Peak or RMS

[Crackle]

What is the official convention for AC voltage figures? If an AC voltage is given as 10v is that 10v Peak?
The course I am taking seems to think so.
"Any exam questions will give you the value as V or Vrms etc."

Thanks
Mike

[ppppenguin]

Always RMS, unless otherwise stated.

[Mark Hennessy]

I'm not sure there is any sort of "official" convention here - so much depends on the context. I'd say that RMS is by far and away the most common - so therefore is the the safest assumption in the absence of any other guidance. But I see so many people making mistakes here that I always advise people to be pedantic about it - so specify whether it's RMS or peak or peak to peak unless it's really clear from the context.

If the context is related to power in any way, then the voltage must be RMS, as calculating watts based on peak or peak to peak will give wrong results. That means that any conversions involving power - such as from dBm to voltage - must be based on RMS quantities.

Most of the time in audio the measurement of interest is RMS, but converting to pk-to-pk and vice-versa is necessary for working out how much signal level you get for a given set of power supply rails, for example. Some program meters measure RMS or average, but others measure peak (most often quasi-peak) because overloading transmitters and other media is something to avoid - but peak levels tell you nothing about how loud an audio signal is, which obviously causes a quite specific sets of challenges in broadcasting.

However, if you're talking about analogue video signals, the RMS is not really meaningful. Peak to peak is the most interesting thing, or even just peak if you're ignoring the sync pulses. It's been a while since I've got my hands dirty with analogue video, so can't remember how APL (average picture level) is calculated, but I think it's a simple average of the values of the active picture area rather than an RMS calculation. I can't see the point of doing RMS for this - RMS is about equivalent DC heating effect*, which is not related to picture content in any way I'm aware of. To keep this thread on track, then a simple correction is welcome if I'm wrong about that, but if anyone wants to expand on it (more than a sentence), then please start a new thread.

* Many here will think I'm stating the utterly obvious - perhaps even being borderline patronising - but I have to cover this with all new recruits, and it's surprising how few know that RMS is the equivalent DC heating effect. <long rant about academia self-edited out!>

Another thing that again might be obvious or worse to some of the folk on here, but my teaching experience tells me it's worth the risk of saying it: the "magic" root-2 that gets you between RMS and peak is only valid for sine waves. Obvious when you understand the "equivalent DC heating effect" point, but well, see my previous paragraph

[peter scott]

Always RMS, unless otherwise stated.

+1

[ppppenguin]

Marks, thanks for the thorough explanation. Especially the point about equivalent heating which is automatically assumed by those who know and easily overlooked by those who have missed that bit of education. Not so much "what is RMS" but "why use RMS at all". Please don't get me going on the nonsense of "RMS power".

Video is indeed a special case where RMS has no useful meaning. Noise is another non-trivial case. Here, the voltage depends on the bandwidth of interest. For video signal to noise ratios there are particular customary measurement methods which won't be of interest to most GVR members. Best read about in Weaver's books on video measurement.

Audio s/n ratios can be another can of worms. I'm sure Mark has studied the various weightings etc with more care than I have.

But coming back to the simple case of a sine wave, unless there is a context to indicate some other measuement, it's RMS. This can fool those who are using a scope and a DVM to monitor the same signal. Why is the meter saying 1V while the scope is saying 2.8V? I think it was Mark who reminded me that some modern digital scopes can calculate the RMS voltage of a waveform, something I found useful a few months ago.

[Mike Watterson]

Also the voltages on valve tester knobs are not what you see on the pins, though might be if you have a true RMS meter, I'd have to re-read how they work to be sure!

Isn't RMS power shorthand for the undistorted (or maybe < 2%) continuous maximum with a sinewave drive on an amp output and on a speaker simply the DC equivalent power that won't burn it out? Perhaps more useful than PMPO on an amp or Peak power on a speaker, both of which are totally devoid of any meaning. I agree, if Watts meant real Watts, the prefix of RMS is meaningless.

