Test and Measurement Forum

Frequency Response Analyzer with Bode Plots

Post discussions on applications you are writing

Re: Frequency Response Analyzer with Bode Plots

Postby dantooth » Mon Jul 04, 2016 3:48 pm

Thanks Aaron.

Yes, agreed, the low frequency response is not usually where we're looking at primarily.
Also you're right that it is +/-50mV full scale is the most sensitive setting for 3206B.

I was also wondering - do you make use of the moving average filter? That increases the effective resolution of the ADC, at the expense of lower overall sampling frequency. Could be useful, maybe at low frequencies where I guess you inject a small signal due to the high gain of the control loop at the low frequencies?
(When I fire up my Picoscope software GUI, then the averaging is found under the Channel Options for that particular channel and the Resolution Enhancement allows you to increase the effective resolution up to 12 bits.)

Actually that reminds me - are you planning on implementing the adjustable stimulus amplitude across the frequency range? I think that is usual method, as I read up on the subject.
I know it's easy for me to ask, but takes time for you...

When I look at various videos of people doing these measurements using the typical gain/phase analyzer's on the market, then they are using scope probes and other cable connections that use co-ax cable, but then have either a long ground leads (in the case of a scope probe) or flying leads with hook-type connectors on the end (in the case of the cable that injects the signal from the transformer). This kind of test set-up picks up a lot of switching noise (from the dc/dc converter) and couples it right into your signal, increasing the noise. To greatly reduce this there are better ways to connect the probes into the circuit using soldered connectors. My picoscope came with some adapters that allow you to remove the probe tip and instead fit something that would connect right into a miniature BNC connector that is soldered onto the PCB. That would keep the scope shielded for right up to the PCB. The link below shows what I'm on about, but the solution used there is a bit makeshift and I think the adapters I mention would be even better.
http://www.eetimes.com/document.asp?doc_id=1273282

Regards Dan
dantooth
Newbie
 
Posts: 0
Joined: Wed Jun 29, 2016 4:56 pm

Re: Frequency Response Analyzer with Bode Plots

Postby hexamer » Tue Jul 12, 2016 10:51 pm

Hi Dan,

Sorry for the delay in responding. I've been studying some other signal processing enhancements related to a topic mentioned earlier in the thread.

I appreciate the recommendation for resolution enhancement. Provided sufficient background noise, I think it could help with small signal cases. Just a note, the resolution enhancement in the PicoScope application is implemented in the PC software. So I'd have to implement equivalent processing in the FRA app. The modular design should make that relatively straightforward though.

Adaptive input signal amplitude is definitely at the top of the priority list - it's probably the most important enhancement for SMPS loop analysis. I've been mulling over various options, but haven't started implementing it yet. While I can't commit to a timeline, I assure you it's pretty important to me!

Thank you for pointing out the importance of probe connections. I like what you're proposing. I assume you're talking about something like placing a U.FL connector on the board and then running a coax of appropriate impedance to the scope. Minimizing apparent ripple and switching transients would also help the dynamic range problem.

Thanks,

Aaron.
hexamer
Advanced User
Advanced User
 
Posts: 62
Joined: Tue Aug 12, 2014 10:09 pm

Re: Frequency Response Analyzer with Bode Plots

Postby dantooth » Wed Jul 13, 2016 7:38 am

HI Aaron
Yes regarding the probes. My probes come with a connector that you can attach to your scope probe, after you have removed the tip and gnd lead. So you can use the scope probe to connect into that connector on your pcb.
(The gnd lead picks up a lot of switching noise and makes the probed signal look noisy when in fact it is noise coupled into the probe lead.)
Regards Dan
dantooth
Newbie
 
Posts: 0
Joined: Wed Jun 29, 2016 4:56 pm

Re: Frequency Response Analyzer with Bode Plots

Postby nano » Wed Aug 17, 2016 7:10 pm

dear hexamer,

nice piece of software!!! Works flawlessly with my 2205a. :D

Wanted to ask, if there is a possibility to work without the gui (i.e. via command line interface, )? I intend to do some scripted/automated measurements based on python and octave.

