JCM900 Low Output

[J M Fahey]

Math aside (your peak power is 2xRMS but that's not the point now) , you have what you have.
Your tubes in your amp put out 36W unclipped, simple as that.
I have *often* measured around 40W unclipped under those conditions, so in the real world you are not that bad.
And I bet in actual stage use that amp works very well.
I wouldn't worry toomuch now about a *number*, if the amp reasonably does its job.
At most, someday you have a new set of different brand tubes you retest it and tell us what you find.
Good luck.

[jrfrond]

Math aside (your peak power is 2xRMS but that's not the point now) , you have what you have.
Your tubes in your amp put out 36W unclipped, simple as that.
I have *often* measured around 40W unclipped under those conditions, so in the real world you are not that bad.
And I bet in actual stage use that amp works very well.
I wouldn't worry toomuch now about a *number*, if the amp reasonably does its job.
At most, someday you have a new set of different brand tubes you retest it and tell us what you find.
Good luck.

This is one of the things I was getting at. I've seen a lot of Marshall's that put out less than their rated power. It doesn't mean there is anything wrong, and sometimes you need to forget about the numbers and evaluate by performance. Marshall doesn't have a corner on this either. Guitar amps can and do vary all over the map. I serviced a SF Fender Twin today that put out 72WRMS, 10% under it's rated power, but behaved and sounded the way a Twin should when played and sounded great. That was with new power tubes and screen resistors. F**k the numbers, it was fine. Amps behave quite differently when under resistive rather than inductive load, and that renders your numbers NOMINAL.

[rf7]

Alright, thanks guys. I did try putting back in some 6L6's, and they also only put out 36Wrms.

[MarkusBass]

I think that you are chasing a ghost. I had exactly the same problem with Marshall amp (I was getting exactly 36W). Since I do not fix valve amps on a daily basis I was thinking that there is still something wrong with the amp (after fixing it). I searched the net, I downloaded EL34 and 6L6 datasheets and I came to a conclusion that this is normal. Now I have Fender Dual Showman (4x6L6) and on the rear of the amp it's written: Speaker 85W. So it looks like 36-40W power out of 2x6L6 seems to be correct. I would ask another question: is there any person that measured 50W RMS from this Marshall (or any 2x6L6 amp), or 100W out of 4x6L6 amp?

Mark

[km6xz]

These are not Hi-Fi amps so do not need to be tested at the FTC level for published specs. So forget the RMS power at specified Z, at specified bandwidth at not more than specified Thd, after a specific pre-test 1/3 power preconditioning to heat and stress them.
The reason is because they are not designed to be clean, they are designed to create Thd as the normal function so hi-fi tests would be meaningless.
Companies rate guitar amps by RMS power into a specified load Z, with no reference to bandwidth, distortion or conditioning. The published power rating is very dependent on waveform/duty cycle. A distorted signal, clipped for example has a higher RMS value than a clean sine wave at the same amplitude. So knowing how to measure the distorted signal's RMS value makes a difference. DMMs can have different AC voltage calibration: peak, peak-with average calibration(most common), average and RMS. RMS can be true RMS or calculated RMS. RMS voltage is the same heating power of a DC source driving the same resistive load. It can be detected a number of way for example by integration of all the instantaneous values of potential times current for all the points under the waveform's envelope. A 10 volt RMS sine wave would have the same heating power as a 10 volt DC supply feeding the same resistive load.
So a couple things are adding confusing. One is the wave form and the other is the calibration or method of arriving at the value by the meter. Two different models TRUE RMS voltmeters can be sensitive to the harmonics, all of them or some of them, so different levels of distortion would result in different readings.
What type of meter, is it Average measuring, calibrated in RMS like most or does it have an actual RMS converter inside? The traditional method is using a thermometer to detect the rise in water temperature cooling a submerged load. That is how RMS values were defined in the first place. Thermocouples can be used for measuring RF power usually...it the net temp rise negates any concern over waveform distortion. The amount or type of distortion will not cause errors in the reading of temperature.
Once you know what the reading really means when it says 17volts for example, you need to know the distortion.
Basically, a tube amp(not a solid state amp which has a power section that is not designed to create distortion....if would be too ugly if it did...can be tested like a hi-fi amp and readings will be less ambiguous since it is not driven into clipping in the test) can create a heck of a lot of distortion in the output which increases true RMS value, heating equivalency value, when increasing drive past onset of clipping(so does a solid state amp but the output sound would be so unpleasant that you wouldn't). A guitar amp with 2 6L6s would would likely show only 35-40 watts at the onset of clipping, but upping the drive, can easily end up with 50 watts in generation and still be quite listenable, in fact having a desirable tone.
There are too many variables here that are unknown to say if your amp is having a problem. If your meter is not true RMS and if you are not driving it past the onset of clipping than you are doing fine, and you can expect it to churn out plenty more power real power using different test methods. If the meter is correct and you are driving into clipping, there might still be a matching problem

[J M Fahey]

I would ask another question: is there any person that measured 50W RMS from this Marshall (or any 2x6L6 amp), or 100W out of 4x6L6 amp?

I never could in most, but I *did* measure almost 100W from a 4x6L6 amp.
Fact is, it was a Matchless Chieftain "120" so in theory (or in practice) they were not delivering either.
I think it managed to practically reach those 100W by having transformers twice the weight as Marshall's.
In a nutshell: it can be done, but it's not easy.
EDIT:
I forgot: I also measured 100W RMS without visible clipping in a Mesa Boogie bass amp, the one with 2x10" speakers, packed in a airport-suitcase-like cabinet , complete with two wheels embedded on the back and a collapsible handle.
Fact is, it used 6 (six) x 6L6 output tubes.
I guess it was the same output stage as a Triple Rectifier, although I'm only gessing here, I've never had one of them naked in my bench.
Now that I remember, "good" Soldanos, as SLO100 or HRD+ (if that's the model name) also come quite close to the mythical 100W RMS.

EDIT2: If you *really* want your 50 or 100W, leave screens at 400V but rise plates to, say, 500V or a little bit more.
Not really needed though, more of a curse than a blessing in my eyes.