How does an RMS detector work?

May 30, 2007 16 Replies

I'm currently delving into audio stuff. I was wondering how exactly does an RMS detector work. I can picture how a peak detector works, but if you send a speech signal to an RMS detector, how does it compute the RMS value?



Also, does attack time dictate the time period for RMS calculations?



It squares the signal (converting it to a unidirectional signal by the transformation), averages that with a low pass filter, and then takes the square root of the low passed signal. RMS means square Root of the Mean of the Square of a signal. Mean is a way of saying average over some time.

The RMS volts is the same as the DC volts that gives the same watts. In fact, the teletronix RMS limiter used an incandescent light bulb as the rms calculator. There is certainly a 'time constant' associated with it... get an old speaker voice coil, hold it between your fingers and run music thru it and turn the volume up and down and it gets hotter and cooler. The 'crest factor' is the peak to RMS ratio. I was surprised to see that the RMS seems to track the avg but 3dB higher for most of the music I've run thru my rms and avg calculator program, so either you can use an RC which gives a good avg and just add 3dB to get the rms, or my program is messed up. Anyone else want to run a wave file thru an rmsser and an averager and see what comes out?

One way is to dump it into a resistive load and measure the temperature rise of the load. By the way, equal dc and rms voltages cause the same temperature rise in a given load.

Also, check this out:

formatting link

"MRW"

** Where did you come across an "rms detector" in audio gear ??

Something made by dBx ??

If so, their use of the term is 100% bollocks.

** A " true rms to DC converter " has a slow response time when used for audio band signals, ie 100mS at least.

Not must use in audio, unless maybe you are trying to compute the heat dissipated in a speaker's voice coil.

....... Phil

A really great and thorough response to your question is in Analog Devices' "RMS-to-DC Application Guide", Second Edition. It's available in .pdf format at this page:

formatting link

The whole thing is over 10MB in .zip file format. Most of your answer will be in secion 1: theory:

formatting link

Take some time, and read it. You'll definitely learn the answer to your question, as well as a lot of other stuff. It's presented in the old Analogue Dialogue method, so even students and newbies can get a lot out of it.

Good luck Chris

page:

formatting link

theory:

formatting link

Wow! Thanks, Chris! I've been looking for something like that. Time to hit the books.

"MRW"

** Shame how it does not answer your original question.

How about you tell us what the heck you are on about.

Is it a " dbx " comp / limiter - or not ?

....... Phil

one way is to digitise and then arithmetically compute the RMS

another is to use a logarythmic amplifier

another is the feed it to a resistor and see how hot it gets.

Bye. Jasen

Correct. Take a MC1496 4quadrant multiplier and build your own RMS meter. Take care for the crest factor!

"Jasen the autistic kiwi Cunthead "

** LOL !!

Love to feed * Jasen the F****it J****ff * a 120 watt soldering iron where the sun don't shine.

...... Phil

Here's an example,

formatting link

jokes on you Phool.

Bye. Jasen

It seems that standard practice is to filter out or somehow exclude the DC value when doing rms measurements. I.E., a DVM or oscilloscope set to "AC" will read Zero when measuring a constant DC signal.

Is there anything like a standard, agreed-upon cutoff frequency used by DVM's and 'scopes for AC/RMS measurements? I'd guess mainly you'd want it well below the lowest audio frequencies of ~20 Hz, so that audio and higher frequencies are included in the measurement.

Mark

"Jasen the autistic kiwi Cunthead "

** ROTFL !!

Love to feed * Jasen the Autistic F****it J****ff * a 120 watt soldering iron where the sun don't shine.

...... Phil

yeah, it makes sense to filter out the DC when doing an AC measurement, and it hard not to filter out the AC when making a DC measurement.

certainly rolloff would want to be below mains frequency.

probably anything that's too high in frequency to read on the DC scale should appear as AC.

ideally the sum of the AC and DC readings should be the RMS of the whole signal (assuming that the DC is advertised as mean and not RMS reading) , I guess that puts the cutoff frequency down below 3Hz somewhere

Bye. Jasen

You don't "Filter out DC". Either you block it with a coupling capacitor which can cause some low frequency roll off, or you use an op amp with an adjustable bias voltage on the other input to remove the DC component.

Some REAL "True RMS Voltmeters" allow you to select AC only, or AC & DC at the push of a button.

Service to my country? Been there, Done that, and I\'ve got my DD214 to prove it. Member of DAV #85. Michael A. Terrell Central Florida

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required