Xilinx Zynq 7000 Sample Request

[diyaudio]

I'm looking at doing a another diy project project next year, I always had my eye on the Xilinx Zynq 7000 series, however I'm curious do Xilinx  even sample them ?

[julianhigginson]

I'm not aware of them sampling much, beyond brand new silicon to huge customers.... but if you have a relationship with your distributor,  wouldn't hurt to ask...

of course, there are all bga parts that will need a DDR3 RAM layout to actually live up to the whole point of having them...  so to evaluate the use of a zynq in pretty much any design it would be faster and cheaper to find an appropriate zynq eval board, that has the connectivity to add whatever other custom things you want, and prototype from there....

[ebclr]

No they don't sample, for simple mortals only fortune 500 can do this, i already have tried without success, they want to sell it, and is more expensive than boards with the zinc on board, the easiest way to have a Zinc is buying a Parallella Board

https://www.parallella.org/board/

[Howardlong]

I agree, it's definitely a non-trivial exercise to get a Zynq running.

Much as it leaves a lot to be desired as a test instrument and mid-level deveopment, the Red Pitaya is the least worst option I've encountered in terms of a reasonable cost development and proof of concept platform for the Zynq. The hardware is reasonably sound, it's just the nonsense around the documentation, much of which is out of date, and the way they try to monetize it by selling half baked apps.

Remember that this is an application processor plus FPGA, so typically you'll probably choose to get Linux running on it, which in itself is a significant project if doing it from scratch. For the project I used it on, we use Red Pitaya boards directly with our own hardware (frequency agile RF quadrature up and down converters). It's very low volume so we'd be nuts to spend months getting our own board working.

Another board I tried was the Zybo, which has a tutorial book http://www.zynqbook.com/downloads/The_Zynq_Book_Tutorials_Aug_15.pdf but I found it was geared towards ticking a few course boxes for students rather than getting you to a position where you could realistically do something with it on your own.

[julianhigginson]

there's more and more zynq options now. the trick is you probably want a 7020 unless your FPGA fabric needs are really low... so that cuts out all the cheapest zynq board options for a general play about platform....

One good low cost solution, depending on your io needs is the new pynq board (released yet?) from digilent. which is an arduino format thing with a pretty good 7020 part. basically it's their older arty board but upgraded to a zynq from the original artix7 part.

there's a crowd funded zynq som that just had disaster after disaster but stool might come through... snickerdoodle. looks good if they can ever make it.

there's also a Chinese option that's available to get now, and seems pretty solid (my friend is using it) it's in a raspberry pi format, but has two reasonably high density connectors at the bottom for plugging into a board of your own design with whatever you need attached. they also sell a more generic som that you can use.... I'm on my phone, so can't get the links right now but will try to do that later.

[diyaudio]

Looks like I might just go for a OMAP or i.mx6 platform, note this is only for educational purposes. so far I found lots of open OMAP and i.MX6 projects.

[Howardlong]

there's more and more zynq options now. the trick is you probably want a 7020 unless your FPGA fabric needs are really low... so that cuts out all the cheapest zynq board options for a general play about platform....

One good low cost solution, depending on your io needs is the new pynq board (released yet?) from digilent. which is an arduino format thing with a pretty good 7020 part. basically it's their older arty board but upgraded to a zynq from the original artix7 part.

there's a crowd funded zynq som that just had disaster after disaster but stool might come through... snickerdoodle. looks good if they can ever make it.

there's also a Chinese option that's available to get now, and seems pretty solid (my friend is using it) it's in a raspberry pi format, but has two reasonably high density connectors at the bottom for plugging into a board of your own design with whatever you need attached. they also sell a more generic som that you can use.... I'm on my phone, so can't get the links right now but will try to do that later.

Although I realise the OP's gone a different path now, one of the key things for all these complex SoC dev boards, irrespective of the SoC itself, is what the support is behind them, in terms of both direct and through the community. The Raspberry Pi series, although not the best spec'd boards out there, wins by miles on this, and it's the difference between having a small paperweight and a device you can actually do something without locking yourself away hermit style for a decade of frustration trying to flash an LED.

[julianhigginson]

Looks like I might just go for a OMAP or i.mx6 platform, note this is only for educational purposes. so far I found lots of open OMAP and i.MX6 projects.

wow... that's a *very* different thing to a zynq!
so, I guess you're going to decide on the project's eventual application sometime later?

I seem to remember Ti killing off a lot of their OMAP line a few years ago (getting out of the smartphone market??) so watch out for that... but yeah there's a lot of doco out there if you want to do something based on the beaglebone black design, for instance...

[tggzzz]

I'm looking at doing a another diy project project next year, I always had my eye on the Xilinx Zynq 7000 series, however I'm curious do Xilinx  even sample them ?

Why do you want a Zynq? You should be aware there is a very steep learning curve, especially with the tools. FPGA programming is worthwhile, but not easy.

No amateur should consider putting a Zynq on a board; they should simply buy one of the many boards out there.

If your interest is using ARMs, then just use one of the many astoundingly cheap ARM dev boards.

If your interest is using FPGAs, then just buy a small simple FPGA dev board. The principal difference between Xilinx FPGA families is in the IO structures (especially SERDES) and fabric size. There is a hard distinction between the "obsolete" ISE dev tools and the "futureproof" Vivado tools.

