Electronics > Projects, Designs, and Technical Stuff

Double Pulse testing project

(1/1)

Glenn0010:
Hi All,

At work we do a lot of pulse testing for IGBTs and want to come up with a rig that does the pulse testing automatically and process all the data automatically.

Ideally on the test side, you would enter desired pulse current and it would run the test until that current is reached both for turn on and turn off.

On the processing side, Ideally it would de-skew the current measurement for the delay of the rogowskis', remove any offset for the rogowski and process all the data. I would also to include Vce sat and Vf measurements in there as well. Processing the data takes quite a lot of time so fully automating it would be great.

The ,first bit is relativity easy to do with a micro controller however getting all the data is a bit trickier.

The easiest way of getting the data would  be to connect the scope to excel somehow and write vba scrips to process the data. The problem that I see with this is that we have some old scopes with really good specs, however I dont think they have the functionality to connect directly and stream data to a PC.

Another way I think may work is to find some high sample rate ADC boards with memory on them so that data can be captured on them and then streamed to a PC at will since I would require a lot of channels. Does anyone have any idea were to start looking in this area?

Cheers!

jbb:
I guess the thing to do is see what interfaces you’ve got on the existing scopes.

Current pulse control shouldn’t be too tricky - I guess you already have something in place. Do you also want to sweep DC link voltage and IGBT temperature?

In terms of Vce,sat measurements you have to contend with the old problem of standing off the OFF voltage (eg 1000V) and making accurate measurements of the ON voltage (eg 3V). We often did these measurements separately with a curve tracer where I used to work.

If you need a digitiser, there are some out there which are quite fast and have 12 or 14 or more bits of resolution.

Finally, cast a fresh eye over the safety of the system. When you automate processes, you might accidentally open up gaps in your safety envelope. Remember that less skilled people may use the automated version...

Glenn0010:

--- Quote from: jbb on December 22, 2019, 07:24:41 am ---I guess the thing to do is see what interfaces you’ve got on the existing scopes.

Current pulse control shouldn’t be too tricky - I guess you already have something in place. Do you also want to sweep DC link voltage and IGBT temperature?

In terms of Vce,sat measurements you have to contend with the old problem of standing off the OFF voltage (eg 1000V) and making accurate measurements of the ON voltage (eg 3V). We often did these measurements separately with a curve tracer where I used to work.

If you need a digitiser, there are some out there which are quite fast and have 12 or 14 or more bits of resolution.

Finally, cast a fresh eye over the safety of the system. When you automate processes, you might accidentally open up gaps in your safety envelope. Remember that less skilled people may use the automated version...

--- End quote ---

Currently control is quite crude, it's just increasing the on time till the desired current is reached. We usually test for worst case and nominal to limit the number of tests, however if we get a faster system, we can afford to sweep through more parameters.

Currently what I am doing is using 2 scopes, one for measurement of voltage and current and the other for Vce, Vf measurement, where I am clipping the high voltage to around 5V which then allows me to take an accurate reading so I'd need even more channels. I was wondering if there are some existing ADC dev boards with onboard memory which can do the job then I can send data serially to the PC.

With regards, to safety, it's definitely a good point. We have quite good processes at work which limit the tests less skilled people can do so this shouldn't be a huge issue.

T3sl4co1l:
BTDT, on a smaller scale:



Schematic: https://www.seventransistorlabs.com/Images/Diode_Recovery_Tester.pdf

I've made minor changes since; most importantly are, the gate driver is enhanced with an emitter follower (PBSS303NX/PX pair) and 2.2R instead of a ferrite bead.  The IGBT is whatever; smaller devices are desirable for faster switching, bigger devices are desirable for higher current or power.  (Right now, I've got a STP19NM50N in it.)  Also a 220pF + 22R damper from GD0V to GND.

Of course for testing IGBTs, you'd use the same pulse generator (or perhaps elaborate it a bit for better controls, or timing precision), a higher gate drive voltage with bias (as applicable), and a proper output stage that's not through a terminal block but is built into a board or busbar to fit the DUT, and equipped with dampers, snubbers, bypass, etc. as needed.

Tim

Navigation

[0] Message Index

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod