Power meter

Please let us know how that works out for you.

John

For sine waves, sure. But in the more general case... Getting a flat response from near-AC (say, 10Hz) to something above 10MHz out of a transformer is the hard part; if you just wind your own CT or Rogowski coil you won't just be able to multiply voltage and current to get an instantaneous power -- it becomes rather messier.

The way it's normally done is to buy a proper current probe, as they're designed to have reasonably flat responses -- Tek more or less invented the DC-100MHz+ current probe, and all the major scope manufacturers these days will sell you one.

Most all modern DSOs can be set to do the math to compute instantaneous and average power; some have the option of various additional software tools to look at harmonics, plot histograms of cycle-by-cycle power differences, etc.

I think the reason no one sells a dedicated meter at "inbetween" frequencies is because there are so many different variations on just exactly what it is you'd like to call "power" that you'd end up with something nearly indistinguishable from a scope anyway.

---Joel

But even that's not too hard, since CTs and amprobes are somewhere between easily made (a single CT good from maybe 1kHz to 10MHz isn't too hard) and expensive but achievable (we have a PEM CWT Rogowski probe, -3dB at 3.5Hz and 16MHz, 600A full scale).

That works, except for phase shift.

Suppose we measure current with the Rogowski, and voltage with a 100x scope probe, set the scope to multiply and average. Average says, let's say, 4kW. Try again with a differential probe (which presumably has some amplifiers inside it, as does the Rogowski), now it reads 5kW. Which one is right?

The power meter on the line says about 5kW (I'd be inclined to trust it, because it's more or less line frequency, and comes with cal certs on the very probes used), but the power supply isn't 100% efficient. A tiny phase shift error on a highly reactive load (~90 degree phase shift) is a huge difference in power measurement, even at low frequencies (~10kHz).

Neither probe we have is even specified in propagation delay -- only way to approximate that is by using the same device on both channels, on the assumption that they are within some manufacturing tolerance. Too bad you can't measure current very precisely with a differential voltage probe...

Well, the Fluke 43 / 430 isn't exactly indistinguishable from a scope -- it has a scope function, but it's very limited by the line frequency nature of its purpose. I want the same thing, it doesn't even need a scope, just power stats. It could have a high bandwidth, slow sampling ADC, good enough to calculate RMS voltage, current, and average and peak power (peak +/- or V*I gets VA total, and thus PF, phase shift and Q where applicable). A fast ADC could be set up as a simple scope to read the waveforms and do some other fancy work, like FFT, harmonics, DPF and so on. Nothing difficult about this, there are probably FPGA + ADC dev boards with orders of magnitude more than enough power, just add the probe interface.

Tim

Which is an argument I just recently had with one of my bosses at work.

For what ever reason, one of my German 100:1 2k scope probe failed, just not getting any readings. Most likely one of many I allowed to use my probes, used it as a tether to stop a falling attached device.

But in any case, this was a 250mhz probe and I requested a replacement which wasn't a problem, then I requested a pair which was a problem in his eyes. It wasn't the money, it was the reasoning which It seems that I didn't get across to him. I've got the replacement probes coming any way, even with a little protest from him, I just pushed a little higher in the food chain.

Jamie

Yes, good point -- you do need to dial in the propagation delays for the various channels.

Probably neither. :-)

Feed a step into both probes simultaneously and measure their prop. delay?

Note that if you are using a differential probe, a non-negligible amount of effort by the manufacturer goes into matching the two probes, particularly for any probe/amp assembly that has decent CMRR specs; most of the manufacturers have app notes on how you go about matching the probes to one another if you ever have to replace one.

All depends on how much voltage you can afford to lose in the shunt resistor?

Something like a LeCroy DA1855A is popular for these applications -- it's not cheap, but it has better CMRR than any of the smaller "handheld" diff probes they offer.

...and all the software. I agree there's nothing fundamentally cutting-edge in building something like this, but there are a lot of devils in the details that tend to make it take awhile to come to fruition.

If you're going to have this thing be triggerable, just getting the data interpolation routines working (to shift traces all to the same point in time so you can average or whatever) is not as trivial as it might first appear. :-)

That all being said, if you do build something like this, I think you'd find plenty of other people who'd be interested in buying one.

---Joel

Pretty much-- which has interesting implications for testing. Could put on a SCR-capacitor-discharge-resistor loop and some software to inverse IIR filter the result accordingly.

For sure, the triggered scope part is an entire DSO. Even just V, I and V*I statistics would be valuable though.

Tim

Tim Williams a écrit :

Phase shift is everywhere... Even your power current shunt with a perfect diff probe will be annoying. Today decent scopes (read, certainly not the cheap Chinese) offers the capability to calibrate the signal path *including* the probes.

There are some cheapish ($1200) new DA1855As on Ebay now

ref: 350505799067

Michael Kellett

Cheapish?

Got one for, shame... 35 Euro Guess I was lucky...