I measured my amp's power output - it doesn't seem right.

[Sonicboom]

I did a search in this forum and came up bust, so here's one for the experts:

My Fender has always seemed to be quiet for a 45w amp, especially compared to my 50w Marshall. I know that there are a host of reasons for a diff. in SPL. Anyway, I connected a scope to the speaker terminal and ran a signal into the amp's input. Regardless of frequency, the sine wave output was clean until about "3" on the volume knob. The amplitude was 10v pk to pk. Anything beyond 3, the signal just started clipping, becoming almost a square wave. The amp did seem to get slightly louder, but not much.

Using 10v pk to pk, I calculated the RMS value, squared the result and divided by the speaker impedance of 8 ohms (result was ~6 watts). I forgot that speaker impedance varies by frequency, so I will redo using an 8 ohm resistor. Is this process of measuring voltage with a scope and calculating power (using a resistor though) accurate?

Thanks.

[imaradiostar]

I've been down this road as well. For years I've played a low power Marshall clone that I built and I was surprised to find it only makes about 12 watts before it distorts- I expected more like 18 to 22. My measurement method is a little different. I route a sine wave from a signal generator into the phase inverter and measure the output to a resistive load using a true RMS voltmeter and monitoring the PI input and amp output with a scope. When I can just see that the input and output waveforms aren't the same I call it "max power." That rms voltage squared divided by the load resistance gives wattage.

I'd wager this is a conservative rating. Cranking the generator up results in distorted output in excess of what would be expected from my amp.

jamie

[imaradiostar]

If you were really only getting 10 pk to pk I'd call that 1.57 watts into 8 ohms.

10/(2*root2)=3.54 v rms.

(3.54*3.54)=12.53

12.53/8 ohms=1.57 watts

Try running a sine wave in and view the output on a scope into an 8 ohm resistor. Any cheap digital multimeter should give a pretty good approximation of true rms as long as you're measuring a true sine wave.

jamie

[Enzo]

DOn't overlook the forest with all the trees blocking your view. I agree with Jamie's math.

WHen we start worrying about impedances and frequencies and stuff, or whether we are measuring with zero distortion or 10% or whatever distortion, we are ignoring the basic problem of 1.5 watts coming from a 45w amp. If the question was why is my 45w amp only making 38 watts it would be different. But a few percentage points either way can't explain 1.5w in place of 45.

Start with the power supply. Is there good B+ on both plate AND screen pins of BOTH power tubes. And also, is the bias on pin 5 of each, within the range of reasonable?

Now look at the small tubes, especially the phase inverter. Pins 1 and 6 of each tube are plates. Is ther something in the 200 v area on each one? I don't know what exactly, but we don;t want to see zero and we don't want to see 350v. And pins 3 and 8 are teh cathodes on each. MAke sure none of those is sitting at zero volts. If your amp has trem, the trem tube will measure funny.

And speaking of trem, make sure your "bug" isn;t hanging up your signal path.

SInce you are scoping, apply a good strong signal to the input. As a Fender, may I assume it has two channels? Normal and reverb? DO both channels result in the same weak signal? If so, then the problem probably lies after where they converge. Scope the grids of the power tubes, how mych signal is there? If there is only8 1/8 of a volt, then we can;t expect the full output to magically appear anyway. What level signal is going into the PI grid?

Pick one channel for the input, and follow the signal stage by stage through the amp. Where does it disappear?

[Sonicboom]

Thanks for the suggestions, Enzo. The amp is a '64 Vibroverb (AB763), and yes, both channels were affected. The 10M resistor feeding the plate on the vibrato ckt. (V5?) was open. I replaced it and the amp cleaned up noticeably. Funny thing, even after replacing the resistor, the voltage on the pin 1 side of the 220K resistor stays at 460v unless the vibrato is engaged at which point it drops to around 290v. I assume that this is "normal" operation - yes? All other DC voltage levels are good.

Plate and screen DC levels on the outputs are fine (~460v DC). The bias voltage is also in the ballpark at 50v or so - can't recall exactly. I replaced the filter caps and power resistors about two months ago.

I still have a couple of nagging issues to clear up though. I see what appears to be a 240 Hz triangular wave on the plates of the output tubes. This is not present on the grids, and I don't remember if it was on the screens. Also, when running a signal into the input and scoping the plates, the output signal has a slight modulation to it. I'm guessing 60 Hz, although the amp is almost dead quite until you crank the vol. to 10 with no input signal; then, there is a faint 60 Hz hum.

Any thoughts on the modulation or triangular wave? Both hardly seem normal.

Thanks.

[g1]

The modulation and triangular wave could be ripple, if so, not a problem if you have no objectionable hum.
Can you try measuring the output with a meter while the scope is showing onset of clipping at the output. Try a higher volume setting and a lower level input signal.

[Chuck H]

If your scope probe has a ground clip, use it. Often you need the scope cable shield grounded at the same 0V as the unit under test to avoid funny ghost impressions on the reading.

Pretty much all ABxxx amps hum a little at higher volumes because of ground scheme and the reverb.

What voltage is the input signal when testing?

Are you testing into a dummy load?

What is the AC voltage across the speaker terminals at the onset of clipping?

[Sonicboom]

Well, the Vibroverb is back to being loud and proud. Just goes to show the importance of a 20 cent part.

