DIY / CHeap Rogowski Coil

[Glenn0010]

Hi,

I'm owrking on an automated double pulse tester, where you set the desired pulse test current and through a current sensor, a uC sorts out the pulses automatically.

I've almost finished the code and now have to start looking at which sensor to use. Ideally it would be a rogowski since it has the least delay. Does anyone know of any cheap rgowski or even a decent DIY method? I'm open to other suggestions as well.

Cheers

[Gyro]

This thread looks as if it could be helpful as it mentions both double pulse detection and Rogowski Coils... https://www.eevblog.com/forum/projects/double-pulse-testing-project/.

A forum search for 'Rogowski' pulls up quite a lot of hits - I remember at least one person trying to build one.

[MagicSmoker]

Making a Rogowski coil is fairly easy; making the output from one usable is relatively difficult. This is because the raw output of a Rogowski coil is proportional to the rate of change of flux through its solenoid coil, so it has to be run through an integrator to produce an output proportional to current that is flat with frequency. Unfortunately, making an integrator that covers a wide frequency range yet exhibits low drift and offset is incredibly difficult. Not impossible, mind you, but much harder than you might expect at first glance (invoking the "How Hard Can It Be" curse here).

Rogowski probes are commonly used for double-pulse testing, but with the caveat that the results should be considered more of relative value, than absolute. The gold standard for high precision, wide-bandwidth current measurement is a Pearson current probe, however. You'll want to be sitting down when check the pricing on them, though they do turn up on eBay from time to time.

[Glenn0010]

Luckily we have a few pearsons at work which I can borrow to develop it. And yes Rogowskis can be a bit finniky with their accuracy and offsets.

Oh and yeah, I didn't relize the pearsons were so deer, jesus!

[MagicSmoker]

Alright, attached is a paper on PCB Rogowski probe designs that I found useful enough to save. I could've sworn I had a few more, but this was at least a couple desktop computers ago so things apparently got lost along the way.

[tautech]

The gold standard for high precision, wide-bandwidth current measurement is a Pearson current probe, however. You'll want to be sitting down when check the pricing on them, though they do turn up on eBay from time to time.

These:
https://www.pearsonelectronics.com/products/clamp-on-current-monitors
Thanks, never heard of them before.

[T3sl4co1l]

Code, what's that?
https://www.seventransistorlabs.com/Images/Diode_Recovery_Tester.pdf

If you're testing IGBT modules, a Rogowski is a good idea; PEM makes good ones.  And obviously you'll need more drive, an isolated +/-15V some-amperes driver would be good and they're available off the shelf as well.

Under say a hundred amperes or so, shunt resistors are fine.  Use SMT chips, several in parallel.  Make a wide connection to minimize loop inductance.  Remaining ESL can be calibrated out with an R+C lowpass filter on the sense port.

I suppose if you need to do automated testing, some code would be called for, but heck, even if you're a manufacturer testing your entire product line, you'll spend way more time on constructing jigs and placing parts into them, than actually running the test.  Set a technician to do it a few days or weeks, and there's your data.

Tim

[Glenn0010]

This is more of a stepping stone project, I had made a motor controller using and ARM7 chip, so this will be an introduction to the new-er STM 32 series.

As for current sensing I'd like a rogwski or something of the sort as its no invasive, thats about it really. Though shunts can definetly work.

So the intial steps of the project would be where you can tell it to do a pulse test at 20A through the pc and it handles thepulse legths automatically (can be a bit tedious). Further more it would be quite good to synchrnize short circuit tests, mainly type 2, where you can tell it to do a short upone reaching 20A current  (ex close a contactor).

In a perfect world, I'd also get the pc to talk to a scope (is it possible? GPIB? need to look into this) so set the scope automatically and run a series of current points, export the data and process it automatically to where you get something like 20A on = 3 mJ for example. But this is a bit far off at the moment.

Any feedback would be appreciated!

Cheers