Choose Zynq if you need tightly coupled processor/logic/memory and/or Linux.

[diyaudio]

I'm looking at doing a another diy project project next year, I always had my eye on the Xilinx Zynq 7000 series, however I'm curious do Xilinx  even sample them ?

Why do you want a Zynq? You should be aware there is a very steep learning curve, especially with the tools. FPGA programming is worthwhile, but not easy.

No amateur should consider putting a Zynq on a board; they should simply buy one of the many boards out there.

If your interest is using ARMs, then just use one of the many astoundingly cheap ARM dev boards.

If your interest is using FPGAs, then just buy a small simple FPGA dev board. The principal difference between Xilinx FPGA families is in the IO structures (especially SERDES) and fabric size. There is a hard distinction between the "obsolete" ISE dev tools and the "futureproof" Vivado tools.

Choose Zynq if you need tightly coupled processor/logic/memory and/or Linux.

I normally take on side line projects like this every year, which means I commit to them for 1 full year and nothing else, I do feel I'm diligent enough to work meticulously on a project like this, I did something similar with a SHARC DSP in the past with 100% success. 

I really do wish they sampled the ZYNQ chips to companies for research.



 

[tggzzz]

I'm looking at doing a another diy project project next year, I always had my eye on the Xilinx Zynq 7000 series, however I'm curious do Xilinx  even sample them ?

Why do you want a Zynq? You should be aware there is a very steep learning curve, especially with the tools. FPGA programming is worthwhile, but not easy.

No amateur should consider putting a Zynq on a board; they should simply buy one of the many boards out there.

If your interest is using ARMs, then just use one of the many astoundingly cheap ARM dev boards.

If your interest is using FPGAs, then just buy a small simple FPGA dev board. The principal difference between Xilinx FPGA families is in the IO structures (especially SERDES) and fabric size. There is a hard distinction between the "obsolete" ISE dev tools and the "futureproof" Vivado tools.

Choose Zynq if you need tightly coupled processor/logic/memory and/or Linux.

I normally take on side line projects like this every year, which means I commit to them for 1 full year and nothing else, I do feel I'm diligent enough to work meticulously on a project like this, I did something similar with a SHARC DSP in the past with 100% success. 

Excellent. Sounds like you are going in with your eyes open. I expect you will find this equally worthwhile.

I really do wish they sampled the ZYNQ chips to companies for research.

Have a look at the Xilinx documentation for PCB layout - the requirements aren't trivial! Much better, IMNSHO to get it already designed, manufactured and tested so that you can concentrate on whatever value you add, i.e. your design:)

There are many such boards available; you might start by having a look at this selection: https://shop.trenz-electronic.de/en/search?sSearch=zynq especially the ArduZynq and ZynqBerry.

[langwadt]

you can get a Zynq board for $99 http://www.myirtech.com/list.asp?id=502

I doubt you could build a board for that even if you got the components for free, just the pcb alone is would probably cost more than
that in prototype quantities

[Howardlong]

you can get a Zynq board for $99 http://www.myirtech.com/list.asp?id=502

I doubt you could build a board for that even if you got the components for free, just the pcb alone is would probably cost more than
that in prototype quantities

As I stated before, take care with any board, there's a world of difference between a SoC board with an extensive ecosystem behind it and one without.

[tggzzz]

Interfaces without at least many dozens of I/O lines with good signal integrity up to at least around 350 MHz are pretty useless for a FPGA evaluation kit.

Very true. Even most of those with enough i/o lines often have too few grounds , or have them carefully grouped together!

However boards with PMOD class interfaces are sufficient for students learning about FPGA system design and implementation.

[langwadt]

you can get a Zynq board for $99 http://www.myirtech.com/list.asp?id=502

I doubt you could build a board for that even if you got the components for free, just the pcb alone is would probably cost more than
that in prototype quantities

As I stated before, take care with any board, there's a world of difference between a SoC board with an extensive ecosystem behind it and one without.

sure but that also goes for any board you build yourself, and that board look to be very similar to the $199 microzed

[langwadt]

Interfaces without at least many dozens of I/O lines with good signal integrity up to at least around 350 MHz are pretty useless for a FPGA evaluation kit.

Very true. Even most of those with enough i/o lines often have too few grounds , or have them carefully grouped together!

However boards with PMOD class interfaces are sufficient for students learning about FPGA system design and implementation.

the microzed uses connectors rated for something like 10Gb/s and have a ground pair differential pair

[tggzzz]

Interfaces without at least many dozens of I/O lines with good signal integrity up to at least around 350 MHz are pretty useless for a FPGA evaluation kit.

Very true. Even most of those with enough i/o lines often have too few grounds , or have them carefully grouped together!

However boards with PMOD class interfaces are sufficient for students learning about FPGA system design and implementation.

the microzed uses connectors rated for something like 10Gb/s and have a ground pair differential pair

Yes. I have an early variant. At the time it was by far the most affordable board with decent I/O, and they have put thought into how the I/O banks can run at whatever voltage is appropriate. There are more options available now, but I haven't evaluated them (except to note many have very poor grounds!).

I have no complaints about the MicroZed itself nor about the (vitally important) support community, but Avnet's ordering and delivery process was significantly defective.