Measuring inrush current

[Mitchekj]

Hey guys, quick question for the gurus.

The nifty thing I use to measure inrush current is out for cal. So, I want to try to measure it using a current sense resistor and a scope. Trying to measure the inrush on a new 60W off-line flyback design.

Was looking for some 0.1 Ohm, but all I have on hand are 0.01 Ohm resistors. The scope is a digital storage 1GS/s Tektronics job.

I get what equates to about 150A when on the ~nanosec time scale, and about 20A when on the 1ms scale. I know 20A sounds like a more plausible reading, but I want to make sure I'm doing this right. Any input?

Thanks!

 

[(*steve*)]

I suspect the inrush current will vary depending on where in the half-cycle the thing is powered up -- or is it running from DC?

For measuring these things, I've used a clamp AC ammeter with a peak hold function.

Not sure how accurate this is though.

 

[Mitchekj]

Good point. Hadn't thought of that. I'd assume I would need to measure at one of the peaks to get a true maximum inrush.

Thinking on it some more, the max inrush would "theoretically" be infinite, yeah? Since the input caps are a direct short to the incoming AC at the exact moment of turn-on when they're completely discharged. There has to be some time constraint there, since the energy involved at this "infinite" current would be small, else we'd theoretically detonate the universe or something.

Bah, now I'm dabbling into this theory stuff, hehe. Anyhow, in the time it would take to source parts for and build some type of peak detection and timing circuit, it'd probably be best to wait the day or so until my stuff makes it back from the cal lab. Or... do I just run 20 or so iterations of my test and take the highest value? Close enough? Just how do they choose the exact moment in time to take this inrush measurement? "Infinite" to "ok, now it's the true value." I must be missing some physics.

 

[(*steve*)]

Theoretically, maybe.

But practically, capacitors have an ESR greater than zero, and that will limit their charge rate. And there are other factors to consider too.

How you measure it depends a great deal on what your measuring device measures And on what the circuit you're measuring looks like. I really don't have enough experience to guide you here.

 

[Resqueline]

What steve says.
150A peak doesn't surprise me, but on nanosecond scale you have to be very careful about the physical measuring setup to avoid inductance messing up the result.
320V / 150A = 2.1 ohms, or 160V / 150A = 1.1 ohms, which seems a ballpark figure for the main cap' ESR.
This extreme (short) peak is the reason NTC's are added as limiting elements.

 

[Mitchekj]

Thanks guys. Very good points! Had forgotten to take ESR into the picture. An ohm or two sound about right, but not sure what the guy spec'd. He doesn't work at our shop. (The Y caps, if I recall the schematic correctly, are the first caps in the picture, from line/neut to pwr gnd.) For the life of me, I can't remember if there's a thermistor in there.

In any case, meters will be back Monday. Interested to see what it reads.

 

[55pilot]

It is a dead issue since your meter should be back now,but I will throw in a few observations.

When measuring things with a scope at high frequencies or low amplitudes, the single biggest source of problems is the ground clip. Its primary hob is to pick up noise and distort the signal. If you really want to measure the type of things you are measuring, you need to properly ground the probe.

Many high end probes come with tiny spring clips that connect to the outside of the probe (after you remove the plastic cover). You end up with a ground pin sticking out about 3mm away from the signal pin. You can fabricate one yourself if you are good at mechanical work.

On the boards I design, I always have an exposed ground pad 2.5mm away from the critical signal test points I need to scope. Whenever I can get away with it, the ground is actually a 0.8mm through hole pad, which makes handling the probe much easier.

---55p

 

[(*steve*)]

On the boards I design, I always have an exposed ground pad 2.5mm away from the critical signal test points I need to scope. Whenever I can get away with it, the ground is actually a 0.8mm through hole pad, which makes handling the probe much easier.

A very good tip.