Maybe Crackle / Mike needs to read the old archived ARRL and RSGB handbooks before some of this modern "courseware". There are online PDFs to download. Physics hasn't changed for Radio & Electronics in the last 100 years and I find the old valve book explanations of the three main RF oscillators (common cathode, common grid and cathode follower amps with feedback) are better than most transistor book explanations. Virtually every current transistor circuit except complementary totem-pole was first documented in valve books. Even op-amps and the balanced multiplier made with valves.

[ppppenguin]

Real component of (RMS voltage * RMS current) = Mean power. "RMS power" is often used as an inaccurate shorthand for mean power.

[Mark Hennessy]

Yes, "Watts RMS" is a marketing term, as "continuous average sine wave power" is a bit of a mouthful

Being charitable, I suppose you could say the term indicates that the quoted value was derived from an RMS voltage measurement. The usual procedure is to wind up the power until the THD measures 1%, record the voltage, then calculate the power based on that and the value of the dummy load used. But some manufacturers might go for 10% THD. Given how abruptly the THD rises as you reach clipping with a typical design with plenty of negative feedback, there isn't really much of a difference between 1% and 10% THD. But for a valve amp, or some other low feedback design, that approach can be quite useful. But with all that in mind, I tend to regard power ratings with no distortion value as incomplete measurements. But power ratings are usually very conservative because power supplies are most often unregulated, so the power you get at 1% or 10% THD will depend on the mains voltage at the time of the test. Plus, manufacturers are very keen to not get caught out by (the small number of) reviewers who actually bother to measure gear. A typical 50W amp might measure 55 or 60 watts on test - perhaps more.

Peak power is the same, only the peak voltage rather than the RMS value was used. Obviously, that's a straightforward doubling of power for a sine wave. Understandably, many are cynical of this, but in practice, music is rarely sinusoidal, and the practice of focusing on "continuous average sine wave power" is a arguably not exactly representative of how an audio amplifier behaves with real signals. But it is an easy test to do, I suppose.

The peak to mean ratio of a sine wave is 3dB, whereas real music can be much larger - perhaps 20dB. So a 100 watt amp that is just hitting clipping on the peaks might only be delivering an average power of 1 watt to the load. This is good news for designers, who can save a lot of money on heat sinking and power supply components - the non-audio engineers will cynically decry that, but those in the industry know that building a 100 watt audio amp that is capable of delivering a 100 watt sine wave continuously would be massively over-engineered (and costly) for the intended application. In a business sense, over-engineering is as bad, perhaps worse, than under-engineering.

But all this means that audio amplifiers deliver higher voltages with music than they do with sine waves, and this introduces the concept of "dynamic headroom", which could be a couple of dB in practice. This is one of several reasons why regulated power supplies in audio power amps with a mains power supply are rare (and actually not a good idea). Again, non-audio cynics will (understandably) scoff, but in practice it's welcome.

All that only scratches the surface, but suffice to say, power is complicated when it comes to audio. But it's one of those nice scenarios where real-world use does actually favour the designers for a change. Not just of amplifiers, but also loudspeaker manufacturers. If you think amplifier power sounds tricky, speaker power is a whole other world!

I'm worried that by discussing audio power specifically, we've straying from the topic here. If this develops further, or if the OP feels it appropriate, I'll break it out into a new thread.

I'm also thinking about moving this thread to "General Chat". Mike, do you have a specific reason for wanting this to be in the private section? If so, I'll respect your wishes, but as a general point, I'm starting to feel that we might use the private areas a bit too much - which means that non-members miss out on potentially interesting and valuable discussions - discussions that might attract new members, perhaps...

[Crackle]

Thanks for your comments and explanations. You can probably guess the reason I queried it was that I too would assume an unqualified 10v in an AC circuit to mean 10v RMS.

Mike

[Craig]

Oddly enough there is no traceable definition of RMS voltage. The only way that NPL/PTB/NIST etc do it is using a thermal transfer standard. Because DC voltage is defined in terms of the Josephson effect to umpteen places of decimals, you compare the heating effect of the AC voltage being calibrated with the heating effect of a DC voltage derived from the Josephson standard.

Craig