Question about digital clamp ammeter reading technique/philosophy/principle

[gar]

120201-1026 EST

A correction to some statements I have probably made without my going back and rereading them.

I believe I may have related RMS current to power without considering the square factor.

When we modulate the sine wave with the on-off square wave, then if it is a 50% duty cycle the power drops in half and the reading of the average reading meter drops in half. But the RMS reading meter only drops to 0.707 times the continuous reading.

If continuous reads 1 A, then the average reading meter reads 0.5 A, and RMS reads 0.707 . If the duty cycle is 1/3, then the average reads 1/3, and RMS reads about 0.58 .

The check for DC current is an important test to determine whether the circuit is working correctly or not. But not a sufficient test.

ArchieMedes is the RMS meter not reading 1/2 the problem you are encountering with the clamp-on probe?

.

 

[PetrosA]

I received another email today with a slightly different take on the matter that I'll post as well.

Hi Peter,

Here is my take. It is a little different.

Typically the SCR shuts off the low amplitude portion of the supply sine wave. That means the circuit draws current only at higher instantaneous voltage. The resultant waveform looks like a sine wave with the wedges cut off, leaving only the humps. The cut offs are very sharp. That means the harmonic contents have very heavy high frequency components.

This is not a challenge for scopes. They have bandwidth in MHz. It is a big problem for current clamps. If you do not want to work out the correlation as outlined by S.Y., your only other choice is to oversize the clamp. Use as big a clamp as possible. (given the money available and the restriction of the physical space) Say use a 1000A clamp. And, find one that has a better bandwidth say 10KHz. Our U1211A is 1000A, 1KHz and would be our best offer. You can try and find out if it is close enough. 10KHz clamp may be hard to find.

Regards,

T.T.

I have a U1211A and if you're nearby I'd be curious enough to get together and try it, but what's the likelihood of that

 

[ArchieMedes]

120201-1026 EST

A correction to some statements I have probably made without my going back and rereading them.

I believe I may have related RMS current to power without considering the square factor.

When we modulate the sine wave with the on-off square wave, then if it is a 50% duty cycle the power drops in half and the reading of the average reading meter drops in half. But the RMS reading meter only drops to 0.707 times the continuous reading.

If continuous reads 1 A, then the average reading meter reads 0.5 A, and RMS reads 0.707 . If the duty cycle is 1/3, then the average reads 1/3, and RMS reads about 0.58 .

The check for DC current is an important test to determine whether the circuit is working correctly or not. But not a sufficient test.

ArchieMedes is the RMS meter not reading 1/2 the problem you are encountering with the clamp-on probe?

.

For the current reading using RMS meter:

actual is 25% but measured is 42.3%
actual is 50% but measured is 61%
actual is 80% but measured is 88% (looks OK)
actual is 100% measured is 100% (OK)

For the voltage reading using RMS meter:

actual is 25% measured is 42%
actual is 50% but measured is 65%
actual is 80% but measured is 88% (looks OK)
actual is 100% measured is 100% (OK)

I believe because the OFF cycles are more frequent in the lower input range (25% and 50%), the meter is having a hard time or not accurately averaging the current and voltage over time. The higher input range has more ON cycles compared to the OFF cycles so the reading is more accurate in the higher range (80% and 100%).
The challenge really is the meter which can average accurately the current or voltage cycles over time to accurately give the reading.

 

[ArchieMedes]

I got a response from my Agilent contact today which I'll post here in case it helps. My Extech contact has passed things on to their engineers and I should hear something soon



I think Gar may be better qualified to expand on this than I, but from what I understand from this an SCR signal would generally be outside the design limitations of an AC only clamp meter, and the performance of an AC/DC clamp would need to be verified before use with a possible correction factor applied to future measurements. If your equipment is to be sold in number, it may be worth designing test methods for troubleshooting which would include make and model of known "accurate" meters to use.

Hope this helps

THANKS for this. I learned a lot. This is very valuable info. As suggested, I will try TRMS clamp meter with AC/DC clamp which should be capable to detect lower frequency signal. It makes sense that the cycling OFF created by the SCR results in the change of the frequency to a lower frequency which not all clamp meters and clamp probes have capability to measure or detect.

 

[T.M.Haja Sahib]

Before removing the DC component the peak voltage is 10 V, and I calculated the average at 1 V and RMS at 3.16 V.

I need your calculation of the RMS value after the DC component is removed, and your calculation of the DC component.

After removing the DC component the DC average is 0, and the full-wave rectified average is 1.8 V.

