Am I measuring the output wattage correctly?

[joulupukki]

Am I doing this right? I don’t have a pure resistive load, but I do have a reactive load (Torpedo Captor X). I’ve got the main output plugged into that. Out of the ext speaker jack I have it connected to a short cable that I’ve got tied to my oscilloscope and my digital multimeter. I’m feeding a 1000Hz signal into the input with the master volume dimed, and just enough volume knob without it clipping/distorting.

If I’m making the calculations correctly, my Mk I version of this amp is putting out 16W of clean output and Mk II is around 21W. Are these calculations correct?

The scope picture shows measuring the Mk II:

 

[Phrygian77]

It is unlikely that your speaker is 8 ohm at 1000Hz.

Edit: Derp der. You're Captor is also probably not 8 ohm at 1kHz either. It has a choke to simulate the impedance rise of a speaker.

 

[bebopbrain]

Looks good. You could try it at another frequency (125 Hz?) to see if the reactive load is throwing things off.

 

[SerpentRuss]

You must have a few extension speakers lying around, why not hook one up and throw a blanket over it (so you can stand the sound level) and see what you get, it would be an interesting comparison to this test. As far as your math goes, if that's what you're asking, I think your calculations are correct.

 

[joulupukki]

Looks good. You could try it at another frequency (125 Hz?) to see if the reactive load is throwing things off.

Good idea. I’ll do that before putting everything away.

You must have a few extension speakers lying around, why not hook one up and throw a blanket over it (so you can stand the sound level) and see what you get, it would be an interesting comparison to this test. As far as your math goes, if that's what you're asking, I think your calculations are correct.

Hehe. My family is gonna LOVE that. But I like that thinking. I’ll see if I can do it quickly without them noticing.

 

[Phrygian77]

@joulupukki it's possible that lower frequencies may be slightly less than 8 ohm. You can just measure what it is at 1KHz. It would be helpful to know the output impedance of your signal generator. Put a 100 to 1k resistor in series with the generator and the Captor. Send 1k into it, and measure the voltage drop across the resistor to calulate current (obviously you should measure the exact resistance first). Then, measure the voltage across the Captor.

 

[tele_savales]

I can't even imagine pushing my not-in-a-cab bench speaker to the onset of clipping w a 1000hz signal, even though I do hate my upstairs neighbor.

Since you've been on a roll with multiple builds going you might as well make a load box. There's a whole lot of different schematics on Aikens site.

 

[tubedude]

The standard for audio measurements is 1kHz resistive. So build/buy a resistive load for amp measurements. Then all results can be used in comparison. Reactive loads are for speaker simulation. They have a rising impedance and are not good for measurements, just reducing levels and generating a signal for recording, reamping, whatever.
The peak to peak voltage is cut in half, multiplied by .707, squared, then divided by your load, usually 8 or 4 Ohms. That gives you RMS power. It seems like a low level because amp manufacturers misrepresent power in various ways. I just built an AB KT88 pair final that puts out 26 WRMS. Similar amps out there will say it has a 70W output. A Twin Reverb on the bench recently put out 33 WRMS. Fender calls it a 100 Watt amp.

 

[joulupukki]

@Phrygian77 interesting idea to figure out the impedance. I‘ll have to give that a shot another time.

I remembered that I have this non-reactive attenuator I built (and never use). As far as I know, it’s a resistive load and turned all the way down measures 8.5 Ohms of resistance. Hooked up to that…

Not too much difference between 125Hz and 1000Hz so I’ll only show the 1000Hz:

Interestingly, why would Mk I not be putting out a smooth sine wave like Mk II?

 

[tele_savales]

A Twin Reverb on the bench recently put out 33 WRMS. Fender calls it a 100 Watt amp.

How is that even possible if the amp had all 4 tubes in it and was biased correctly? Not trying to be combative, I just don't understand that at all.

 

[tubedude]

That's what I'm saying. No mfgr. uses RMS as a spec. Your 100 Watt amp is really much less. But don't be depressed, it's still powerful AF, it's just a numbers game.

 

[Phrygian77]

That's what I'm saying. No mfgr. uses RMS as a spec. Your 100 Watt amp is really much less. But don't be depressed, it's still powerful AF, it's just a numbers game.

I'm also skeptical of your numbers, as they're about half of what they should be.

 

[corliss1]

I also disagree with @tubedude 's numbers Those amps really can make 100W all day long.

@joulupukki - your math seems fine to me.

For fun, let's look at the '68 Fender Twin Reverb schematic. The modern service manuals include the AC voltage component along with the DC values. They are putting in a wimpy 10mV AC for their test, for whatever reason. Fine, that output shows 2.3VAC at the speaker jack. So we get:

2.3^2 / 4ohm = 5.29/4 = 1.3W

A wimpy input gives us a wimpy output. Now...let's put in an actual guitar signal. I've found that vintage humbuckers are somewhere around 100mV, give or take - it's a nice easy number to work with. Fender's number of 10mV going in and 2.3V out let's us make a nice, clean ratio:

2.3 / 0.010

which gets us about 230X amplification.

So...my 100mVAC in would be about 23VAC at the speaker jack (that's .100 x 230 = 23). That seems like a reasonable number to me - let's do the math:

23VAC^2 / 4ohm = 529/4 = 132W

So you want RMS from that? Fine, let's get our magic .707 number:

23VAC x .707 = 16VAC RMS at the jack

16^2 / 4ohm = 256 / 4ohm = 64

Maybe we have some Dimarzios instead, and we get to the 85W range pretty quickly. If your Twin is at 33W...something is up.

Another thought worth noting, is that your meter very likely reads the AC component in RMS already, which means we likely don't need to do the .707 math at all if you're taking your AC reading with most any multimeter (although check your user manual to be sure)

I *think* I got all that right, but the numbers got blurry after a bit.................

