SCR Current measurement 3

[nhcf]

A basic question - the line current at the output of a 3 phase SCR heating controller (phase angle fired) is being measured with a standard digital clamp-on multimeter.

Will the non-sinusoidal waveform cause any measurement errors by the meter? I am thinking the reading should be reasonably accurate.

 

[VEBill]

It depends on the meter and waveform.

Google: fluke true-rms

 

[Comcokid]

Take a True-RMS meter like the Fluke 325 (a random google selection). The data sheet indicates the Fluke accuracy for frequencies above 65 Hz is 2.5%. It is advertised as True RMS, but the datasheet indicates measurement bandwidth is only 5 to 500 Hz. It doesn't state, but usually the frequency at which the bandwidth is listed is -3db down on accuracy, or 0.707.

The Fourier series of a partial half-sine as found in a SCR circuit is going to have harmonics that the meter won't see. But the meter will see the lower harmonics which is where most of the current [power] is at. Since you indicate it is a heating controller, the load is resistive and will not contribute to creating additional harmonics. My guess-itmate is that your readings will probably be off only 5% to 7% max.

 

[jraef]

There is no high frequency involved in phase angle firing of an SCR, the sine wave is partial, not recreated, but is always at the same frequency as the line. A good quality True RMS meter should work fine.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[ScottyUK]

jraef,

I think comcokid is referring to the harmonic components of the distored waveform - the meter will need a reasonable bandwidth to measure the distorted waveform although the higher harmonics do drop away fairly quickly.

I'm not sure how Fluke specify their instruments though - does it state the maximum fundamental frequency at which it can maintain acceptable measurement accuracy for (say) a square wave, or does it state the actual -3dB bandwidth?

 

[djs]

Also is the output AC or DC?

 

[waross]

"3 phase SCR heating controller" Usually AC.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Comcokid]

ScottyUK is correct about what I was trying to state. I design switch-mode power converters for aircraft, and I've seen many engineers make measurement mistakes by trying to use a hand-held meter to measure waveforms with a fundamental or significant harmonics into the tens-of-kHz range without realizing the bandwidth or accuracy limitations of most meters.

 

[nhcf]

thanks for the replies. yes it is an AC waveform.

 

[nhcf]

Assuming the meter/probe is 'true RMS' with a range of 5 - 500 Hz - does this mean it will account for 1st/3rd/5th/7th harmonics (60 Hz waveform), but won't pick up any higher than that? So as long as my waveform is predominantly low-order harmonics, it will provide a reasonably accurate reading? assuming I meet their crest factor specs?

 

[VEBill]

2nd, 4th, 6th, and 8th too. Just in case your waveform contains any even harmonics.

A bandwidth spec is typically given at the -3dB points. So it'll be significantly inaccurate before the defined frequency bandwidth limits.

The waveform you've described should be, almost has to be, predominately low frequency. You could just listen for any higher frequency harmonics, as you'd probably hear something whining if any were of significant amplitude.

 

[ScottyUK]

A quick glance at the calibration manual suggests the response is pretty much flat out to 400HZ, the highest calibration frequency quoted. It would help if the -3dB point was stated, then at least we could understand the capability of the instrument. Of course the response may not follow a classical linear response at all and it could be corrected digitally giving a sharp cut-off frequency above which the readings are garbage. Fluke's manuals are so dumbed down these days that I don't think it will be possible to be sure without a call to Fluke themselves, and then good luck finding someone who can answer!

 

[waross]

If I really needed to know the true RMS I would be considering "rolling my own" RMS ammeter.
Does anyone remember the "Hot Wire" motor starting relays?
With a suitable resistive shunt and a DC calibration the heating effect should give a true RMS indication at any frequency or complex waveform.
Even if this method is never implimented, the suggestion may be used as justification for accepting a less than perfect reading from another instrument.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Comcokid]

I've done that before. Install a shunt resistor in the circuit of a type that can take some heat(you want it to get warm, not super hot). Probe the resistor with a thermocouple and get the temperature. Then put the resistor on a DC power supply and adjust the supply until you get the same heat readings. You want the resistor under identical conditions of ambient and connection (same thermal conductivity environment). Solve P=V^2/R and you have the average power which will equal the rms power in the circuit.

 

[LionelHutz]

Knowing what the reading will be used for could help.

In practice, any decent quality true RMS clamp meter will give an accurate enough reading almost every time.

The testing of utility metering probably requires something better. Offhand, I can't really think of any other reason you'd need more accuracy.

 

[nhcf]

Thanks for the replies. The accuracy discussed for a true RMS meter is sufficient the measurements we are taking.

For me this discussion turned into a good review of fourier series and AC waveform theory.

 

[mudandsnow]

You could try measuring the peak value and calculating the rms. Try it under min, max and medium load to see how accurate your meter is throughout the range of possible phase angles. If there's much error at low phase angle then your calculation is probably wrong.

 

[zappedagain]

Off topic, but this thread takes me back to a test fixture long ago that was drifting all over the place when I hired in. The systems's A/D and a multimeter proved the UUT was drifting all over, so the project manager was arguing with the design engineers (on the other side of the world) about the root of the issue; were the units bad or not? I showed the users what was really going on at the output of their test fixture opamp when I put an oscilloscope on it, an oscillation that the A/D and multimeter were clueless about. Luckily the op amps were socketed, so after replacing them with amps that could handle a capacitive load I was a hero! Ah, the good old days...

Z

 

[Skogsgurra]

Had that in Siemens LOGO! a few years ago.

Some of the 324 opamps could handle capacitive loads and others could not. Siemens never understood what was happening. I think that the fix (extra pull down resistor) is standard now.

Yes, I was a local hero, too.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.