So imagine a pcb trace carrying the unknown load current. Place four test probes in a line along that trace. Measure the voltage between the inner pair and inject a current through the outer pair such that the voltage measured is reduced to zero. The current you have to inject will be the same magnitude but opposite direction to the unknown load current.
piglet
AAAAHHHHHH!!!!!! What a perfect way to destroy electronics! Forcing current into a print, where the current to be measured, can change without prior notice, and your current is going the other way......
No with six connections I think it's OK. If the input current changes, then the voltage across the trace changes, and the external circuit compensates. It's a lot of work to measure a current. But a new idea for me, Thanks piglet.
George H.
And how do you measure the injected current?
"You can't fool me! It's turtles all the way down."
It's not hard to figure where current is going: just measure voltage drops. What's sometimes difficult is quantifying it.
Just now we're laying out a 10-layer board with two ground planes and three power planes. There are 22 power supplies. Most of the power distribution will be interestingly-shaped interleaved pours, not traces, on various layers.
That magnetic gadget would be hopelessly confused. Multiple currents and various return paths would make it useless.
My favorite tool for tracing unusual current flows is my Flir E45 thermal imager. It cost $12,000.
We just demoed a cool new thermal imager, mounted on a nice little stand, with its own display and also USB interfaced for pics or movies. They're going to let us keep it.
Let's see you use it to find a stray 100 mA DC on a crowded circuit board. :(
Names, please. And results, later.
You don't need to cancel the trace current--just injecting a calibrated a.c. current between the two probes would be good enough to measure the trace resistance, then compute the d.c. current from the d.c. drop.
Cheers, James Arthur
We don't believe it until there are pictures.
You can add a current sense resistor to that current... as large a turtle as you choose. The idea is not not disrupt the measured current with a resistor. Heck, this could be done with an external device like a meter. You just need four probe points which wouldn't be too hard with a good design. It would need to be sharp enough to pierce the solder mask easily. The only danger to the electronics would be if you probed two different traces... lol
Maybe that should be one probe with four sharp contacts.
I'm not sure I can name names yet. It's still in development.
We've become sort of a beta tester for these people. They send us units and get our feedback, and we can keep them. Or maybe they just like coming to San Francisco.
Just power the board down and measure the trace resistance!
Another trick is to measure the drop in a trace, add a dummy load resistor, note the delta-v, and do the math.
Or measure the trace resistance on a bare board from the same batch.
Sure, but we're talking an automated instrument do that all in one go, without having to make a separate 4-wire resistance measurement.
I like the magnetic probe for qualitative info. Maybe a Hall probe could do it too.
We need a FLIR for magnetic flux.
Cheers, James Arthur
It would be cool to see the mag fields around a board.
OK next you'll be synchronously detecting the AC. :^) George H.
Of course, that's the reason for using a.c.!
Cheers, James Arthur
Hello James,
That would called an EMSCAN.
- Wolfgang.
AWESOME.
Cheers, James Arthur
That's easy. Just sprinkle it with iron fillings. ;->
Oh come along now. You know it's not intended for that sort of usage!
Have something to add? Share your thoughts — no account required.
Ask the community — no account required