I have a '66 Super Reverb that I pulled out for comparison. I was wondering if it had the same weird signals on the plates as my VV. Both amps have a triangular shaped wave on the plates (with or w/o an input signal), which cannot be heard in either amp. In fact, after moving a few wires around on the VV, it is now much quieter than before. No hiss and only a very faint hum when the vol. is on 10. Your ear has to be right by the speaker to hear it - not recommended in case someone spanks an A chord.

All in all, I think the SR is slightly louder than the VV, but not by much. The VV gets nastier while the SR has a bit more headroom. Scoping the output signal of each amp (at the speaker), both are equal in amplitude. I don't have a signal generator, so I improvised with a guitar and an eBow. This actually worked really well since the input signal is automatically at guitar level.

I have to say, the VV sounds the best I've ever heard it. Apparently, the resistor was bad when I bought it.

Thanks, guys, for the help.

[Enzo]

I think if you look at it, your triangle waveform is really what we call a "sawtooth" waveform. The leading edge is much steeper than the trailing edge. It is the ripple on the B+ supply, and is quite normal.

And if you were to set your scope zero trace down at the bottom of the screen, and put the scope on DC, so the entire DC B+ level would stay on the screen up near the top, I bet that sawtooth would be too small to see. In fact with near ful power audio on the screen, reading at those plates, now how large is that ripple sawtooth looking?

I don't know if there is a term for it or not, but I call it "trace chasing." People not used to staring at a scope, start turning up the scope vertical gain until they see some sort of waveform. You turn it up enough, and eventually you WILL see some sort of waveform. The trick is to put it in context. All readings from any test equipment need to be in context. If you have 200 volts of signal on those plates, a volt of ripple is not very significant. But crank the screen up so the ripple fills the view, and all of a sudden it looks horrendous. Like those electron microscope pictures of fleas, that look like giant space monsters. Or the bass player for the Red Hot Chili Peppers, who also looks like a space alien.

[Sonicboom]

Absolutely true. Since I' haven't scoped out a tube amp, or done any signal chasing for that matter, in 25+ years, this was a great re-learning experience. In the grand scheme of things, the ripple was pretty low, but I had just finished working on another piece of gear (an audio matrix switch) where the noise I needed to eliminate was from cheap power supply regulators with a high noise floor. Talk about small signals; my vintage Heathkit scope could barely measure it, but "in context" to the small input signal, it was large enough to have audible hiss.

I guess you could call the ripple I saw a sawtooth wave, but the leading edge was only slightly longer than the trailing edge. Not what I would have expected.

Great fun. Now, I only have one amp left to fix and 7 pedals to build.

[imaradiostar]

I'm confused. How much clean output is it making now, just for the sake of discussion?

jamie

[MarkusBass]

I also would like to know. And what was exactly the problem? Don't tell me that it was 10M open resistor in the vibrato circuit .

Mark

[imaradiostar]

I think if you look at it, your triangle waveform is really what we call a "sawtooth" waveform. The leading edge is much steeper than the trailing edge. It is the ripple on the B+ supply, and is quite normal.

And if you were to set your scope zero trace down at the bottom of the screen, and put the scope on DC, so the entire DC B+ level would stay on the screen up near the top, I bet that sawtooth would be too small to see. In fact with near ful power audio on the screen, reading at those plates, now how large is that ripple sawtooth looking?

I don't know if there is a term for it or not, but I call it "trace chasing." People not used to staring at a scope, start turning up the scope vertical gain until they see some sort of waveform. You turn it up enough, and eventually you WILL see some sort of waveform. The trick is to put it in context. All readings from any test equipment need to be in context. If you have 200 volts of signal on those plates, a volt of ripple is not very significant. But crank the screen up so the ripple fills the view, and all of a sudden it looks horrendous. Like those electron microscope pictures of fleas, that look like giant space monsters. Or the bass player for the Red Hot Chili Peppers, who also looks like a space alien.

We were working on a REALLY high voltage power amp at work that used high frequency switching with some mosfets and a giant pot core with litz wire to boost about 50 volts up to produce a pair of 1200 volt DC supplies for a bridged class d output. Whenever we fired it up all of the ungrounded scope probes on the other side of the room would go NUTS. This is obviously an extreme case but if you look hard enough you can find all kinds of wrong signals instead of the right ones!

Jamie

[Enzo]

My shop is a half mile from a local AM radio station transmitter. If I leave my probe sitting there, I can turn up the scope gain and easily watch the modulation envelope of this AM station. For some reason it impresses people when I show them that, then turn up the radio so they can hear the sound they are looking at.

[Sonicboom]

Okay, I finally had time to get back to this project. After more listening tests, something still appears to be off (besides me). So, I used an 8 ohm resistive load and measured 18 v p-p before the onset of clipping from a .35v p-p input signal. If I do my math right this time, I'm looking at ~5 watts. Huh?!!

Going back over what I did before, I replaced a bad 10M resistor, but I forgot that I also re-flowed a couple solder joints on one of the preamp tubes. I don't remember which tube though.

FWIW, other people I've asked on forums who have both VVs and Super Reverbs have said that the VVs should be fairly close in volume. Mine is not. The SR is definitely louder. Looks like it's back to the bench with it.