Measuring output power

[bsco]

Got a question....I am trying to measure the output power of a Marshall VS100R Combo which is rated for 100W into 4 ohms...The amp has an 8 ohm speaker......so it's total power out should be 50W.....?????anyway, there is a clean channel, and overdrive 1 channel and an overdrive 2 channel....I have all the amp tone controls set at 12 o'clock.....all gains and volumes at max....reverb, effects mix etc....are set at 0........I measure 12.5 VAC at the output....square that and divide by 8 and I get roughly 19.5W...a far cry from 50W.....am I doing something wrong here?? I was using the clean channel for the test.....when I use the overdrive channels the output voltage only increases by a volt or two and if pushed to max input signal, will show a highly distorted signal.....so what am I doing wrong here......or do I have a problem with the amp itself?? Any advise is highly welcome......
Cheers,
Bernie
P.S. I remember Enzo once told me that the best of people in all fields seek advice from other people.....I am not saying I am the best....far from it......but I take Enzo's advice very seriously.......so hopefully somebody can shed some light on this for me.......

[g1]

Go in to the FX return. For this model you will need to insert a dummy plug in the regular input jack to be able to do this (stupid design). Adjust to get your max. output before clipping on the scope, then calculate power output.

[bsco]

OK. Thanks g-one.....I'll check it out....does this procedure apply to all amps with both clean and overdrive channels???
Just curious......
Cheers

[Jazz P Bass]

For the power test, are you using a signal generator?
If so, what amplitude/
I usually use 100mv/ 1K.

[bsco]

For the power test, are you using a signal generator?
If so, what amplitude/
I usually use 100mv/ 1K.

Hi Jazz.....I was using a signal generator.....had the amp volume set to max and adjusted the generator output till the waveform clips then back off slightly.....please correct me if this is not correct......can't take any pictures as my wife is away visiting family and she has the digital camera with her......

[nsubulysses]

Bsco you seem to be on the right track. Sorry if this is redundant BUT my route is set amp volume settings on 10, set tone knobs on 5. In some amps you will not get a clean sine wave reproduction if tone knobs are cranked. For some amps (Sound Citys and some other amps with active preamps) you will not get full power unless you put the tone knobs on 10, but that is kind of rare.

Anyway, with amp volume all the way up bring up the volume on the signal gen until it reaches clipping. You can measure peak-peak and multiply it by .3535 to get your RMS voltage, or just put an AC voltmeter on the dummy load to get your RMS AC voltage. to calculate power do RMS^2/load impedance.

For example, 20VAC with 8 ohm load would be 20 x 20 / 8 = 50W power.

1K sine wave is good. Avoid using sine waves that are multiples of the line frequency, 60Hz (120hz, 240, 600, etc.). You will get some weird effects that are misleading but I forgot what they are. I'm sure someone can comment. I once was testing an amp with a 600Hz signal by accident and I thought there was a problem but there wasn't, and I forgot exactly what happened.

[eschertron]

If you have input gain at max, you may be seeing distortion and clipping in the preamp. No good. You want to make sure you have max unclipped waveform on the power tubes to measure the power correctly. One way to do that reliably is to scope the wave form at the grids and verify a sine wave. Well, G-one's advice to skip the preamp entirely is good, too. That'll usually get it done. Set MV to max, and everything else that matters to where you need it to get the appropriate sine wave at the power tube grids.

[bsco]

If you have input gain at max, you may be seeing distortion and clipping in the preamp. No good. You want to make sure you have max unclipped waveform on the power tubes to measure the power correctly. One way to do that reliably is to scope the wave form at the grids and verify a sine wave. Well, G-one's advice to skip the preamp entirely is good, too. That'll usually get it done. Set MV to max, and everything else that matters to where you need it to get the appropriate sine wave at the power tube grids.

Yes. I am going to try g-one's method first...... but with a 8 ohm speaker in the amp the max power would only be 50W not 100 right??

[eschertron]

Yes. I am going to try g-one's method first...... but with a 8 ohm speaker in the amp the max power would only be 50W not 100 right??

I say you are right. If the output transistors are going from rail to rail, the voltage doesn't change (20v RMS per the example above) but the speaker load does. Twice the resistance, half the effective power. (using P = E^2/R)

edit: As long as you're not expecting to find some component failure that affects power output, it's really all about the voltage rails for a SS amp.

[nevetslab]

In a solid state amp, it not having an output transformer, I'd say yes. Using the 4 ohm tap with an 8 ohm load, somewhat unloading the output stage, it may not be a 1:2 relationship. If all were working right, you'd probably get a little more out, rather than less with twice the rated load.

but yeah....12.5VAC is way under what you should see. Assuming as eschertron hinted you're not clipping in the preamp. Jacking in as g-one suggests thru the Effects Return, you're looking at a linear power amp circuit.

[loudthud]

Some meters don't measure accurately at 1KHz. You can test this by just connecting the meter to the generator and sweeping the frequency.

