writing a single line of HDL code - this is a godsend when you're bringing up your board for the first time as it allows you to quickly verify that it works - and if it doesn't you can be 99% sure it's board issue, not that of a bitstream nor firmware for the CPU. Also few versions back they've added ability to directly instantiate your own RTL modules in the system diagram without creating and packaging it as IP (which was pain in the behind!), which makes integrating your own components a breeze - and self-documenting way, as diagram itself is a very good documentation/
Bottom line - for today's beginners - unless they need to be able to work with pre-7 series devices I would strongly recommend to learn Vivado and forget ISE ever existed.

Offline legacy

  • Super Contributor
  • ***
  • !
  • Posts: 4415
  • Country: ch
Re: Spartan 3E board
« Reply #17 on: November 22, 2018, 03:13:07 pm »
I would strongly recommend to learn Vivado and forget ISE ever existed.

I do find this comment very irritating.
 

Online asmi

  • Super Contributor
  • ***
  • Posts: 1188
  • Country: ca
Re: Spartan 3E board
« Reply #18 on: November 22, 2018, 03:23:06 pm »
I do find this comment very irritating.
Why is that?

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 7526
  • Country: us
Re: Spartan 3E board
« Reply #19 on: November 22, 2018, 03:54:58 pm »
I have a couple of Spartan 2 boards and a few Spartan 3(E) boards.  My CPU project runs on a Spartan 3E board and, while I have thought about porting it to an Artix 7 board, I just can't build up any interest in such a project.  What I have works and doing something twice is 'production' and I don't do production.  So, I will likely always have to use ISE to alter that projects.  I haven't done much with the project for the last 11 years except port from S3 to S3E and alter the IO ports and even that was a long time back.

I also have a couple of Arty boards and a couple of the Nexys 4 DDR boards and these require Vivado.  I didn't like the startup curve for Vivado and I still don't truly understand the constraints file but after I built a real project (LC3 RISC processor), I got a little way up the path.  Some day I will admit to actually preferring Vivado.  It's a nice tool.  There's a thing about putting aside a well understood tool and taking up a new tool that actually increases the difficulty.

It's a toss-up whether I prefer ISE which I have used for the better part of 15 years or Vivado which I have used for just a couple of projects.  I think the nod would go to ISE because I am more comfortable.  But that's because of history.  Looking forward, Vivado is the way to go.  ISE is a dead issue, Xilinx is no longer upgrading it, what it is is all it will ever be.  I certainly wouldn't point a newcomer at ISE UNLESS the board in question was sufficiently loaded with gadgets and attractively priced.  Otherwise, it's Artix-7 and Vivado.

The ARTY-7 board is modern, uses Vivado but doesn't come with a lot of gadgets.  It makes up for that with Arduino compatible headers and a lot of DDR3L ram (I prefer the static ram on the Sparten 3E boards).  I thought it used to cost $99 but now I see it's $119.  Maybe there is something else in that price range.  OTOH, it uses the 35T chip and I don't know what to make of that.  I would always prefer the 100T chip but I haven't really tried to stuff a large project in the 35T.  If it fits, that's all that matters.

Like it or not, moving forward it is going to be Artix 7 and Vivado, especially for newcomers to the field.  Unless, of course, their university is stuck back in time.  That happens...
 

Online asmi

  • Super Contributor
  • ***
  • Posts: 1188
  • Country: ca
Re: Spartan 3E board
« Reply #20 on: November 22, 2018, 05:02:55 pm »
Another issue with using old devboards is that peripheral chips tend to become obsolete, which will become an obstacle should you ever decide to take your project on a dedicated board - and that is certainly my goal as I always design projects so that they will eventually be implemented on dedicated boards. Well-designed devboards are a good tool to aid initial R&D and prototyping, but at some point development will have to move on a dedicated board so that system-wide integration testing can happen as there is more into the design process than just FPGA bitstream and MCU/CPU software.

Offline legacy

  • Super Contributor
  • ***
  • !
  • Posts: 4415
  • Country: ch
Re: Spartan 3E board
« Reply #21 on: November 22, 2018, 05:59:54 pm »
The Nintendo Gameboy ADV's dev board is still based on a Spartan2, and it's a neat and funny idea if someone wants to learn how to develop a video game on a real portable console  :D

 

Offline KE5FX

  • Super Contributor
  • ***
  • Posts: 1243
  • Country: us
    • KE5FX.COM
Re: Spartan 3E board
« Reply #22 on: November 22, 2018, 11:13:23 pm »
The Nintendo Gameboy ADV's dev board is still based on a Spartan2, and it's a neat and funny idea if someone wants to learn how to develop a video game on a real portable console  :D

Yeah, there's no question but that you can still do great things with Spartan3 and older parts, just as you can build a working computer by sticking a Z80 and a bunch of wires into a solderless breadboard.  That doesn't mean it's a good use of limited educational and/or R&D time.

It is true that flexibility in I/O standards has gotten somewhat worse since Spartan6, and it is also true that XDC constraints are a tire fire.  I don't mean to sound like a Xilinx evangelist.  But the original poster didn't leave us much to go on, in terms of what they are trying to accomplish.  Advising them to start two full generations behind the curve doesn't seem like a good idea.
 

Offline legacy

  • Super Contributor
  • ***
  • !
  • Posts: 4415
  • Country: ch
Re: Spartan 3E board
« Reply #23 on: November 23, 2018, 12:01:56 am »
Advising them to start two full generations behind the curve doesn't seem like a good idea.

And advising to stay with the last technology, just because it's super-fast, super-amazing, super-wonderfull ...  is just like being too obsessed with what the market needs to promote.

Besides, THE point was the cost of these boards! Since Spartan3 is obsolete, boards come very cheap!
 

Online asmi

  • Super Contributor
  • ***
  • Posts: 1188
  • Country: ca
Re: Spartan 3E board
« Reply #24 on: November 23, 2018, 01:11:18 am »
It is true that flexibility in I/O standards has gotten somewhat worse since Spartan6,
Can you please elaborate on that? The only standard that is missing on 7 series is 2.5 V (sans (B)LVDS), but I can't think of any chip that would require it except DDR1, which itself is long obsolete.
and it is also true that XDC constraints are a tire fire.
Well there is GUI for creating/editing timing constraints, although typing them directly is probably easier for experienced designers.
« Last Edit: November 23, 2018, 01:14:42 am by asmi »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf