In Preparation for the AVNET MiniZed (Xilinx Zynq XC7Z007S SoC) RoadTest. What Do You Need?

How about a webinar on Zynq?

Scott

Hi Randall,

It looks like this board would be perfect for an introductory course on how to program and use Xilinx FPGA to do small projects.

If a short little training set of classes, similar to the Arduino Starter Kit, was available, I can see a lot of people who would be interested in dipping their toes into this device.

The board price is attractive, but some projects need to be added to show the range of capability available on this particular chip.

My basic question would be what is the complexity limit for building a solution on this board and what the next range of board costs should I need to go larger?

DAB

Hi Randall,

a great idea will be a sort of 10 project in 10 days (maybe something like already happened with PSoC4 but there was larger). I mean, instead of a bare webinar a piece of knowledge base together with a project growing from dummy to higher quality. I approach this idea as probably this maybe a greater appeal to road tester as this world is relatively new in the scenario of makers (at least this is my impression).

A mix between the mentioned 100 projects in 100 days and the Arduino cookbook. Yes, the board cookbook, why not?

Enrico

DAB,

AFAIK from the most recent posts here and there about FPGA and Xilinx it seems to me it is one of the most recent (or the most) entries in the scenario of suppliers products offered by Element14/Farnell. To be precise I have the perception that it is a new range of products and sounds almost revolutionary in the FPGA wrold, just because these board and others all branded Xilinx. You can find the entire range of base components in the article In Preparation for the AVNET MiniZed (Xilinx Zynq XC7Z007S SoC) RoadTest. What Do You Need?

What left me surprised is the price difference between the IC models:

The PGA Zynq-7000 Family 23000 Cells 667MHz 28nm 1V 400-Pin WCBGA Tray cost about 60 Euro

while the FPGA Spartan-6 LX Family 3840 Cells 45nm (CMOS) Technology 1.2V 144-Pin TQFP (maybe the lover version?) cost about 20 Euro

and the FPGA Artix-7 16640 Cells 28nm Technology 1V 236-Pin FPGA cost few less than 300 Euro

I have difficult to collocate the IC of this board (that is of the family of the first IC I think) respect what is in the mentioned presentation page.

Enrico

DAB wrote:

 

It looks like this board would be perfect for an introductory course on how to program and use Xilinx FPGA to do small projects.

 

My basic question would be what is the complexity limit for building a solution on this board and what the next range of board costs should I need to go larger?

At US$89, the MiniZed is an attractive price for a Xilinx Zynq with ARM Cortex-A9 and 512MB DDR.  That said, in my opinion a Xilinx Zynq is not a good introductory FPGA.  It's a complex architecture and you have to deal with the added complexity of a "motherboard on chip", which usually runs Linux. Learning about FPGAs with a simple board like the DEO-NANO or a less-expensive entry-level FPGA board is probably enough learning curve steepness.  But of course, chacun a son goût (YMMV).

Regarding what other boards are available, take a look at some of the other Zed boards at the MiniZed link.

Enrico Miglino wrote:

 

I have difficult to collocate the IC of this board (that is of the family of the first IC I think) respect what is in the mentioned presentation page.

For your amusement, here's a quote from an early Zynq article: "Zynq-7000 EPP Sets Stage for New Era of Innovations", Xcell Journal, Second Quarter, 2011:

With a starting price point below $15, we are really making it hard for companies to justify the cost and risk of designing any ASIC that is not extremely high volume.

Xilinx has had a dreadful time getting Zynq cost down.  Given the Zynq list prices and typical FPGA development board volume, I can only conclude that the MiniZed must be heavily subsidized.  Smart chip vendors subsidize development boards: if more people can afford to try your board and they enjoy the experience, you'll get more people designing with your chip.  Lattice must do this with the US$22 iCEstick: it costs $20 just for a board with an FTDI FT2232H and USB connector, and that's before you add the iCE40 FPGA, LEDs, and PMOD connector.  Cypress has some nice PSoC prototyping boards for $4 or $10.

The original ZedBoard cost $395, but you could only buy one at that price.  If you wanted multiple boards, it cost $495.

I wonder how hot that Zynq gets, particularly if it's under an Arduino shield.  I remember that Parallella, which has a Zynq as well as its Epiphany parallel processing chip, had a lot of heat problems early on.

I'd like to echo John's comment that this board is not really suitable for initial learning about FPGAs - the Zynq chips are very, very difficult to get to grips with. There is all the difficulty of the series 7 FPGA and an ARM applications processor with a sophisticated interface to the FPGA as well.

Don't misunderstand - the Zynq is a great chip - but it is complicated. The basic documentation for the memory interface on the series 7 FPGA (that's the little tiny bit in the corner of the chip that talks to DDR3 RAM) runs to 678 pages.

The Zedboard doesn't, in my opinion, have a useful IO configuration  - it doesn't have Ethernet so the ARM, which typically runs Linux, has no useful IO port. The Arduino interface and 2 PMODs is about enough to connect a serious ADC and some debugging pins (I know this because I've used a similar boards (Arty and Nexys Video)).

So it's a bit of tiger in a very small box.

If you want to do more than just run some demo stuff where you point and click though many pages of detailed instructions then you will need to to do some serious work and it will take a long time.

If you have decided to use this chip for a project and need to get the hang of it and the tools, the board is a cheap way to get started - otherwise I can't see much point in it. If you want to learn about the Xilinx 7 series FPGA then an ARTY is a better route (Artix FPGA rather than Zynq so no ARM on board).

MK

Michael,

considering your opinion trustable - as I do - as it is also following logic and a minimal knowledge of the FPGA world, why do they assume that this is (one of) the best solution for projects oriented to hobbyists and makers? Just based this consideration on the price of this board respect he rest of the family or competitors?

I have approached in past a couple of very difficult problems based on FPGA. I found this kind of logic a sort of miracle for the possibilities and flexibility but what I discovered in that period is that when you join a class of devices it is not easy to abandon them: most of the FPGA compilation is done by the software logic "interpreting" your design to create the gates. And this sort of fuzzy behaviour vary by platform to platform making almost difficult to move the knowledge to different FPGA family products. But maybe my knowledge is a bit outdated...

The point is that under this vision, if this producer finds the right way to evangelise a wide group of users investing on them this contributes creating a robust market. As far as I know at the date, what can be considered the best FPGA approach model is where the better and cheaper development platforms are offered including examples etc. reaching as much as possible new potential users.

Enrico

Enrico Miglino wrote:

 

... why do they assume that this is (one of) the best solution for projects oriented to hobbyists and makers? ...

Enrico, I don't see this claim being made.

The description says it's for entry-level Zynq developers.