[bsco]

In a solid state amp, it not having an output transformer, I'd say yes. Using the 4 ohm tap with an 8 ohm load, somewhat unloading the output stage, it may not be a 1:2 relationship. If all were working right, you'd probably get a little more out, rather than less with twice the rated load.

but yeah....12.5VAC is way under what you should see. Assuming as eschertron hinted you're not clipping in the preamp. Jacking in as g-one suggests thru the Effects Return, you're looking at a linear power amp circuit.

Actually, there is no 4 ohm out....it only specifies that the amp will produce 100W of power into a 4 ohm load....the amp comes with a 8 ohm speaker.....and there is no master volume control on this....I am going to run the sig. gen into the effects return like g-one suggested and put a shorting jack into the input...and see what happens.....I should see 20v rms on my meter.....wait...the meter is not an rms meter....just a regular DMM.....I will have to take the p-p reading and convert to rms.....707xpeak or 1.414xp-p.....I will post my findings.....shortly.......
Cheers

[J M Fahey]

In theory, 8 ohms power should be half 4 ohms one.

In practice it's not , but somewhat more, because power supply voltage varies with load, so driving 8 ohms you have higher rail voltages than when driving 4 ohms.

That amp, which if I don't remember wrong has idle +/-40V rails , can put out some 25 or 26V RMS ... unloaded , which come from:
40Vpk - 4V (drop across driver and output transistors and it's in the datasheet) x .707=25.5V RMS

Now with rated 4 ohms load +/-V rails drop, both because of resistive losses in the power transformer and because important ripple appears, these amps use the skimpiest filter caps they can get away with, 2200uFx50V per rail.
You can easily have around 3V ripple there which of course must be substracted in any honest power measurement.

That is what nsublysses mentions: if you use a frequency which is multiple of line frequency (we have a lot of that, because our 50Hz mains means 400Hz, and specially very popular 1KHz are exact multiples) you will have a very fake reading and won't notice it, because measuring signal will be "synced" with a given part of ripple waveform, easy to have 3 or 4 V measured errors which on a 20V something signal mean a lot.
So on Monday you measure 85W output and on Tuesday you measure 115W ... same amp same load

And since meanwhile you switched fuse brands , you publish in DIYForum that "the new expensive fuse increased power by 20% and I can prove it"

So you must vary test signal frequency until when clipping you clearly see ripple modulating signal peaks, the back off a little until it's real clean.

Easy done with a signal generator, not so with a downloaded fixed frequency test signal.

Back to measurements, that amp should provide 20V RMS into 4 ohms meaning:
20V*1.41 + 4V (loss) +3V (ripple)=+/-35V rails

Into 8 ohms, supply will sag half, ripple will also be half, so same amp can probably provide some 22V RMS or a little more, meaning around 60/65W RMS .

Its granddaddy, basically same amp concept Marshall Valvestate 8080, claimed "80W RMS into 8 ohms" which was .... let's be kind .... "optimistic" , much closer to 70W (if that much) , but real loud and usable because it had a "real" speaker, Celestion G1275T , same one furnished in the standard 4x12" and a much more important factor than a couple Volts more or less.

So on your VS100 you should have some 20V RMS into 4 ohms and some 22V RMS into 8 ; 12 something volts definitely sound like a problem.

To play it safe:
a) get in as close to power amp as possible, Loop return is acceptable, if you have a MIX control set it fully towards the effects side, you may also have a sensitivity/level control (say, +4/-10dB ) set it to max sensitivity so you make certain your oscillator can drive it.
b) scope the output until it starts clipping ugly, showing ripple, etc. , the back up until top is almost clean.
c) now scope power amp input, to see whether it's getting clean signal or not.
Otherwise clipping in an earlier stage will drive you crazy

Good luck

EDIT: of course, this is long and complicated, so you can also use the Gerald Weber "I don't use scopes" method:

[eschertron]

I should see 20v rms on my meter.....wait...the meter is not an rms meter....just a regular DMM.....I will have to take the p-p reading and convert to rms...

if you can select either peak or average readings on your meter, you've got a nice one! And you can check your math too; for a sine wave, peak should be 1.414x as you said. My meters are all the simple kind, and will return an average volts reading, like 118vac or so, when I test the household mains.

So for a sine wave, average and RMS will be the same. For any other waveform, you would need a "true RMS" meter to get an accurate reading. That's why sine waves are popular with technicians

[The Dude]

EDIT: of course, this is long and complicated, so you can also use the Gerald Weber "I don't use scopes" method:

OMG Juan. That is hilarious! There are just so many things wrong with that, I wouldn't know where to start. Perhaps most f'd up is that he makes no attempt to play something close to a constant tone for "accurate" (used loosely) measurement. Instead he's jamming licks. Well, that'll get you a consistent output voltage now won't it? Why would you want to bother with one of the two scopes or the tone generator sitting on the bench in the background?