Strip lights go out when I turn on the saw

[Nozzle]

You guys have taken all the words out of my mouth ref the difference between *true* RMS and my boggo standard cheapie meter. The relationship between peek volts and RMS volts for a true sine wave is squareroot of 2 i.e. x0.707... but mains power is anything like a good sine wave! That's another story..

I believe the RC network in the meter, to give an acurate "RMS" measurement also has to have the correct frequency applied, and it'll be tuned for 50Hz or 60Hz depeneding on which country the meter is sold in. Add into this the rating of these component will drift, and frequency is not always 50.000Hz.. it all adds up to a measurement from my meter that has little to be believed!

I'll get onto the network people and ask them to undertake this survey. I bet it'll take ages to even get them out..

Nozzle

 

[flyingsparks]

I'll get onto the network people and ask them to undertake this survey. I bet it'll take ages to even get them out..

Nozzle

A low voltage problem, I would imagine they will be out the same day..

 

[Spark123]

Anything supplied from the national grid ought to be 50Hz +/- 1% i.e. 49.5Hz to 50.5Hz which isn't going to play up much with a bog standard multimeter.

 

[JohnW2]

You guys have taken all the words out of my mouth ref the difference between *true* RMS and my boggo standard cheapie meter. The relationship between peek volts and RMS volts for a true sine wave is squareroot of 2 i.e. x0.707... but mains power is anything like a good sine wave! That's another story.. I believe the RC network in the meter, to give an acurate "RMS" measurement also has to have the correct frequency applied, and it'll be tuned for 50Hz or 60Hz depeneding on which country the meter is sold in. Add into this the rating of these component will drift, and frequency is not always 50.000Hz.. it all adds up to a measurement from my meter that has little to be believed

I think you're probably being unduly pessimistic in terms of voltage measurements in electrical installations. The waveform of UK electricity supplies will usually be very close to a true sine wave and, as you have been told, will necessarily have a frequency very close to 50Hz (certainly between 49 and 51Hz).

"Bog standard cheapie meters" will generally have a specification indicating that they can measure ('calculated') RMS voltages with an accuracy of ±1% to ±2% over a frequency range of 40Hz up to several hundred Hz. That is more than good enough for nearly all measurements in a domestic electrical installation. 'True RMS' meters will obviously maintain much the same accuracy regardless of waveform but, for waveforms close to true sine wave, the spec of even very expensive ones is not usually that much 'better' than the cheapo ones.

Kind Regards, John

 

[ban-all-sheds]

Anybody got a scope handy (mines not), and a dimmer switch?

Would be interesting to see how far away from a sine wave you have to get before a BSDMM starts giving badly inaccurate results.

 

[JohnW2]

Anybody got a scope handy (mines not), and a dimmer switch? Would be interesting to see how far away from a sine wave you have to get before a BSDMM starts giving badly inaccurate results.

I'm not quite sure how one would do that. How would one determine the true RMS of the chopped waveform in order to determine when/whether the meter was giving 'badly inaccurate results'?

It's also a pretty unfair test, since the sort of waveform one gets from a dimmer is much more extreme than anything one would ever see 'generally' within an electrical installation.

Kind Regards, John

 

[plugwash]

Anybody got a scope handy (mines not), and a dimmer switch? Would be interesting to see how far away from a sine wave you have to get before a BSDMM starts giving badly inaccurate results.

I'm not quite sure how one would do that. How would one determine the true RMS of the chopped waveform in order to determine when/whether the meter was giving 'badly inaccurate results'?

Most modern scopes i've used have an option to calculate the RMS of the displayed waveform.

You would need a suitable probe of course. Don't go probing mains with a regular scope proble.

 

[JohnW2]

I'm not quite sure how one would do that. How would one determine the true RMS of the chopped waveform in order to determine when/whether the meter was giving 'badly inaccurate results'?

Most modern scopes i've used have an option to calculate the RMS of the displayed waveform.

None of mine are that modern or that sophisticated! Indeed, they are all analogue ones which don't calculate anything

However, soon after I wrote the above, I realised that it only takes a little bit of calculus to determine the RMS of of a sine wave 'chopped' at any given point in the cycle by a dimmer. If I use a scope to determine the duration of the 'chop' (as accurately as I can, visually), and assume that the supply is a sine wave, I can therefore calculate the RMS. If/when I have a little spare time, I'll do the experiment.

Kind Regards, John

 

[JohnW2]

Anybody got a scope handy (mines not), and a dimmer switch? Would be interesting to see how far away from a sine wave you have to get before a BSDMM starts giving badly inaccurate results.

