20 millivolts on chassis

[jan.didden]

Can you confirm that it really is a hum and not a buzz?
A buzz has the same mains frequency but has some added higher harmonics and sounds more 'ragged'. Hum sounds 'clean' and 'round', sort of.

This is important, because if it is buzz it is caused by a circuit detail after the rectified supply.
Hum is caused by mains interference from cabling or defective screening but not from the supply.

jan didden

[DF96]

Originally Posted by chokesrule

I am measuring dc voltages

I assumed AC. Ignore everything I said.

[chokesrule]

I would guess its a buzz. I can't do it right now but I will build a simpler heater supply shown on the Valve Wizard site see what happens. If I turn on the heater circuit with just the amp on at full vol. As the heaters warm the buzz grows. The cabling is heavily shielded mic wire

[jneutron]

Originally Posted by DF96

I assumed AC. Ignore everything I said.

Not necessarily. His DVM may be picking up an asymmetrical magfield and sees some DC in it. Half wave and full wave current pulses can do that.

The grounding system seems really complex from the description.

chokesrule...any chance of a drawing or pic? Also, what is the AC voltage from point to point?

Each box should have only 1 path to safety ground.

To accurately measure the ac voltage from the chassis to the wall socket ground, you have to wrap the dvm wire around the line cord the entire length, and it cannot have the same twist pitch as the internal wires of the PC.

jn

[chokesrule]

I got rid of the dc in the chassis. Zero. The separate chassis earths which ran directly to ground and the star I sent them only to the star . In the morgan Jones schematic (page 374)theres an earth point just after the common mode choke. I sent these to the star before I just sent them to the mains. This killed all nearly all the hum in the left channel , and only a little in the right so obviously I have another problem there. I think I going to rewire the whole thing put the star where the mains earth is. Its the best schematic I draw up , ask me to draw any more detail I will end up cutting an ear off.

[chokesrule]

been messing this evening. The reason why I thought I had faint buzz and then thought I had much more when I originally listened was because the first time I had the interconnects unplugged and the second time: plugged. Although the schematic shows a onboard dac in the chassis :for testing right now, I have a cd player running. By rerouting some of the earth routes to the star I had indeed reduced the hum quite well. BUT!!!! one of the rca plugs was not isolated from the chassis and THAT is why I had one channel worse than the other. I dont now if it is possible to erradicate all hum/buzz from a valve output stage. You can only hear it at max. When do i ever have it at max?So is possible? I have no voltage reading when I test the chassis

[DF96]

Originally Posted by chokesrule

I dont now if it is possible to erradicate all hum/buzz from a valve output stage.

The strict answer is no, you can't eradicate all hum from any circuit. The practical answer is yes, you can reduce hum so that it is not normally audible. You need a combination of the right circuits and the right implementation (e.g. grounding details).

[chokesrule]

I have been further experimenting with earths and have had a massive result. I connected the shielding of the interconnects between the cd and output/tripath stage chassis and at full volume as far as I can make out nearly all the hum/buzz has vanished and all I hear without straining is white noise the hum/buzz was masking it. Glad I did not just settle for what I had in my last post. BUT! now how do I kill white noise???is there ever an end??

[DF96]

White noise comes mainly from active devices and resistances. Some (thermal noise) is almost unavoidable but can be reduced by appropriate design. Other (e.g. excess noise, flicker noise, shot noise) can be reduced by careful choice of components (e.g. resistor type). Thermal noise can be calculated; the others may have to be measured or estimated/guessed.

[AndrewT]

White noise and similar will be exaggerated if you have too much gain in the system.

If you find that you need lots of attenuation to avoid clipping of the output then you probably have too much gain.

read:
What is Gain Structure? - diyAudio