Other DSO manufacturers besides GW-Instek that use Xilinx Zynq-7000 series

[pascal_sweden]

I wanted to find out through this thread, if there are other DSO manufacturers out there,
besides GW-Instek, that use the Xilinx Zynq-7000 in their digital oscilloscopes?

The Zynq-7000 architecture is very promising, and results in a very responsive digital oscilloscope.
I really like that GW-Instek selected this architecture, unfortunately I don't like the design of their scopes,
and therefore I am on the lookout for another DSO manufacturer that uses this architecture, and that has better design taste

I know about the Red Pitaya product, built on a Zynq-7000, but this is not what I am looking for.
There has to be another DSO manufacturer, that uses this architecture, but I haven't found one yet.

[saturation]

On the $6000 NI Virtual Bench Dave identified Zynq SoC on board.

https://www.eevblog.com/forum/blog/eevblog-875-ni-virtualbench-teardown/msg934441/#msg934441

Signal path also did a review of the lower end $2000 NI, but didn't go far enough to identify the core CPU so don't know if this is Zynq powered.



On Instek, there are 11 variants on the platform: 6 for the 2000E and 5 for the 1000B.

[pascal_sweden]

Thanks for the update!

That instrument classifies however under the USB device section, and not under the stand-alone DSO section

Moreover that device is not in the price range, that I am looking for.

So hereby I further specify my search:

I am on the lookout for other DSO manufacturers besides GW-Instek, that produce stand-alone digital storage scopes which are built around the Xilinx Zynq-7000 SoC architecture, and that come in a 500-1500 USD price range

Besides the Xilinx Zynq-7000 SoC architecture, the newer Xilinx Zynq UltraScale MPSoC architecture is a valid candidate as well, in the stand-alone DSO that I am looking for

[Someone]

Are you wanting to make an alternative firmware image but want to use a cheap product as a base? Otherwise the underlying architecture is mostly irrelevant.

[nctnico]

Otherwise the underlying architecture is mostly irrelevant.

Not quite. Equipment with lots of processing horsepower makes processing tasks like decoding, filtering, math, FFT, measurements, etc very fast for very low cost. Just look at Dave's FFT comparison and see how the GW Instek scope runs circles around every other scope he tested. There is a good reason higher end scopes are based on PCs: it is all about processing power.

[Kilrah]

Yeah but there's nothing miraculous about the Zynq, it's an awesome platform but in the end it's nothing but a dual core CPU and an FPGA in the same package. So there's no reason a different solution that packs a comparatively powerful CPU and FPGA in discrete packages can't do just as good.

The Zynq may save a little in parts and/or assembly costs meaning a bit more power than a discrete solution for the same price, but I'd tend to think any advantage most likely comes from good IP/software design and optimization.

[pascal_sweden]

The Zynq SoCs and MPSoCs from Xilinx are a natural fit for design teams consisting
of software and FPGA hardware engineers. However, teams with limited or no
hardware engineers, have been challenged due to the RTL (VHDL or Verilog)
development expertise needed, to take full advantage of the benefit of the device.

To resolve this challenge and enable more design teams to take advantage of Zynq
devices, Xilinx has introduced SDSoC, a new C/C++ development environment.
The SDSoC development environment enables the broader community of embedded
software developers, to leverage the power of hardware and software
‘all programmable’ devices.

The SmartConnect interconnect optimization technology attacks the IP interconnect
bottleneck at the system level, by automatically optimizing the interconnect for the
specific design’s requirements around throughput, latency, and area while providing
the optimal performance/watt.

[EEVblog]

Interesting question, I don't recall seeing any other bench scopes...

[H.O]

The Zynq SoCs and MPSoCs from Xilinx are a natural fit for design teams consisting
of software and FPGA hardware engineers. However, teams with limited or no
hardware engineers, have been challenged due to the RTL (VHDL or Verilog)
development expertise needed, to take full advantage of the benefit of the device.

