Hum Reduction - Heater DC offset

[Diablo]

My old Capitol amp has a little more hum than I'd like, and I've isolated it to the power amp section. There's no schematic, but the PT has two sets of filament heaters, one for the power amp and one for the preamp, and it's a quad of 7591 tubes in push-pull cathode bias. The preamp uses a balancing rheostat between the filaments to minimize hum. The poweramp filaments are DC offset by connection to one pair of output tubes' cathode bias resistor. There is no center tap for the heaters on the PT. Would I likely be better off adding a pair of 100 ohm resistors from the heaters to ground or to the cathode bias resistor? The amp is point to point wired and I don't think there's an easy way to change the ground scheme.

[Enzo]

Why not determine where the hum is coming from? If the hum is power supply ripple, messing with the heaters will have no effect. Since the amp is push pull, remove the phase invereter tube. Hum remain or go? If it goes away, then the powr stage is innocent. If it stays, then how much ripple is on the screen node?

If the power stage is the hum source, and it is cathode bias, then we have no bias supply to fail. But the tubes are suspect. Don;t worry about the impedance, but yank two of the power tubes, one from push and one from pull. Does the remaining pair still have the same hum? Now remove those two and install teh ones we yanked first. Does that pair also hum?

And you know what? Pull all four power tubes. is ther still any hum?


And I should have asked first, is the hum 60Hz or 120Hz?


And as to your question, no reason you couldn;t add a couple resistors, I'd leave the cathode voltage offset there. My vote is it probably won;t change anything, but who knows.

[Diablo]

Why not determine where the hum is coming from? If the hum is power supply ripple, messing with the heaters will have no effect. Since the amp is push pull, remove the phase invereter tube. Hum remain or go? If it goes away, then the powr stage is innocent. If it stays, then how much ripple is on the screen node?

If the power stage is the hum source, and it is cathode bias, then we have no bias supply to fail. But the tubes are suspect. Don;t worry about the impedance, but yank two of the power tubes, one from push and one from pull. Does the remaining pair still have the same hum? Now remove those two and install teh ones we yanked first. Does that pair also hum?

And you know what? Pull all four power tubes. is ther still any hum?


And I should have asked first, is the hum 60Hz or 120Hz?


And as to your question, no reason you couldn;t add a couple resistors, I'd leave the cathode voltage offset there. My vote is it probably won;t change anything, but who knows.

It's 120 Hz hum, and it occurs with no instrument plugged in and is not effected by the volume controls. The amp is stereo - two PIs, two OTs, with a pair of tubes for each output. The left side hum is as loud as the right side hum. Pulling either PI reduces the hum. I substituted new PI tubes and the hum loudness is the same as the old tubes. If I pull all the output tubes, the hum goes away. I substituted new output tubes, and the hum is the same as with old output tubes. The ripple on the plates is 5.7 volts. The ripple on the screens is 0.21 volts on the left side and 0.34 volts on the right side. The amp has all new e-caps. I tried jumpering in two 40 mfd caps in parallel with the first filter, and there was no change in the hum. Originally the amp came with a fried GZ34, so I'm now using a Weber copper cap 5U4G (solid state rectifer + resistor + inrush current limiter). I chose the 5U4G to drop a little plate voltage. I added the two 100 ohm balancing resistors for the power amp heater filaments and connected them to one cathode bias resistor - You were right Enzo.....it made no difference to the hum. The right side power amp uses a long shielded cable from the preamp to the PI. Moving the cable around effected the hum, so I replaced the cable - the hum diminished. I chopsticked the wires and couldn't find anything suspicious. The amp is quieter, maybe I'm just being too picky.

[Enzo]

Ah, stereo, wasn't thinking that.

120Hz hum comes from power supply ripple, usually ground return currents sharing copper with signal grounds.

If pulling PIs drops hum, that's a clue, does pulling both end it?

It may be normal, but maybe still cureable. Where is the PT HV CT grounded? And the main filter caps? Those should be grounded together. SOmething as innocent as a 3" hunk of wire in a ground return can matter. For example, if the CT wire goes to a terminal strip, and a wire from ther to chassis, and the circuit ground goes to that terminal instead of chassis... that is all it takes.

Push pull stages will tend to cancel out B+ ripple, so pulling the PI isolates them to find out. But B+ is single ended in the rest of the amp, no cancellation. SO where the circuits from the PI on back get their ground matters a lot.