something like:
#1 define measurement parameters (start, stop, etc.)
#2 call fra4picoscope module, sweep and return array or save .csv dataset directly
#3 ...postprocessing

Any ideas? Thakns a lot in advance!
nano
nano
Newbie
 
Posts: 0
Joined: Wed Aug 17, 2016 7:00 pm

Re: Frequency Response Analyzer with Bode Plots

Postby hexamer » Mon Aug 22, 2016 2:55 am

nano wrote:dear hexamer,

nice piece of software!!! Works flawlessly with my 2205a. :D

Wanted to ask, if there is a possibility to work without the gui (i.e. via command line interface, )? I intend to do some scripted/automated measurements based on python and octave.

something like:
#1 define measurement parameters (start, stop, etc.)
#2 call fra4picoscope module, sweep and return array or save .csv dataset directly
#3 ...postprocessing

Any ideas? Thakns a lot in advance!
nano


Nano,

Thank you for the kind comments.

If I'm understanding you correctly, your asking for something like what was discussed in this post: topic14311-75.html#p55541

If there were a Dll, I assume it would be accessible from Python, then you could do pretty much whatever you wanted to setup/control the FRA, gather data, then post-process.

Please let me know if this is not what you're thinking.

Thanks,

Aaron.
hexamer
Advanced User
Advanced User
 
Posts: 62
Joined: Tue Aug 12, 2014 10:09 pm

Re: Frequency Response Analyzer with Bode Plots

Postby nano » Mon Aug 22, 2016 7:23 pm

Hello!

yes exactly, i intend to do automated measurements -scripted via python or octave. Most of my lab-equipment is already controlled via python, therefore python is preferred. It's a personal project for my little home-lab - nothing commercial.

That DLL-Wrapper Option seems the best solution for me. Do you intend to implement such a tool some day?

Otherwise I'll have to recode the whole DFT-extraction, which seems to be "not so straight forward", concerning some details.

I appreciate your advice!
nano
Newbie
 
Posts: 0
Joined: Wed Aug 17, 2016 7:00 pm

Re: Frequency Response Analyzer with Bode Plots

Postby hexamer » Mon Aug 22, 2016 8:44 pm

I think a Dll is realistic based on the way the program is designed. It is pretty modular, but I'm sure some changes would be required since I've not been building it that way - probably a little more complicated than just repackaging. I'll try some quick-n-dirty repackaging just to see how close it is. If it's pretty close I can include a sub project to build a Dll. But I probably wouldn't be able to put much effort into testing it in the near term.
hexamer
Advanced User
Advanced User
 
Posts: 62
Joined: Tue Aug 12, 2014 10:09 pm

Re: Frequency Response Analyzer with Bode Plots

Postby hexamer » Tue Aug 23, 2016 2:27 am

I tried a few things out and found that modularity is probably not going to be much of an issue. Isolating and building things like the PicoScopeFRA, ScopeSelector, and the Scope implementation classes (e.g. ps2000aImpl) into a Dll is easily achievable with minor changes.

Initially I was thinking of exporting whole classes, which would give a Dll client access to the most functionality. However after encountering some trouble exporting some aspects of the classes and then researching some more, I realized that such an approach would likely only be usable from a C++ client. I could use something like SWIG to provide a more universal interface, but I'm not sure I want to climb that learning curve, and it would probably be useful to build a custom C wrapper around those classes anyway to simplify operations. I'm pretty sure a C interface would be accessible to must programming languages. That's what I'll investigate next.
hexamer
Advanced User
Advanced User
 
Posts: 62
Joined: Tue Aug 12, 2014 10:09 pm

Re: Frequency Response Analyzer with Bode Plots

Postby nano » Tue Aug 23, 2016 7:56 pm

Thanks a lot! I am looking forward to your future findings! :D
nano
Newbie
 
Posts: 0
Joined: Wed Aug 17, 2016 7:00 pm

Previous

Return to Applications

Who is online

Users browsing this forum: No registered users and 0 guests