Please help me in getting rid of the buzz.

[keithgreenhalgh]

I doubt it a problem with the power transformer. Try pulling the 6L6 tubes and tell me if you still have hum.I have more questions for you but try this first.

[rdk845]

I removed the 12AX7 and the buzz disappeared. But with 12ax7 there and the buzz was there even when input was grounded.

Could anyone let me know what is causing buzz from this?

Thanks!

[rdk845]

Originally posted by keithgreenhalgh
I doubt it a problem with the power transformer. Try pulling the 6L6 tubes and tell me if you still have hum.I have more questions for you but try this first.

Without the 6l6g and with 12Ax7 in, the buzz was still there but became very very faint.

Will look forward to your input.

Thanks!

[keithgreenhalgh]

The 100ufd 400volt cap conected to the 12ax7 would have about 180-200 volts on it but you said you had a 16volt cap in series with a resistor to gnd. this doesn't make sense because the cap would fail. The 100ufd 400volt cap is what gives the plate of the 12ax7 a pure d.c. voltage. without this capacitator the 12ax7 would amplifiy the hum from the power supply.Check the wiring again and If you have a voltmeter ,measure the plate voltage. It should be 180-200 volt range.Any capacitor connected to gnd generally allows a 100% greater working voltage so it should be rated at 400volts and 100ufd sounds about right for the value.

[SY]

How are the heaters referenced to ground? Are they biased up or (as the schematic suggests) just left to float? Floating heaters = bad idea.

Put a scope on the input tube's rails and see what your ripple looks like.

[rdk845]

You are right, my mistake. The 100/16V cap is actually connected to the cathode of the 12AX7 not to the plate side.

So on the actual board, the cathod of 12AX7 is connected to 91K feedback capacitor as well as 1K resistor and 100/16V cap. Both of 1Kand 100/16/v are connected to ground through 1 100 ohm resistor.

The plate voltage for 12AX7 is 147V.

12AX7 plates of both channels are connected to single set of 100 / 400V and 10K ohm resistor.

Now the buzz is mostly on the left side. Right side is quite faint.

Will wait for your reply.

[rdk845]

Originally posted by SY
How are the heaters referenced to ground? Are they biased up or (as the schematic suggests) just left to float? Floating heaters = bad idea.

Put a scope on the input tube's rails and see what your ripple looks like.

The schemtic omits this but the one side of 6.3V secondary is connected to the ground. The 10V rectifier circuit for 845 heater is referenced to ground via 100 ohm hum pot and 1 ohm cathode resistor.

Wish I had a scope. May be some day....

[SY]

Now the buzz is mostly on the left side. Right side is quite faint.

What did you change to cause this? It would seem that you've eliminated supply ripple as the principal culprit.

The schematic omits this but the one side of 6.3V secondary is connected to the ground.

That is the single worst way to do it. You can greatly improve things (and maybe cure your buzz?) by disconnecting that ground, connecting two equal resistors of roughly 100 ohms in series, then put the seriesed resistors across the heater winding. Their junction forms a virtual centertap. That centertap can then be connected to a +50V or so reference, formed by connecting an appropriately sized voltage divider across the B+, and bypassing the lower leg to ground.

You also want to have the heaters grounded against common-mode AC or RF by bypassing each end of the heater to the chassis through a ceramic disc cap. At minimum, do this at the input tube.

This sounds fussy, and it is, but this can make an obvious sonic difference.

[rdk845]

Originally posted by SY

What did you change to cause this? It would seem that you've eliminated supply ripple as the principal culprit.

The amp did it itself. It appears the buzz changes over time as amp is turned on and off.

Originally posted by SY

You also want to have the heaters grounded against common-mode AC or RF by bypassing each end of the heater to the chassis through a ceramic disc cap. At minimum, do this at the input tube.

By bypassing you mean, one end of ceramic cap to the chassis and the other to an end of 6.3 winding, right? So this must be done after removing the ground from 6.3 winding? What kind of capacitance value the ceramic disc cap should have?

Originally posted by SY

connecting two equal resistors of roughly 100 ohms in series, then put the seriesed resistors across the heater winding. Their junction forms a virtual centertap. That centertap can then be connected to a +50V or so reference, formed by connecting an appropriately sized voltage divider across the B+, and bypassing the lower leg to ground.

I¡¯m not sure how this works, but it seems to float the heater circuit at 50V DC from ground. Could you explain why this is better?

Thanks.

[SY]

Does the buzz vary if you change the 12AX7?

Biasing up the heaters helps alleviate leakage between the heater and the cathode. There's a really complete and understandable explanation in Morgan Jones's "Valve Amplifiers," 3rd edition, which you ought to have.

What's important and nearly always neglected is to take care of common mode noise, which is the point of those ceramic caps. I'd put them right at the tube sockets.