To resolve this challenge and enable more design teams to take advantage of Zynq
devices, Xilinx has introduced SDSoC, a new C/C++ development environment.
The SDSoC development environment enables the broader community of embedded
software developers, to leverage the power of hardware and software
‘all programmable’ devices.

The SmartConnect interconnect optimization technology attacks the IP interconnect
bottleneck at the system level, by automatically optimizing the interconnect for the
specific design’s requirements around throughput, latency, and area while providing
the optimal performance/watt.

And here's the source of those statments link, and link.

[Someone]

Yeah but there's nothing miraculous about the Zynq, it's an awesome platform but in the end it's nothing but a dual core CPU and an FPGA in the same package. So there's no reason a different solution that packs a comparatively powerful CPU and FPGA in discrete packages can't do just as good.

You got it Also, if you buy a scope you can test its performance for the important uses you have, it doesn't matter if there is a desktop intel CPU (Lecroy) or an FPGA (Tek, Rigol, etc) or an ASIC (Keysight) what matters is the end result.

The Zynq may save a little in parts and/or assembly costs meaning a bit more power than a discrete solution for the same price, but I'd tend to think any advantage most likely comes from good IP/software design and optimization.

Price of a zynq part has been much higher than comparable separate CPU + FPGA parts, so its been targeting space constrained applications. With serious competitors creeping into the market prices might come back to a realistic point or who knows!

To resolve this challenge and enable more design teams to take advantage of Zynq
devices, Xilinx has introduced SDSoC, a new C/C++ development environment.
The SDSoC development environment enables the broader community of embedded
software developers, to leverage the power of hardware and software
‘all programmable’ devices.

Don't believe the hype, high level synthesis has been around a long time and still isn't able to extract the same efficiencies as poorly written and designed HDL. Even Altium had a half decent attempt at it with little impact:
https://en.wikipedia.org/wiki/High-level_synthesis

[nctnico]

I recall a thread where someone tried Xilinx's C to FPGA compiler and it seemed to work quite well. OTOH what matters most nowadays is speed because resources are abundant. When I started with FPGAs there where devices with 32 or so block rams of 512 bytes (or so) each. You really had to plan the usage at the design stage. In a current project even a low cost FPGA has over 100 blockrams holding 4kB each.

[pascal_sweden]

High-Level Synthesis has become a reality!

Please take a look at CyberWorkbench from NEC:

C-based integrated environment for System LSI design, that provides an "All-in-C" based design flow.

Digital designs can be dominated by either control or data path logic. This tool uses powerful synthesis
algorithms to generate control and data flow logic allowing the generation of any type of digital design.

Here is an impressive video from NEC about CyberWorkbench:
http://www.nec.com/en/global/prod/cwb/video/index.html

Enjoy!

[coppice]

High-Level Synthesis has become a reality!

Please take a look at CyberWorkbench from NEC:

C-based integrated environment for System LSI design, that provides an "All-in-C" based design flow.

Digital designs can be dominated by either control or data path logic. This tool uses powerful synthesis
algorithms to generate control and data flow logic allowing the generation of any type of digital design.

Here is an impressive video from NEC about CyberWorkbench:
http://www.nec.com/en/global/prod/cwb/video/index.html

The videos have been than impressive for over 20 years.

[pascal_sweden]

If you want to improve your HLS skills this summer, then I recommend you to buy this book

http://www.hlsbluebook.com/

[ebastler]

The Zynq SoCs and MPSoCs from Xilinx are a natural fit for design teams consisting
of software and FPGA hardware engineers...

High-Level Synthesis has become a reality!
...
Digital designs can be dominated by either control or data path logic. This tool uses powerful synthesis
algorithms to generate control and data flow logic allowing the generation of any type of digital design
...

Pascal -- please do yourself and us a favor: If you quote some manufacturer's marketing blurb, please put in in quotes and explicitly state the source. That also applies if you transcribe marketing blurb from a manufacturer's video ;-)

I am not opposed to relaying manufacturer information here. But I think it is good practice to make it very clear which parts are your personal (hopefully unbiased) assessment, and which ones are quoted from a manufacturer who obviously is far from unbiased.