[Diablo]

Thanks Enzo, I'll take another look at my ground scheme. The PT CT is grounded at the rectifier tube socket metal base. The main filter caps are grounded to the chassis about 3 inches away from the rectifier tube socket. Originally the amp had two multi-cap cans for the main, screen, and PI filters. I recapped the amp with individual caps on terminal strips with the grounds within an inch of the original multicap chassis ground. The main, screen, and PI filter caps all share the same ground point on the terminal strip and then to the chassis. I added a 16 gauge wire between the terminal strip ground and the rectifier socket ground - but that didn't knock down any of the hum. For the most part, I'm using the same ground path that was part of the original amp design. It's point to point wired with various terminal strips grounded locally to the chassis. I have no schematic available and there is a separate preamp chassis and power amp chassis. Two shielded rca cables join the preamp to the two PIs in the power amp. There is also a multiple wire cable between the two chassis that includes the preamp B+ and 6.3 volt heater wires and 120 volts for the on-off switch. The preamp also has three filter caps, which are grounded locally on terminal strips and then to the preamp chassis. I'm not sure, but I think the multi wire cable may include a ground wire between the two chassis.

[jrfrond]

It would seem to me that the grounding scheme for that particular amp was well thought out by whomever debugged the original design. "Floating" the heater CT and using separate ground schemes for pre and power tubes was probably NOT part of the original design, and was probably done so that the two heater systems wouldn't share ground currents, as Enzo said.

Grounding and hum-tracing/removal is the bitchiest part of amp design. My guess is that, due to the more complex ground design, the amp has a problem NOT related to heater grounding.

You said "more hum than you'd like". Do you have any idea what the original S/N specs were? How many dB down is the hum? Unless you have some concrete figures, you could just be chasing your tail.

[Diablo]

You said "more hum than you'd like". Do you have any idea what the original S/N specs were? How many dB down is the hum? Unless you have some concrete figures, you could just be chasing your tail.

Actually I have reduced the hum by replacing the shielded cable to one side PI, and the amp is quiet enough. I don't have any S/N specs on this amp. I can't find ANY information on this amp. No schematics, no references on the internet to this company. I think it was built by Lectrolab in 1961, but I haven't seen any reference to this particular model.

To answer Enzo's question - I pulled both PIs and the hum is reduced a little. Pulling the rectifier kills all hum. Pulling the output tubes kills all hum.

[Diablo]

Last night I changed most of the power amp over to star grounding. I separated the ground connections for the main filter caps, screen caps, and PI caps on the terminal strips, and eliminated the two local grounds to the chassis. I ran separate 18 gauge wires from each cap to a star ground near the rectifier tube. I also disconnected the local chassis grounds for the two cathode bias resistors, and ran separate 18 gauge wires over to the star ground. I connected the PT center tap to the star ground. The two PI circuits still use two local chassis grounds, each one is about 8 inches away from the star ground. The speaker jacks are also locally grounded on opposite ends of the chassis about 9 inches away from the star ground. . The experiment didn't make the amp any quieter or any noisier. I can live with the small amount of remaining hum.

[txstrat]

I've got a (ground-)hum issue in my latest B15 copy. While searching the web I've learned the PI draws a good amount of current.
It might be of help to ground the PI at the star ground too. Wouldn't hurt to try IMHO.
Just my $.02

[Alex R]

Could just be unbalanced tubes (I mean the Capitol amp) - some push-pull amps are very sensitive to this. A bias balance pot would help (but it's cathode bias...). With amps of the AC30 type I often sit with a box of tubes plugging them in to an empty position and listening as the amp approaches balance and the hum reduces. Does the second/fourth tube fail to remove the hum completely? If so you need one that draws a bit more. Or does it first remove the hum... and then pull the amp over into imbalance on the other side and put a bit of hum back again? If so, find a tube that draws a bit less.

[Diablo]

Thanks Alex for the suggestion. I'll try to move some tubes around. Unfortunately, I don't have a bunch of 7591 tubes available for matching.

[Alex R]

Well you have four, so work on one side and you've three to swap in. You'll know soon enough if matching is your problem; plug in a second tube that pulls more current than the one that's in there working and you'll hear the hum go down to nothing before it comes back up a bit. You'll be able to grade the tubes 1-2-3-4 for current draw and match them as close as is possible.