Before doing it 'properly' (per my post to plugwash), I thought I'd look at an extreme case of a non-sinuoidal waveform, by using just a power diode and resistor to get rid of one of the half cycles - which obviously would reduce the true RMS (measured over any number of complete cycles) by half. Using the cheapest and nastiest DMM I had to hand, I got 256V (about average for my installation) for the full supply and 144V with the half-wave-rectified version. 146V is obviously a fair way from the correct answer of 128V (12.5% error), but not dramatically wrong. This suggests to me that the meter's readings would probably remain pretty accurate with the very modest deviations from a true sine wave that one would expect to see in a domestic electricity supply. The proper experiment will hopefully confirm in due course.

Kind Regards, John

 

[plugwash]

Before doing it 'properly' (per my post to plugwash), I thought I'd look at an extreme case of a non-sinuoidal waveform, by using just a power diode and resistor to get rid of one of the half cycles - which obviously would reduce the true RMS (measured over any number of complete cycles) by half.

WRONG.

The RMS is the square root of the mean of the square of the signal.

Taking only one half cycle of a sinewave will halve the mean square value and hence reduce the RMS value by a factor of sqrt(2) or about 1.41. So the RMS value of your rectified signal would be about 182V which means your meter significantly underread the voltage.

 

[ban-all-sheds]

However, soon after I wrote the above, I realised that it only takes a little bit of calculus to determine the RMS of of a sine wave 'chopped' at any given point in the cycle by a dimmer. If I use a scope to determine the duration of the 'chop' (as accurately as I can, visually), and assume that the supply is a sine wave, I can therefore calculate the RMS. If/when I have a little spare time, I'll do the experiment.

You could do that.

Or just count the grid squares under the trace, and do a bit of Mk1 eyeball estimation of the bisected ones.

 

[JohnW2]

However, soon after I wrote the above, I realised that it only takes a little bit of calculus to determine the RMS of of a sine wave 'chopped' at any given point in the cycle by a dimmer. If I use a scope to determine the duration of the 'chop' (as accurately as I can, visually), and assume that the supply is a sine wave, I can therefore calculate the RMS. If/when I have a little spare time, I'll do the experiment.

You could do that. ... Or just count the grid squares under the trace, and do a bit of Mk1 eyeball estimation of the bisected ones.

[when you say 'under the trace', I presume you mean 'between the trace and the V=0 line] That would give the mean, not the RMS, wouldn't it?

Kind Regards, John

 

[Spark123]

RMS is pretty pointless for anything other than a proper sine wave, it is basically a meaningful expression of an average for the waveform.
If you half clip and only have the positive half cycle then the average is going to be below that of RMS.

 

[JohnW2]

Before doing it 'properly' (per my post to plugwash), I thought I'd look at an extreme case of a non-sinuoidal waveform, by using just a power diode and resistor to get rid of one of the half cycles - which obviously would reduce the true RMS (measured over any number of complete cycles) by half.

WRONG. ... The RMS is the square root of the mean of the square of the signal. ... Taking only one half cycle of a sinewave will halve the mean square value and hence reduce the RMS value by a factor of sqrt(2) or about 1.41. So the RMS value of your rectified signal would be about 182V which means your meter significantly underread the voltage.

Whoops - so much for haste However, there's a need for far more !!! Having mistyped my supply voltage as 256V (rather than 246V, which is what it actually was), I then did my calculations using that erroneous figure. The true RMS should therefore have been about 174V, such that my 146V represented an under-reading of about 16%

However, I was actually a bit extreme in my test, since, as I said, I used the cheapest and nastiest (and extremely old) meter I had in my drawer. Using a much more modern, but very cheap, one, I get 244V for the supply and 158V for the half-wave rectified voltage, an under-reading of about 9% as compared with the true rectified RMS of about 173V.

Interestingly, another modern, very cheap, one I tried gave identical results for both the supply and the rectified voltage (and a zero reading for the rectifed voltage if I reversed the diode) - so the meter itself is presumably only half-wave rectifying and looking at just one polarity of half-cycles.

Again, interestingly, my Fluke 1652 gives a much lower reading for the rectified voltage (about 131V) - hence much further from the true RMS than any of the cheapo DMMs.

However, we must bear in mind that this is an extreme deviation from a sine wave.

Kind Regards, ohn

 

[JohnW2]

RMS is pretty pointless for anything other than a proper sine wave, it is basically a meaningful expression of an average for the waveform.

I don't think that's really true. We don't use RMS just for the hell of it - and,in some senses,it is even more valuable for non-sinusoidal waveforms (since there is a simple arithmetic relationship between RMS and mean for a sine wave). IIRC, RMS current multiplied by RMS voltage will always give true power, regardless of waveform, whereas the same is not true of any other sort of average (such as the mean).

Kind Regards, John.