After removal of the DC component what is the peak positive value and the peak negative value..

Unless it is sine wave you have in mind,a few more data are required:the rms value of wave form after removing the DC component.The value of DC component is also required. With that data,the rms value of total waveform including DC component can be calculated.So tell me is it sine wave you are talking about in which case the problem is elementary?

The result is all over the map. It depends upon the waveform.

An average/peak reading meter can not measure a non-sinusoidal,perodic wave form accurately.

 

[gar]

120202-0817 EST

T.M.:

I provided you a simple rectangular pulse on which to perform some calculations. This was:

A repetitive positive 10 V pulse (peak is 10 V) that is on for 10% of a period, and is 0 V the remainder of the period. Period being the time of one cycle. The pulse is a steady 10 V during its on time.

.

 

[gar]

120202-0825 EST

ArchieMedes:

What does actual 25%, 50%, 80%, and 100% mean.

I believe 100% means the applied voltage is a continuous sine wave.

Does 80% mean 4 full cycles on and 1 full cycle off? This should produce an RMS reading of 89%. At 60 Hz most meters will easily have a long enough averaging period for this kind of input. Average reading meter will read 80%.

Does 50% mean 1 full cycle on and 1 full cycle off? This should produce an RMS reading of 70.7%. Average reading meter will read 50%.

Does 25% mean 1 full cycle on and 3 full cycles off? This will produce an RMS reading of 50% and an average reading of 25%.

If the on period is longer than 1 full cycle, then you may start to have problems with the averaging time constant in the meter. This will show up as a fluttering in the reading.

I have mocked up a circuit to generate your kind of signal and I tested it with both average reading meters Fluke 27, Simpson 270, and RMS meters Beckman 4410, Simpson 880. My tests were at 50% and 33% duty cycles.

The Simpson 270 is a slightly more accurate 260 and uses a bridge rectifier for AC to DC conversion and a 50 microampere Weston-d'Arsonval type DC meter movement. Means it is a mechanical pointer type meter. This is an average reading type of meter.

The Simpson 880 is a true RMS electrodynamometer type meter. Again a mechanical pointer type of meter.

The Fluke 27 is an electronic digital meter of the averaging reading type. Basically meaning it uses the average DC output of a rectifier.

The Beckman 4410 is an electronic true RMS meter and I do not know their means of RMS measurement.

In the mocked up measurements the DC component was in the volt range compared to a 100 V RMS input when full cycle periods existed for both on and off periods. All readings were of the voltage across the load resistance. Load resistance was 230 ohms. The switch was an IR SP2110. I can't find a good data sheet on this. But it appears to turn on and off at zero crossings. The available specs indicate a maximum turn on and off time of 8.3 milliseconds. This implies control of both on and off at zero crossings. Also, I could get discrete switching from near zero DC to a substantial DC component as I adjusted the triggering timing.

For the lower duty cycles, 50 and 25%, is the on time longer than one full cycle?

Try a Simpson 260 measuring voltage.

.

 

[PetrosA]

THANKS for this. I learned a lot. This is very valuable info. As suggested, I will try TRMS clamp meter with AC/DC clamp which should be capable to detect lower frequency signal. It makes sense that the cycling OFF created by the SCR results in the change of the frequency to a lower frequency which not all clamp meters and clamp probes have capability to measure or detect.

The other opinion suggests that you would need a clamp with high frequency bandwidth. To cover both bases, you would need a clamp capable of measuring current over a large bandwidth. From the specs I had to time to look at last night, the best option would be a good DMM with a clamp attachment rated at least up to 10 kHz. These will typically have BNC terminations for use with oscilloscopes but should come with an adapter that will work with a DMM.

Here are a few I found:

Up to 100kHz:
http://www.fluke.com/fluke/usen/accessories/current-clamps/i1000s.htm?PID=56280

http://www.aemc.com/products/html/moreinfo.asp?id=20205&dbname=products



Up to 50kHz:
http://www.aemc.com/products/html/moreinfo.asp?id=20206&dbname=products



Up to 10kHz:
http://www.home.agilent.com/agilent...T&cc=US&pselect=SR.PM-Search Results.Overview

http://www.fluke.com/Fluke/usen/Accessories/Current-Clamps/i1010.htm?PID=56281

http://www.fluke.com/Fluke/usen/Accessories/Current-Clamps/i800.htm?PID=56324

 

[gar]

120202-0934 EST

PetrosA:

I do not see any need for good high frequency bandwidth.

