Magnatone Noisy PI

[Diablo]

I picked up a Magnatone M13 amp, and it has a little too much hiss and hum for my taste. I grounded out the grids on all the preamp tubes in sequence, and traced the excess noise to the phase inverter tube (12DW7). I also grounded out the grids on the EL34s by using a capacitor to ground. I swapped out the PI tube, and there's no change in the noise floor. All of the resistors are carbon comps and original to the amp, built around 1965. So my plan is to swap out the carbon comp resistors on the PI socket. Is there any clever way to test which resistor is creating the noise, so I can just replace that one? Or should I take the shotgun approach and just replace all of them?

http://www.magnatoneamps.com/schemat...natone_m13.jpg

[g1]

Is there any clever way to test which resistor is creating the noise, so I can just replace that one?

Freeze spray or a heat gun might help identify which particular resistor is creating the noise.

[dwmorrin]

If ac grounding the grids of the power tubes didn't make the noise go away (they were grounded one at a time?), then that makes me think either the 270k on V6B or the 1M on V6A are the most likely candidates.

If a resistor is noisy in the plate circuit, will that noise show up in the cathode circuit? I've never checked for that before...

[Diablo]

If ac grounding the grids of the power tubes didn't make the noise go away (they were grounded one at a time?), then that makes me think either the 270k on V6B or the 1M on V6A are the most likely candidates.

If a resistor is noisy in the plate circuit, will that noise show up in the cathode circuit? I've never checked for that before...

I did the ac grounding of both power tube grids at the same time, and that did make the noise go away. Then I removed the ac ground from the power tubes and grounded out the grid of the PI, and that also made the noise go away. Grounding out the grids of all the other preamp tubes didn't make the noise go away. So, I concluded that the noise is in the PI tube circuit. Anyway, I took the shotgun approach, and replaced all the resistors and e-cap o the PI socket. Seems a little better, but not perfect. Unfortunately, this amp has a lot of local grounds on the chassis, and maybe that is raising the noise floor?

[ReadyTeddy]

Did you try individually grounding the grids of U6 with a capacitor? I note that the second half of the tube is a cathodyne PI whose gain is one. Because of low gain, they are not known for being noisy, and have often been used in hifi amps, where low noise is a necessity. The first half of the tube is a high-gain stage, which might seem like the more likely source of noise. This stage also has high-value resistors in its grid circuit that will contribute their Johnson noise, even if they are perfect resistors.

[Diablo]

Did you try individually grounding the grids of U6 with a capacitor? I note that the second half of the tube is a cathodyne PI whose gain is one. Because of low gain, they are not known for being noisy, and have often been used in hifi amps, where low noise is a necessity. The first half of the tube is a high-gain stage, which might seem like the more likely source of noise. This stage also has high-value resistors in its grid circuit that will contribute their Johnson noise, even if they are perfect resistors.

Yes, I grounded the individual grids of V6 and I think you're right about the 1st triode being the major noise maker since it's the high gain stage with those high-value resistors.

I have to stop working on this amp today because I'm burnt out on it. I replaced all the e-caps, figuring that could only help, since they're 48 years old. Then I turn the amp on and I have no volume! Try channel 1, 2, and 3 - still nada. I re-inspect all my work, check continuity on every solder joint I made, trace the wiring, check voltage on every tube. Run to Radio Shack to get some new test lead clips after one of mine breaks. Do pop tests on the grids of each tube, and it makes noises. Plug my guitar back in and discover that I'm an idiot - I had the guitar volume turned all the way down. Nothing wrong with the amp, and I wasted 3 hours troubleshooting it....

[Chuck H]

Happens to all of us sometimes! I once lost a sale on a used Bessa Moogie amp because I depressed the "mute" switch hidden under the rear chassis while bringing it into my living room... "I don't understand? It was just working a second ago!" I had done numerous mods to the amp and knew the switch was there. It just hadn't occurred to me at the time since I never used it and it was well hidden.

[Diablo]

Thanks Chuck for the sympathy. Anyway, I was poking around in the amp to see if I could make the reverb a little quieter, and found that there's 120 V - dc on the grids (2&7) of the reverb driver tube (12AU7 - Valve V10 on the schematic). I thought that the coupling cap from the plate of V9B might be shorted, so I pulled V10, and there's no voltage now on pins 2&7 of the socket. So that tells me that the 120 volts must be from the grid charging from the flow of electrons inside the tube. Is this normal at idle?

[dwmorrin]

Wouldn't that mean you'd need more than 120V at the cathode... meaning that you'd have ~450mA current there? (121V/270Ω cathode resistor)
My guess is that would be abnormal. Try changing the tube and check the voltages again.

[Diablo]

Wouldn't that mean you'd need more than 120V at the cathode... meaning that you'd have ~450mA current there? (121V/270Ω cathode resistor)
My guess is that would be abnormal. Try changing the tube and check the voltages again.

Yes, I tried a brand new 5814 tube (industrial 12AU7) thinking the tube must be shorted internally.
Funny thing...it also measured 125 volts on the grids, and the reverb works well with either tube.

[dwmorrin]

And what about the other terminals of the tube? What do they read?

[g1]

^^^^^^What he said. What is the voltage at the cathode? Then measure the voltage from cathode to grid and do the math. Measuring from grid to ground can often give false readings.

[Diablo]

This amp is haunted. I go back in and measure all the resistors on V10, and they all are reasonably close to face values. Then I measure voltage drop across the across the 270 ohm cathode bias resistor and get 6.7 volts. I recheck the grids (pin 2&7) and now measure 0 volts to ground.

One other fact. When I first measured pin 2&7 to ground and measured 125 volts, the amp would buzz loudly when I connected the ground.
And the reverb is still working just fine.

I measured the grid to cathode voltage as -6.2 volts.
I measured the plate to cathode voltage as 247 volts and calculate 6.13 watts dissipation on the plates. Looks like the tube is pushed pretty hard - 3.03 watts is the absolute maximum for each plate, according to the tube spec sheet for a 5814.

[Diablo]

Another update on trying to make the reverb quieter. Using the foot pedal to ground out the reverb, makes the hum disappear. I tried moving the pan away from the amp, but that didn't do anything. I made a sheet steel bottom shield for the pan, and that didn't do anything. I made up a new set of reverb cables with better shielding, and that didn't do anything. I plugged the reverb output cable into a new pan, sitting away from the amp, and that didn't do anything. I took another crack at moving wires with chopsticks near the reverb send and return connections in the amp, and eureka, I could make the hum louder and quieter. The problem turned out to be the 120V-AC wiring that runs from the main amp chassis plug to the on/off switch and the safety interlock switch. The wires weren't twisted and run too close to the final filter capacitor for the high voltage that supplies the reverb circuit. I had to cut the wires and splice in some extra length so I could twist them and reroute away from the filter cap. I would call it a bad Magnatone layout, to run the AC line voltage near the sensitive reverb return circuit.

[Chuck H]

Nice one! Sorry it took such a rundown of the more common stuff first. But good catch ultimately

I have a buzzy reverb to troubleshoot when I can get to it. Customer never turns the reverb up to where it's a problem, but it's a problem that should be fixed anyway. One of my builds too I'm expecting the same sort of trouble running it down.