It wasn't long ago that I posted about saying "Hello" to FPGA thanks to a prize received from one of the many FPGA quizzes on element14. As it turns out, the random number generator must have been smiling down at me because I was selected as yet another prize-winner for the FPGA Technology Quiz. As someone who visits the community on a regular basis, the surprise was spoiled as others living closer to the USA seemed to have received their parcels earlier ... mine only arrived on Friday.
The box was bigger than normal and a little crunched from the round-the-world journey ...
Enclosed within is the Exploring Zynq MPSoC with PYNQ and Machine Learning Applications book by Lousie Crockett, David Northcote, Chris Ramsay, Frazer Robinson and Bob Stewart, published by the University of Strathclyde, Glasgow. There is also a mini screwdriver set which is a clone of the popular Xiaomi Wiha set which I do also own. The latter definitely comes in handy.
The book itself is a rather substantial reference, over 600 pages in total. While you don't get a hard-cover or glossy pages inside, there is colour print which is a very nice feature. The book covers a lot of ground with regard to the Zynq MPSoC, which is the latest generation of "hybrid" FPGA and CPUs combined into a single SoC. It seems that over the years, purely FPGA solutions have slowly given way to hybrids, such as the original Zynq which incorporated ARM cores with FPGA logic units. This latest iteration takes things a little further, by integrating a fully-featured SoC rather than just an ARM core, so think of it as a mid-range smartphone SoC with an FPGA glued onto it. It's quite a bit more complex, but the benefits can be amazing, as the familiarity of programming for CPUs and having access to a proper GPU and video acceleration opens up new applications. In fact, already, the applications for these hybrid chips have reached even oscilloscopes as it seems there are Rigol units based on Zynq chips already. I read up on the specifications of the Zynq Ultrascale+ RFSoCs - the ADC/DACs are quite impressive indeed for SDR and other applications.
But perhaps the most surprising thing to me is that the book itself is published under open source terms! So if you weren't lucky and didn't win, you can download the full eBook PDF version via their website at https://www.zynq-mpsoc-book.com/ . While you're at it, you might also want to grab the prior book that covers the Xilinx Zynq called "The Zynq Book" at The Zynq Book . What a world we live in! I feel so grateful to see such good material being openly available, as it seems a lot of the world is headed this way because of AI, edge-processing and the like. Whether I'll have the chance to put it into practice will be something that remains to be seen.
As for the screwdriver set, the bits seem to differ a bit from the included ones in the Xiaomi Wiha set but it has otherwise a very similar form. The outer casing is tastefully branded with Xilinx and element14 logos. The bits themselves don't seem bad at all, although the finish is on the driver does not seem as consistent, the spinny thumb-rest on the back of the driver doesn't spin as smoothly nor does the catch that holds the sliding tray hold as strongly as in the Xiaomi version. Rarely does any of the cloned kits achieve the same level of quality, but at least the thought behind this kit is excellent, as every engineer is going to have to repair or disassemble something at some point. I can see myself getting a lot of use out of the kit!
Turning the bits around so the sizes can be read, I found the bit marked S2 SL4.0, a slotted bit, seemed to be out-of-tolerance. Incidentally, I picked up this bit first to do up a terminal connector block when I realised it would not slide into the driver and got jammed. Measurements show that the bit was 3.94mm from flat-to-flat, which is a little bigger than the 3.91mm from flat-to-flat for the other bits. All the other bits slid into the driver just fine, retained by the internal magnet, so it seems that it was just the issue with the one bit. To think that 30um would make that much difference!
As always, thanks to element14 and Xilinx for this useful prize!