 

[tubedude]

You first should convert peak to peak voltage to VRMS, then square it and divide by the load resistance.
RMS is .707 of the PEAK Voltage, not peak to peak. So half the peak to peak voltage to get peak, THEN multiply by .707 to get VRMS, square it and divide by the load.
I generally use 500mV as my input voltage. Fender is using a lower number likely to insure the amplified signal is not clipped, so as to get a good number to check that the gain stages are operating correctly. Remember, this procedure was developed in the 1950's when few people had a scope. So it's done with a meter only. Without the benefit of an oscilloscope it's hard to know the onset of clipping, so the input is safe sided to get a valid number to verify operation, not as a full output test.
With a scope, you can use whatever voltage is needed to get the maximum clean signal for determining full output.
So crank your amp to get max clean sine wave output @ 1kHz. Measure the Peak to Peak height of the displayed sine wave, divide in half to get the Peak voltage, multiply by .707 to get VRMS, square it , then divide by the load impedance to get RMS power.

 

[tubedude]

The problem with using a meter is the assumption that a displayed RMS value is unclipped. That's not the case. It could be a square wave. All an RMS reding on a meter means is the input is converted to its DC equivalent. It can't tell if the signal is clipped. Measure output with an RMS meter of max clean waveform. As you increase the signal progressively toward a squarewave the number will increase. If it was measuring just clean RMS voltage the number would stay the same.
True RMS meters are converting the input, clean or dirty, mix of AC and DC, to the DC equivalent voltage that will generate the same amount of heat in a load. It doesn't read correctly for our purposes unless the signal is an undistorted sine wave.

 

[tubedude]

Good idea. I’ll do that before putting everything away.


Hehe. My family is gonna LOVE that. But I like that thinking. I’ll see if I can do it quickly without them noticing.

oh, they'll notice.

 

[tubedude]

@Phrygian77 interesting idea to figure out the impedance. I‘ll have to give that a shot another time.

I remembered that I have this non-reactive attenuator I built (and never use). As far as I know, it’s a resistive load and turned all the way down measures 8.5 Ohms of resistance. Hooked up to that…

View attachment 1099701

Not too much difference between 125Hz and 1000Hz so I’ll only show the 1000Hz:
View attachment 1099702

Interestingly, why would Mk I not be putting out a smooth sine wave like Mk II?
View attachment 1099703 View attachment 1099704

Your Mk1 waveform looks like some rectification spikes from a noise source. Is it repeatable? Is there a phone, TV, or game console nearby?
If the external environment is the same, and the results are repeatable, you might have internal coupling of rectifier pulses. Maybe time to open it up and scope the power supply waveforms. This is a what mark 1, Boogie?
Can you post a schematic?

 

[joulupukki]

Your Mk1 waveform looks like some rectification spikes from a noise source. Is it repeatable? Is there a phone, TV, or game console nearby?
If the external environment is the same, and the results are repeatable, you might have internal coupling of rectifier pulses. Maybe time to open it up and scope the power supply waveforms. This is a what mark 1, Boogie?
Can you post a schematic?

@tubedude thank you for all your good answers.

These amps are single-channel AB763 6V6 amps with a few modifications from the original circuit (like bias wiggle tremolo). I’ve just dubbed them the Mk I and Mk II so I can tell them apart when talking about them. I built them from scratch using the width of a Princeton Reverb with concepts borrowed from Rob Robinette’s Blackvibe 6V6 and Hoffman’s 1-channel AB763.

There are a couple differences in implementation between the two…

Mk I:
- Uses a stand-up power transformer from Musical Power Supplies (PT330M, 660Vac center tapped 140mA: https://www.musicalpowersupplies.com/app/download/7115014784/PT330M_REVD_DWG.pdf)
- Uses a 40/20/20 can capacitor for everything but the B+4, which I’ve used a 22uF F&T
- B+ wires fly over the top
- Grounding more traditional Fender style with PT ground separate from preamp ground
- Uses a stainless steel chassis I salvaged from the front panel of a tankless water hear (never again – it was such a pain to drill!)
- Uses generic CF and MF resistors and Mallory 150 caps

Mk II:
- Uses a lay down power transformer (Mojotone Deluxe Reverb PT, significantly higher voltages, 708Vac center tapped 150mA: https://www.mojotone.com/Deluxe-Reverb-Power-Transformer-125P23B)
- Uses Nichicon radials for reservoir and filter caps with each section of the circuit grounded at the negative terminal of the cap
- B+ and most runners go under the board
- Uses a chassis I made out of 5052 0.090 aluminum sheet
- Uses Dale MF resistors and upgraded SoZo Blue, some Jupiter, and Mallory 150 caps

The schematic is attached. I’ve also attached the two different layouts except the layout for the Mk I doesn’t show a couple of actual updates (like elevated heater center tap and the abandoned 4th cap in the can cap that I replaced with the 22uF F&T in the preamp).

Mk I definitely has more “buzz” at full tilt. I‘ve always chalked that up to thinking it’s the PT (audibly louder even w/o a speaker). But perhaps it’s something else and related to the non-smooth sine wave that I’m seeing via the scope.

 

[Paul G.]

I had a tech back in the day that would measure output of various amps. He said Twin Reverbs put out an average 67WRMS. Which, BTW, is a lot.

 

[joulupukki]

I can’t seem to duplicate the spikes like I could yesterday in Mk I. The only thing that changed is that …

I took MK I’s GZ34 and put it in Mk II to see if MK II would be happy with it. It was. So I put it back into Mk I. Since then, it seems to be much cleaner. Maybe there was some slight corrosion on the rectifier tube pins or on the socket? Weird.