Demonstrating Solver32's Eye Diagram feature with the LiteVNA
Вставка
- Опубліковано 7 лют 2025
- In this video I go over how to use the new eye diagram feature built into Solver32 with the LiteVNA. I pull out my vintage LeCroy 7200 to verify the results.
For more details on creating eye diagrams on a VNA refer to the following:
www.eevblog.co...
A few papers
www.salukitec....
scdn.rohde-sch...
www.signalinte...
teledynelecroy....
Links to a few professional videos
Rohde Schwarz, Understanding Signal Integrity
• Understanding Signal I...
Anritsu, Signal Integrity testing with the ShockLine MS46524B VNA
• Signal Integrity testi...
Keysight, USB Type C Cable Compliance Test with Keysight VNA & Enhanced TDR Application
• USB Type C Cable Compl...
Keysight, Signal Integrity at 400G and Beyond
• Signal Integrity at 40...
Hi Joe, I really enjoyed this video. A few years ago is was very involved in this industry.
May I recommend two more resources in addition to the commercial resources you mentioned in the first part of the video?
First, Packet Micro in San Jose has done quite a bit is interesting work on eye diagrams and associated topics in collaboration with EMStat.
Second AtaiTec, also in San Jose, has done pioneering work in the field of fixture de-embedding,
Keep up the great work, Joe!
I spent some time searching UA-cam for similar videos and was surprised how little information I found. Even the major manufactures offered very little as far as demonstrations. Feel free to post what ever info you like. I have YT set to block links but if you like, I can manually approve them. Another option would be to use EEVBLOG. I've posted a few additional details about this software as I continue to improve it prior to the release.
@@joesmith-je3tq Hi Joe, up to two three years ago most this kind of educational content happened at traditional conferences (DesignCon, EDICon), digital/print trade journals (Signal Integrity Journal) or company webinars. This may have changed, I don’t know.
Maybe you can partner with all-known personalities such as Eric Bogatin, Heidi Barnes, Al Neves, Steve Sandler, Ching-Chao Huang Richard Zai or Don DeGroot. I don’t believe any of them has a UA-cam channel.
@@christopherscholz Once these low cost VNAs became popular, I had several people approach me about making my software public. My plan is to keep the videos and software available free of charge for those interested. I don't run ads and do not solicit members. I have no patreon account and don't accept donations. I'm not sure what a partnership would offer them or myself. The software is a fairly small project and doesn't require a lot of resources. The VNAs I support are in the $50-$150 and while they are a great tool for education, they certainly have their limitations. I imagine the people you mentioned would have access to the top-of-the-line equipment for any presentations they do. I did try to get one of the PCB layout channels that had ADI to do a demo using the low cost VNAs to verify the results from their simulation. I thought that may make for an interesting video but I don't believe they ever did anything with it.
Wouldn't it be faster if instead of IFFT and convolution you did FFT of the PRBS (needs to be done only once) then .* it with your S's and then do the IFFT. I.e. no computationally expensive convolutions: .* is linear, IFFT is N*log(N). BTW, thanks for your videos! Really appreciated!
[EDIT:] Just realized: convolution is for a single symbol only (pulse response). Although the response will likely overlap consecutive symbols - it will probably be much shorter than the entire PRBS sequence (so no point doing what I proposed above).
The biggest problem I ran into was how to do the mask compare fast. I think I bought my PC in 2016. It's old. Replacing it would certainly improve things. Another would be to change over to the 64-bit LabView but that's not going to happen after NI changed their licensing.
@@joesmith-je3tq People simply assume that the mask is square, then lay the trace outside of it as fast as possible (i.e. to the left and to the right of the mask's time interval) and only when inside the mask's time interval do the checks.
👍🏼
Any chance using a GPU would help with processing?
I've written code in Cuda and called the DLL with LabView before. That part is certainly doable however it would make my software less generic. I don't want to force users to a certain brand and minimum required video card.
Which site can i get uni t ut139c diagram.