[coppice]

If you want to improve your HLS skills this summer, then I recommend you to buy this book

http://www.hlsbluebook.com/

Is this recommendation based on your experience of using the book to help you get an actual project successfully completed?

[saturation]

You're welcome, alas I can't recall anyone else which a full bench DSO using any of the Zynq technology.  I too am interested in its capabilities and its one reason I bought the Instek.

I recall the Zybo test board has an DSO demo app, and that's all I think there is as of this time, and all but the Instek are USB instruments.

While the Instek may not appeal to some aesthetically, IMHO its it the current main showcase of Zynq performance as a DSO.   

Prior reviews have shown its speed, performance, far reduced parts count, and finally, reliability.   All that translates to at least a very functional product with much lower manufacturing cost.  It is solid performer I would be confident to use as a test device in front of peers and even critical experiments.

Thanks for the update!

That instrument classifies however under the USB device section, and not under the stand-alone DSO section

Moreover that device is not in the price range, that I am looking for.

So hereby I further specify my search:

I am on the lookout for other DSO manufacturers besides GW-Instek, that produce stand-alone digital storage scopes which are built around the Xilinx Zynq-7000 SoC architecture, and that come in a 500-1500 USD price range

Besides the Xilinx Zynq-7000 SoC architecture, the newer Xilinx Zynq UltraScale MPSoC architecture is a valid candidate as well, in the stand-alone DSO that I am looking for

[pascal_sweden]

Pascal -- please do yourself and us a favor: If you quote some manufacturer's marketing blurb, please put in in quotes and explicitly state the source. That also applies if you transcribe marketing blurb from a manufacturer's video ;-)

I am not opposed to relaying manufacturer information here. But I think it is good practice to make it very clear which parts are your personal (hopefully unbiased) assessment, and which ones are quoted from a manufacturer who obviously is far from unbiased.

Your advice has been noted, and taken into account! Thanks for the advice

[pascal_sweden]

If you want to improve your HLS skills this summer, then I recommend you to buy this book

http://www.hlsbluebook.com/

Is this recommendation based on your experience of using the book to help you get an actual project successfully completed?

I haven't used the book yet for an actual project, but I read a lot of good remarks about it.

But again, to follow ebastler's advice, my recommendation is probably biased because of reviews, that possibly are controlled by the manufacturer, so it is probably better to withdraw my recommendation, and just inform you about the availability of the book, and decide yourself if it is any good or not,
until I have actually used the book extensively and verified personally that it helped me to take an actual project to completion, with a significant reduction of time to market

[pascal_sweden]

Maybe one day we will see a DSO from NEC, designed with CyberWorkbench

The "All-in-C" Digital Storage Oscilloscope from NEC =)

And hereby the feedback expected from other forum users on my post:
"Keep on dreaming Pascal!"

[coppice]

Maybe one day we will see a DSO from NEC, designed with CyberWorkbench

The "All-in-C" Digital Storage Oscilloscope from NEC =)

And hereby the feedback expected from other forum users on my post:
"Keep on dreaming Pascal!"

NEC hasn't existed as a semiconductor company for several years. CyberWorkbench is over 20 years old. If it hasn't affected the DSO market to date, why would you expect that to change?

[pascal_sweden]

NEC has indeed kept this tool in-house for all these years, but they have further perfectionized the tool over the years.

Only since 2011, they released it as a commercial product outside Japan, to get some traction for their EDA tool on the US market, and compete with Synopsys and Cadence:
http://www.eetimes.com/document.asp?doc_id=1264211

And more recently they have reinforced that effort by partnering up with Aldec:
https://www.aldec.com/en/company/news/2013-07-31/199

Whether they were successful in this attempt or not, is another question.

However, it has to be said, that NEC was the first company, to invest substantial amount of money in HLS projects, that led to an actual commercial tape out of a chip.

This is greatly thanks to the effort of Dr. Kazutoshi Wakabayashi, who is an internal promotor and believer of HLS in the company.