Nor do I see any need for low frequency capability unless the measurement continues to be done by measuring current. In this case if one wants DC component measurement capability, then an AC/DC probe (probably Hall effect) is needed.

The problem may be in the averaging time constant of the meter, not in the extraction of the raw signal.

.

 

[gar]

120202-0943 EST

ArchieMedes:

Are your on times many contiguous full cycles, or just one for 50% and below? Obviously for percentages above 50 you need to use more than one contiguous full cycle.

Try a Simpson 260. If the averaging time constant is not long enough, then I believe you will see flutter of the needle. Using a sine wave input to the 260 in AC (no coupling capacitor) at 2 Hz the needle swings peak-to-peak about 10%. At 0.2 Hz it is more like a 2.5 to 7 V peak-to-peak swing with the setting having been 5 V at at a higher frequency.

The Fluke 27 reads about 4.5 V from the 5 V input at 2 Hz, about 3 V at 1 Hz, and jumping around above and below 0.5 V at 0.2 Hz.

.

 

[T.M.Haja Sahib]

120202-0817 EST

T.M.:

I provided you a simple rectangular pulse on which to perform some calculations. This was:

A repetitive positive 10 V pulse (peak is 10 V) that is on for 10% of a period, and is 0 V the remainder of the period. Period being the time of one cycle. The pulse is a steady 10 V during its on time.

.

gar
Without providing numerial value for one cycle,how do you expect me to calculate the required rms value?

 

[gar]

120202-1117 EST

T.M.:

You have all the information necessary. But I will be more specific.

A series circuit consisting of:

A 10 V battery.
The positive terminal of the battery is connected to a switch.
The output of the switch connects to a load resistor of R ohms.
And that resistor returns to the negative end of the battery.

The switch is closed for 0.1 seconds, and repeats on a cyclic basis once every 1 second. 10% duty cycle.

What is the RMS voltage across the resistor?

.

 

[ArchieMedes]

120202-0943 EST

ArchieMedes:

Are your on times many contiguous full cycles, or just one for 50% and below? Obviously for percentages above 50 you need to use more than one contiguous full cycle.

Try a Simpson 260. If the averaging time constant is not long enough, then I believe you will see flutter of the needle. Using a sine wave input to the 260 in AC (no coupling capacitor) at 2 Hz the needle swings peak-to-peak about 10%. At 0.2 Hz it is more like a 2.5 to 7 V peak-to-peak swing with the setting having been 5 V at at a higher frequency.

The Fluke 27 reads about 4.5 V from the 5 V input at 2 Hz, about 3 V at 1 Hz, and jumping around above and below 0.5 V at 0.2 Hz.

.

Gar, thanks for bearing with me on this problem. I am attaching the supposed output waveform of the SCR. The bold/heavily outlined portion of the waveform means it is the ON cycle. The lighter portion of the waveform means OFF so you can consider this is flat or 0V (I am not sure if this can be considered DC component or DC part of the waveform).
So a 25% means 1 cycle ON and 3 cycle OFF
50% means 1 cycle ON and 1 cycle OFF
75% means 3 cycle ON and 1 cycle OFF
100% are all cycles are ON

The waveform is not distorted at all because it is zero crossing triggered. There must be some kind of an equipment which can average this ON and OFF cycling accurately. An equipment that considers the OFF cycle while measuring/averaging over time.

 

[gar]

120202-2148 EST

ArchieMedes:

Your waveforms are exactly what I assumed you were describing.

If at 100% duty cycle, then what is the RMS voltage across the load?
At the same time measure the DC voltage across the load. Should be near zero,

At 50% and using a Simpson 260 meter in the AC position (not output) what is the AC reading?
What is the DC voltage reading.

Also do this at 25%. If you do not have a Simpson, then do it with a Fluke 27.

.

 

[T.M.Haja Sahib]

120202-1117 EST

T.M.:

You have all the information necessary. But I will be more specific.

A series circuit consisting of:

A 10 V battery.
The positive terminal of the battery is connected to a switch.
The output of the switch connects to a load resistor of R ohms.
And that resistor returns to the negative end of the battery.

The switch is closed for 0.1 seconds, and repeats on a cyclic basis once every 1 second. 10% duty cycle.

What is the RMS voltage across the resistor?

.

3.16V

 

[ArchieMedes]

120202-2148 EST

ArchieMedes:

Your waveforms are exactly what I assumed you were describing.

