A useful addition to your toolkit

I have a clamp-on ammeter that pretty much does that, although it just indicates amps, and doesn't allow waveform snooping. 60 Hz waveforms aren't terribly interesting.

My real problem with current measurement is DC, on PC boards. We want to know how much current, say, an FPGA is using. Sometimes I include current shunts in a layout, but sometimes I don't.

One can use existing switcher inductors as current shunts. I wish I had a PCB trace current probe, but that's probably not posssible. You can measure millivolt and microvolt drops across traces and vias.

Indeed they're not. But your meter is presumably *only* designed for use at 60Hz, I would imagine. Hook it up to a 100Hz signal and you'll see nothing at all in all probability. ;-)

Do they even exist? That would be amazing but no doubt *way* beyond what I can justify to splash out on as a mere hobbyist.

A 1" long, 20 mil wide 1oz trace will be about 25 milliohms. 1 amp makes 25 millivolts, and lots of cheapish DVMs will resolve that well enough. You probably need a bench DVM with microvolt resolution to measure, say, 1 amp running through a via, but you could build a little microvolt meter or amp pretty easily. PCB trace and via resistances need to be calibrated, which is only a minor nuisance.

Here are some pcb-trace shunts, down near the connector:

One of the great mysteries of electronics is "where is the current going?" Sometimes a thermal imager helps figure that out. A little magnetometer would be fun, not hard to do these days.

Tell that to a condition monitoring engineer!

Reading the blurb on the cited current transformer reminded me of a puzzle for all clamp meters... Surely they are no good clamped on a normal bi-directional power cable in which the current is both coming and going: the magnetic fields will near-perfectly cancel out. Surely you have to split the cable to measure current?

Mike.

They do make special current probes specifically for doing traces on a PC board. without cutting traces or requiring a loop of wire.

Looks expensive, and I'd guess not very accurate.

It is expensive. At $795 for the basic probe. Data sheet: Instruction manual:

The expense we can handle, if, for example, it'll give us adequate help in debugging our switching power supply designs. But the calibrating scheme for trace current measurements looks iffy. But maybe along with other measurements, power in, voltage and power out, etc., it could do the job. It is fast enough to look at inductor currents.

I generally use a ferrite (pi filter) on the supplies to uCs and DSPs for EMI, then substitute a shunt to measure the power during test.

The old HP current probe was a great debugging tool.

I've got a plastic box, AC in/out and a double banana jack across an R in the neutral line. I've got an ohm in there (solder in 0.1 ohm if needed) and measure Vrms with DMM

For an outer layer you don't really know the thickness... but I think for an inner layer the thickness is more consistent. (typically 1/2 oz?) So maybe a better current shunt... George H.

I don't design AC power supplies any more. I can buy an entire nice PFC switcher for less than I could buy a 60 Hz power transformer. Since the supplies all come with UL/CE/VDE/etc stickers, I don't care about their input current waveforms.

Seems to me it would be a *great* tool for finding where the currents are going. I thought that would be a good thing?

It is. But some people have difficulty accepting anything more complicated than a light bulb, or more expensive than a pack of cigarettes.

You mean like thermal cameras?

If it's a kelvin connection to the pcb trace then you could neatly disregard the trace resistance by injecting a servo current to bring the IR drop to zero.

piglet

piglet

For the *real* cheapskates out there, there's this alternative:

Claims to be able to handle up to 100A!

Huh, push current back through the voltage leads of the Kelvin connection. I drew that but now I've got all the current flowing through the voltage leads... seems like I need yet another pair of leads.. A six wire "Kelvin" connection. Or I didn't draw it right.

George H.