To learn more about the man behind HLS at NEC, take a look at this article:
www.nec.com/en/global/techrep/journal/g05/n03/pdf/a242.pdf

Hereby an overview of the publications from Dr. Kazutoshi Wakabayashi:
https://www.researchgate.net/profile/Kazutoshi_Wakabayashi/publications

The most recent publication dates back to January 2015, which confirms that HLS is still an active topic within the NEC organization.

[coppice]

However, it has to be said, that NEC was the first company, to invest substantial amount of money in HLS projects, that led to an actual commercial tape out of a chip.

Who says that, and what dates do they give?

[pascal_sweden]

Did you read the backgrounder about Wakabayashi that I provided before?
www.nec.com/en/global/techrep/journal/g05/n03/pdf/a242.pdf

On the link below, they indicate that they are number one in Japan, and mention some actual customers:
https://www.semiwiki.com/forum/content/2542-hw-prototyping-hls-dac.html

Hereby some extra references to books and articles that cover HLS, which are worth looking at IMO:

Book 1: High-Level Synthesis - from Algorithm to Digital Circuit
http://www.springer.com/us/book/9781402085871

Book 2: Embedded System Design - Modeling, Synthesis and Verification
http://www.springer.com/gp/book/9781441905031

Book 3: High-Level Synthesis Blue Book
http://www.hlsbluebook.com/

Book 4: ESL Models and their Application
http://www.springer.com/us/book/9781441909640

Article 1: High-Level Synthesis for FPGAs - From Prototyping to Deployment
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5737854&newsearch=true&queryText=High-level%20synthesis%20for%20FPGAs:%20From%20prototyping%20to%20deployment.&fname=&lname=&title=&volume=&issue=&spage=

Article 2: C-based SoC design flow and EDA tools: an ASIC and system vendor perspective
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=898829&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D898829

Hereby some references to past seminars about HLS, which possibly are organized again in the future:

HLS seminar 1: C-Based High Level Synthesis Technical Webinar - Organized by TechDream in Santa Clara
http://tech-dream.com/hls_webinar/

NEC highlighted some success stories on this seminar.

HLS seminar 2: The New Wave of the High Level Synthesis - Organized by Electronic Chips & Systems design Initiative
http://ecsi.org/workshop2008/new-wave-high-level-synthesis

NEC was also present on this seminar.

If you know about any upcoming HLS seminars in Europe (preferably free seminars), please provide more details in this thread

[pascal_sweden]

If the current DSO manufacturers are not taking up and adopting the Xilinq-7000 series in their scopes,
there is another solution!

Here is another option that I see:

Use everything from an existing oscilloscope, including the analog front-end, the controls and knobs, the display, the power supply, but bypass the main processor and insert a Xilinx Zynq-7000 instead. This could in fact be the Zybo board from Digilent.

Would it be easy to redirect the output of the analog front-end? Or would it be possible to tri-state the main FPGA and take over control through a Xilinx Zynq-7000? Would there be space inside to add an extra Zybo board without replacing an existing board?

Using this modification, you could retro-fit an existing oscilloscope and convert it into a powerful oscilloscope with super powers that beats the GW-Instek GDS-2000E

Hereby some examples of new model numbers that I propose, where Z stands for Zynq-7000 option, or R stands for Retro-fit option =)

Rigol DS1052E-Z
Rigol DS1054Z-Z
Siglent SDS1052DL-Z
Siglent SDS1102X-Z
Owon SDS7102-Z

Rigol DS1052E-R
Rigol DS1054Z-R
Siglent SDS1052DL-R
Siglent SDS1102X-R
Owon SDS7102-R

It should be possible, to open up the scope just once for the retro-fit, mount the Zybo board, a JTAG cable and other debugging cables into the scope, and after close it again.

The external design of the scope stays fully intact, while at the same time you would have full access to the scope, to control and monitor it from the JTAG interface and the other debugging cables.

Special labels or decals could be made, with a nice big "Zynq-7000" or the wording "Retro-fit", to proudly show the transformation

Maybe a small batch of these scopes can be produced, for an FPGA and DSP class room