If at 100% duty cycle, then what is the RMS voltage across the load?
At the same time measure the DC voltage across the load. Should be near zero,

At 50% and using a Simpson 260 meter in the AC position (not output) what is the AC reading?
What is the DC voltage reading.

Also do this at 25%. If you do not have a Simpson, then do it with a Fluke 27.

.

I dont have a Simpson. I will do that with a fluke. I would just like to know it the meter will not be damage if I use the DC voltage setting while measuring AC voltage. If it will not damage the meter, should I put it on a higher voltage selector setting (higher than the actual AV voltage value of the load)?

 

[gar]

120203-0752 EST

ArchieMedes:

Applying an AC input to a DC meter will not harm the meter on any range that it does not cause the needle to pin, shake too much, or for mechanical or electronic meters some internal voltage or power dissipation limit.

Thus, if your source voltage is 120 V RMS sine wave, then a Simpson 260 on DC and the 250 V range will not be harmed. If the AC was full-wave rectified the reading would be about 108 V and no harm. If half-wave rectified 54 V, and this would pin the Simpson on the 50 V range, but not by much and would cause no harm.

On a Fluke digital the input impedance is always constant and thus, no power dissipation or needle pining problem. No needle problem because there is no needle. Most Flukes have auto-ranging, and therefore it may hunt. Thus, set it to a fixed range. Unfortunately when I switch my Fluke 27 between AC and DC it seems to go back to auto-ranging. The Simpson 260 doesn't have this problem.

Assume there is no DC component, then the theoretical readings for different integral half cycle periods for the off time are:
1 on 1.5 off is 63.2%
1 on 3.5 off is 47.1%
1 on 4.0 off is 44.7%
1 on 4.5 off is 42.6%

Note: Assuming no DC component, then your
42.0, 42.3% look like 1 on and 4.5 off, and
61, 65% look close to 1 on and 1.5 off.

Using an AC average reading meter instead of an RMS may provide a sharper differentiation of the different signals.

.

 

[gar]

120203-0924 EST

T.M.:

That value you could get from post #40. What is a more accurate value?

Then what are the answers to the other questions I ask in that post for which I did not give answers?

.

 

[Besoeker]

For the current reading using RMS meter:

actual is 25% but measured is 42.3%
actual is 50% but measured is 61%
actual is 80% but measured is 88% (looks OK)
actual is 100% measured is 100% (OK)

For the voltage reading using RMS meter:

actual is 25% measured is 42%
actual is 50% but measured is 65%
actual is 80% but measured is 88% (looks OK)
actual is 100% measured is 100% (OK)

I believe because the OFF cycles are more frequent in the lower input range (25% and 50%), the meter is having a hard time or not accurately averaging the current and voltage over time. The higher input range has more ON cycles compared to the OFF cycles so the reading is more accurate in the higher range (80% and 100%).
The challenge really is the meter which can average accurately the current or voltage cycles over time to accurately give the reading.

Latter end of last year we manufactured a few (8 or so in the 150kW-250kW range) burst firing systems for heater controls. I didn't particularly concern myself with power measurements since it was closed loop and the feedback was from temperature.

But here are a couple of pics of some calcs I did this afternoon.
I used 120V and 100A so that they could be plotted on the same axis. Obviously, fully on, that gives 12kW for a resistive load.



At 50% duty cycle:



The average power is then 6kW.

Vrms is 84.8V (120/.sqrt(2))
Irms is 70.7A for the same reason.

And 84.8 * 70.7 gives 6kW

I can't tell why your voltmeter and ammeter read what they read.
What I can say is that a 50% duty cycle doesn't result in 50% Vrms or 50% Irms and you shouldn't expect meter readings to show that.

 

[ArchieMedes]

120201-1026 EST

A correction to some statements I have probably made without my going back and rereading them.

I believe I may have related RMS current to power without considering the square factor.

When we modulate the sine wave with the on-off square wave, then if it is a 50% duty cycle the power drops in half and the reading of the average reading meter drops in half. But the RMS reading meter only drops to 0.707 times the continuous reading.

If continuous reads 1 A, then the average reading meter reads 0.5 A, and RMS reads 0.707 . If the duty cycle is 1/3, then the average reads 1/3, and RMS reads about 0.58 .

The check for DC current is an important test to determine whether the circuit is working correctly or not. But not a sufficient test.

ArchieMedes is the RMS meter not reading 1/2 the problem you are encountering with the clamp-on probe?

.

Gar,
With reference to the above highlighted text, how come the RMS reads higher than the average meter? What is the theory behind?