• 0
gutobaldi

FPGA board Eclypse Z7 for power electronics application - advice

Question

I'm a newbie here and I’m working on a inverter test bench project where I have two three-phase inverters connected through an inductive load. The idea is to emulate in real time the behave of an electrical machine. To be clearer, the first inverter is going to be tested (Device Under Test) and the second one plus the inductive load must behave like an electrical machine. To do so, we are going to use a FPGA board, which must have the following specifications:

- Capable to drive both inverters switching at 50kHz (each inverter has 6 MOSFETs switching at this frequency)

- 20 digital I/O

- 4 ADC with 16 bits (ideally) and 20MHz at least. The ADCs can be integrated or not in the FPGA board

- Capable to communicate in real time with Matlab/Simulink

- The board will be placed inside the test bench, in a temperature around 50°C

We know that we are going to use Vivado to the VHDL coding, but we are not sure about the ADCs, regarding the Eclypse Z7 with the two Zmod ADCs. We want to code the least possible in VHDL (no VHDL coding if possible), so my questions are:

1) Are we going to have to code the ADC data acquisition?

2) Is the VHDL code generation done automatically by Matlab? I do have the toolbox for HDL coding.

Our budget is around €1000,00.

I would like to know if the board Eclypse Z7 with the two Zmod ADCs is a good choice for the application and if you have another advices it would be highly appreciated.

I hope I made myself clear. Thank you!

 

Share this post


Link to post
Share on other sites

4 answers to this question

Recommended Posts

  • 0

Your requirement for 4 ADC channels plus 20 GPIO lines will be tough using the EclypseZ7. At the moment you only have a choice of 2 SYZYGY ADC pods, one from Digilent and one from Opal Kelly. It's been quite a few years since I used the FPGA/MATLAB toolchain and then it was limited to particular platforms. This might be an issue.

I suspect that the best shot for your project involves custom hardware and software design and development, but I haven't looked around for the simplest options available.

Can the Eclypse-Z7 handle about 160 MB/s of ADC data? If you can do processing in HDL to reduce the data rate considerably perhaps. I don't see a quick and dirty MATLAB solution here; though that not a particularly well informed opinion.

Unfortunately, the decision to make Digilent's first SYZYGY platform Z7000 based makes it unsuitable for a wide range of applications. A Genesys2 with SYZYGY would have been a home run, though perhaps not for MATLAB HDL dependent projects.

Requiring a 50 C environment complicates your hardware acquisition work considerably as I doubt that you'll find vendors willing to spec their boards at that temperature, at least for a low budget project.

Share this post


Link to post
Share on other sites
  • 0
1 hour ago, zygot said:

Your requirement for 4 ADC channels plus 20 GPIO lines will be tough using the EclypseZ7

Yes, you are right, I did not pay attention to the number of GPIO.

1 hour ago, zygot said:

It's been quite a few years since I used the FPGA/MATLAB toolchain and then it was limited to particular platforms. This might be an issue

I've talked to some people about this, and I think that we can do the communication between the FPGA and Matlab/Simulink without a problem.

1 hour ago, zygot said:

I suspect that the best shot for your project involves custom hardware and software design and development

Can you be a little more specific about this?

1 hour ago, zygot said:

Can the Eclypse-Z7 handle about 160 MB/s of ADC data? If you can do processing in HDL to reduce the data rate considerably perhaps

I have no idea and that is exactly what we don't want to do.

1 hour ago, zygot said:

Requiring a 50 C environment

It is more like a worst case, I don't think we are getting there.

Maybe we can handle the project with only 2 ADC. Do you mind to explain me what is the issue with Matlab? There is one more thing that I haven't talked about, which is the model we are going to implement in Simulink. The signals going to the MOSFET's gate are coming from this model and the ADC data will be used in the Simulink's model in a closed loop regulation.

Anyway, the number of GPIOs has already discarded this board. And thanks for your answer!

Share this post


Link to post
Share on other sites
  • 0
Posted (edited)

No doubt that FPGA vendor support for MATLAB/SIMULINK has improved since my experiences using it so I'm probably not the one who should be providing guidance on that. My comment about your project likely requiring custom hardware, HDL, and possibly software is more of a hunch based on experiences with general purpose hardware and support. There are a lot of variables to consider. ZYNQ, and the requisite AXI infrastructure adds it's own complexity, especially if it is a data bottleneck.  ADC conditioning is a complexity specific to any application. Generally, general purpose off the shelf hardware isn't ideal for particular applications. For your class of applications isolation from out of spec signal anomalies is frequently a concern.Anything involving high currents high voltages and inductive loads makes me think about a lot of potential for problems. These are usually taken care of in the hardware design phase and general purpose hardware is designed for a restricted range of applications by definition.

If you can find off the shelf components designed to fit your objective then concentrating on the mathematical model is easy. If you have to scrounge parts then the burden is on you to figure out all of the details and make sure that you are covered. Usually, this involves more work than anticipated, especially of you don't have prior experience with similar projects. I realize that all of this is rather general and unspecific but that's what I'm limited to here. 

One problem with MATLAB is that while it's affordable in an academic setting few companies in the commercial world are willing to pay for it so there's a gap in experience using it between the those worlds.

Edited by zygot

Share this post


Link to post
Share on other sites
  • 0
Posted (edited)

I should note that SYZYGY is still in its infancy. On the right platform there's no reason why 1 standard  pod couldn't have 4 100 MHz ADC channels leaving a second port open for GPIO. Some forward thinking vendor will eventually figure this out but user's will have to wait. That doesn't make what's currently available unusable for a limited range of applications.

[edit] Do check out Opal Kelly's website. They offer breakout boards for prototyping.

Edited by zygot

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now