Rough 6L6 wattage output with given B+

[EFK]

Hopefully not TOO much of a sidetrack, but if you just stick your meter across the speaker terminals and measure AC voltage, square it and divide by the speaker load (which everyone in internet-land tells you to do), *just how close* is one REALLY getting to the output of the amp? I'm not talking RMS output, I'm just indicating plain old max output when the amp is "dimed." And this is assuming a matched load, of course. I suspect the results of doing this are usually optimistic using a plain old cheap DVM?

[Steve Conner]

A cheap DVM's AC ranges will rectify the waveform, take the average value, and multiply by 1.11, the "form factor" of a sine wave.

If we assume that our dimed amp puts out a square wave, with a form factor of 1.0, then the calculated power will be high by a factor of 1.11 squared, which is 1.23, or 23%.

If you drive the amp to the edge of clipping with a sine wave, then the DVM should give a reasonably accurate answer. A true RMS meter should give the right answer irrespective of the waveform. I use the meter section of my distortion analyser, since it's normally hooked across the dummy load already.

[pdf64]

Further to that, all DMMs will only respond up to a certain frequency, due to the sampling rate etc; for a cheap DMM, that may be quite a low frequency, eg 200Hz. So assuming a square wave input, a cheap meter may read low due to the extended harmonic content that it isn't catching.
And the impedance of a speaker varies significantly, from a bit under nominal to maybe ten times nominal (at bass resonance), rising to about twice nominal by about 2kHz. The high output impedance of a cranked tube amp will track that, and put out more voltage at above nominal impedance.
And with a high power amp, hearing damage is a distinct likelihood!
So the method will give a rough 'ball park' idea of power output, but is way too inaccurate to judge whether there was a performance problem, such as worn tubes, caps etc.
Pete.
So a

[imaradiostar]

I use the meter section of my distortion analyser, since it's normally hooked across the dummy load already.

What model do you have? I just acquired a dead HP 339a- need to fix it soon.

My usual power output measurement method is to use two traces on my scope with one on the input and one on the output and run the oscillator level up till just before I can see a difference in the two. I read the output voltage using a true rms bench meter and get a pretty accurate result. It usually is very close to the results with an inexpensive meter as long as I don't let the waveform get away from a sine shape, which makes sense. The true rms meter is nice when looking for the distorted output power at "11".

JT

[Steve Conner]

It's a Tektronix DA4084, the military version of the AA501.


Lord knows what the military need a distortion analyser for, but there it is. It came without an oscillator section, so I built my own, which was quite a project in its own right.

[pdf64]

'My usual power output measurement method is to use two traces on my scope with one on the input and one on the output and run the oscillator level up till just before I can see a difference in the two'

My feeling is that works ok on amps with global negative feedback on the power amp, in that the input and output waveforms are closely similar up until the peaks get clipped.
However, on those that don't (eg Vox AC30) the waveform can get 'rounded off / fattened up / compressed' some way before the output has reached its limit. That's when I feel a distortion meter may help to give a better comparitive measurement, in the AC30 may have 20% THD (my guess) when it hits the limit, whereas a eg Deluxe Reverb may be only 5% THD.
It could be perceived that all that extra harmonic content would give the Vox a falsely high Vrms, and so power output, compared to the Fender.
An alternitive view is that for an instrument amp, it's the harsh, 'high order' harmonics that result from clipping which are most significant for the user's perception of power, and that the large 'low order' harmonic content from the Vox is an acceptable part of its clean output. and so should be included in the power output measurement.
Whatever, I'd like a distortion analyser / meter to help me make that judgement.
Pete.

[Steve Conner]

Well, that opens up a whole can of worms as to what distortion level a guitar amp's power output should be quoted at. But either way a distortion analyser is a very handy tool for R&D. However, for guitar amp use you could get by with a PC soundcard and some audio analysis software. This might even be a better option. One issue with my analyser at least, is that it struggles to lock to a waveform with more than 10% THD.

I had to get a real distortion analyser because my latest hi-fi amp design was making less distortion than my best soundcard.

[big_teee]

So Idle current wattage estimating, is lower than real output Wattage?
That is my question.
So my Marshall type clone, that has a loaded Plate Voltage of 492V x .036ma =17.7 watts x2 =35.4 watts.
So is this any indication of what actual wattage of the amp is?
Please keep it simple!
T

[pdf64]

'for guitar amp use you could get by with a PC soundcard and some audio analysis software'

Good idea, it would be nice to see some evidence for my hunches regarding low and high order harmonics.
Any recommendations for freeware / shareware audio analysis software, ideally to run on a Win7 netbook? And a good downloading host site that doesn't include dodgy malware as part of the package.
Pete

[Steve Conner]

Big Tee asked to keep it simple, so here goes:

There is no relation between idle dissipation and output power.

If you play with the bias on a fixed bias amp, you can see for yourself that it doesn't affect the maximum output power available before clipping. It affects the gain of the output stage: as you make the bias colder you'll see that you need more drive to push it to the point of clipping. But it doesn't change the power level where clipping starts.

I use Spectrum Lab for my audio analysis needs. It's freeware and can be downloaded direct from the author's website. The catch is that it's incredibly complicated and tricky to set up.

[imaradiostar]

I'm going to slightly complicate what he said and add that if the power supply is sagging, to some small extent you'll get a change in power as you alter output bias because the available supply voltage changes. Certainly not worth worrying about, of course.

As much as I love building for big power, why does it really matter? You probably would be hard pressed to hear the difference between 40 and 45 watts anyway. I've spent a lot of time with a scope and voltmeter playing with parameters and in the end I understand things a little better but my ears are still the same!

jamie