Homemade amp only hums with volume turned down

[elipsey]

Hi. It hums a bit with volume pot turned down, and the hum gradually dissapears as the pot approaches max. It's not that loud, but it bugs me.

It goes away completely when the master volume is turned down. It's still there when V1 is pulled, but it goes away when V2 (cathode follower pair) is pulled. I tried connecting the (shielded) lead straight from V2a grid to volume pot wiper to rule out the gain switch wiring, and also disconnected the bright switch wiring, but it still hums. I have tried messing with the lead dressing around the tubes in nearby stages. I have also checked the ground connection on the volume pot, and tried moving some of the ground bus around, and I'm kind of running out of ideas.

Any suggestions?

A schematic and pictures are attached. Thanks very much for your time.

[xtian]

Sounds like a groud loop, as you've guessed. In the photos, do I see some stuff grounded to a buss bar connected to backs of pots? Other grounds running to a star ground point? Still more going back to the reservoir cap ground?

Please describe how your grounds are wired, especially the cathode grounds for V1 and V2, the ground of the volume pot, and the grounds of all those shielded cables.

[elipsey]

Sounds like a groud loop, as you've guessed. In the photos, do I see some stuff grounded to a buss bar connected to backs of pots? Other grounds running to a star ground point? Still more going back to the reservoir cap ground?

Please describe how your grounds are wired, especially the cathode grounds for V1 and V2, the ground of the volume pot, and the grounds of all those shielded cables.

Thanks very much. I'll come back shortly to try to describe it better; if I don't finish cooking dinner my family will begin to make far more annoying noises than then one described here.

[elipsey]

Ok, I will begin by describing the grounding scheme in general, and then describe the ground bus that the volume pot is connected to.

There is a star ground near the PT. Connected directly to this star ground are the high voltage center tap, the ground of the cap can, the bias supply ground, the grounds from master volume pot and presence control, and two other ground busses which I will describe shortly. The star point is a terminal strip with all terminals soldered to a piece of bare copper wire. The terminal strip is bolted, and grounded, to one leg of the PT. My intention was to attach high current grounds here to keep ground currents out of the chassis.

I did use a virtual center tap for the filament supply, as no physical center tap was provided. The virtual CT is connected the chassis at another leg of the PT. The speaker jack is grounded directly to the OT but is not an isolated type.

There are two ground busses.

One is a stranded 20AWG wire soldered to the back of the pots -- you'll see in the pictures that the end connected to the main star ground has green insulation. Connected to this bus are the grounds from the isolated input jacks (blue wire), grounds from V1A and V1B catode resistors, the signal ground from the two 470k voltage divider resistors between V1B and the cathode follower (V2) input grid, and several sheild drain wires. All sheilds are grounded on one end only. The two shielded leads that connect the isolated effects loop jacks to the master volume wiper and PI input are grounded to the chassis at a preamp tube socket mounting bolt near the jacks. All other sheilds are grounded to the bus wire on the back of the pots.

Finally, a seperate bus grounds the V2 catodes as well as the 8uf preamp power supply decoupling cap directly to the main star ground. In the pictures you see a black PVC insulated wire connecting these from the (-) end of the cap on the perf board. While I was flailing around trying to quiet the amp down, I moved these off of the other bus, thinking that the later stages might need to be grounded seperately from the input jacks and V1 cathodes. The MV and presence cap are grounded seperately for the same reason -- you'll see one end of the .1uf Mallory cap hanging off the ground leg of the MV pot, and a black PVC wire connecting that to the main star point. All of these used to be on the first bus, but I can't really remember if I made things better or worse.

Wheh, I think that covers it. More pictures are attached.

[eschertron]

You've discounted heater-induced hum in V2? How are the tubes' heaters wired, ie, in what order from PT to end? Just a thought for your consideration.

[elipsey]

You've discounted heater-induced hum in V2? How are the tubes' heaters wired, ie, in what order from PT to end? Just a thought for your consideration.

Thanks. I wouldn't say I've discounted it, but I'm not sure how to test

The heaters are wired with 20AWG solid wire -- it's the green and white wire in the photo. Starting from the pilot light, the order of tubes on the filiment supply is: 1,2) power tubes, 3) PI, 4) cathode follower (V2), 5) V1a and b. The sixth tube is the rectifier, which is on a seperate 5V supply of course. The filiment supply wire is twisted together and laid flat into the rear corner of the chassis as far from the other socket wiring as I could get it. I have prodded at the socket lead dress, but I haven't found anything so far. V2 is the socket in the middle of the photo.

As an experiment, I also tried bypassing V2A cathode resistor with a 25uf cap but that just made things worse. I didn't want the gain anyway. (I'm not sure about trying to bypass the V2b, that's a lot of voltage?) I will try other tubes at V2, just in case.

Any other ideas to test this?

Thanks

[xtian]

Your buss wire, that's connected to the backs of the pots AND to the star ground, is a loop--it's connected to your chassis at multiple points (all the pots are making contact with the chassis).

[elipsey]

Your buss wire, that's connected to the backs of the pots AND to the star ground, is a loop--it's connected to your chassis at multiple points (all the pots are making contact with the chassis).

Right, I'll try seperating the pots from the bus. Thanks

[big_teee]

As mentioned, grounding has numerous ground loops.
If you have you shielded leads grounded on both ends, that causes additional ground loops.
Shields are normally only grounded on one appropriate end, that way a noise signal can't travel across the grounded shield.
T

[eschertron]

As an experiment, I also tried bypassing V2A cathode resistor with a 25uf cap but that just made things worse.

That's telling in that the more strongly coupled V2 cathode is to ground, the noisier. You're on the right track.

edit: what happens if you add a 1M (or less) resistor from the grid of the valve right to ground at the same point the cathode resistor sits? It looks to me that the other grid leak paths may be 'far away' on the ground bus.

[elipsey]

That's telling in that the more strongly coupled V2 cathode is to ground, the noisier. You're on the right track.

edit: what happens if you add a 1M (or less) resistor from the grid of the valve right to ground at the same point the cathode resistor sits? It looks to me that the other grid leak paths may be 'far away' on the ground bus.

Nice. I'll try it before I take the bus apart.

[elipsey]

Nice. I'll try it before I take the bus apart.

Ok. I soldered one end of an 820k resistor volume pot wiper, which is connected directly to the V2A grid. I tried clipping the other end of the resistor on and off from the ground of the V2 cathode resistors with a test lead, while the amp was on (carefully ), with volume turned down. This had no discernible effect on the hum.

[Steve A.]

Here is the schematic from Post #1...

[elipsey]

OK, here's what I did:




The volume pot is out in the pictures, but I tried it with everything put back together, and it worked great.

I looked at Ceriatone builds and tried to imitate their grounding, more or less. I built this amp to try out some different circuits and I have been hacking on it for a long time, so things got kind of messy. I'm pleasantly suprised at how low the noise floor is now. It still hums a little when switched to high gain mode, which is similar to a jcm800 preamp circuit. It's maybe a little noisier in that case than I would expect from a modern comercial high gain amp, but it doesn't seem excessive. Maybe there's a limit to how quiet you can make a circuit like that when it's point to point. I have never played a real 800, so I'm not really sure what's normal.

It's a tad quieter with V1B bypassed, but barely. I can compensate for the extra volume by adjusting the voltage divider after that stage. In general, should I expect any difference in tone from bypassing a cold cathode stage, other than more volume into the next stage?

Any other ideas about quiting this thing down just a bit more?