Help with Preamp/Amp hum

[lt_texan]

I haven't posted much lately as I'm been busy with other things, but I recently rebuilt my phono/pre in a new shielded chassis and have created a problem with hum. Any help would be very much appreciated.

The phono and pre are scratch builds around ECC83/phono and 12B4/pre. The phonostage has no noticeable hum. The pre does. The pre is transformer out with a Sowter 9705s 4:1 40H 30mA.

The amp is a WAD6550.

I've attached a simple drawing of the grounding. You can see I have no network at the transformer out. It goes directly to the RCA and the ground is earthed via a 10R signal lift in the amplifier chassis (in accordance with WAD instructions).

Out of the speakers, I'm getting about +25db's around 100-120Hz. I can measure 21 and 29mV AC at the amp speaker posts. (No noticeable difference if it is CD or Phono selected.)

Trying to track the hum down, I've found the following:

1. Shorting the interconnect neg at the preamp to ground causes even more hum. (Causes an earth loop?)
2. Amp with the inputs shorted to ground gives no hum (the amp is really quiet).
3. Measuring AC at the end of the disconnected interconnect (amp side) with the preamp on, I get almost nothing, 0.3mV.
4. Measuring AC at B+ prior to the pre output transformer is 0.2mV.
5. Shorting the grid on the 12B4 pre made no change in hum.

The phono/pre power supply is in a separate chassis with all grounding from the phono/pre chassis back to that.

I used a bus bar (heavy copper wire) in the pre with all the small signals at one end starting with the phonostage, then pre and then filtering (some final stages in the pre chassis) and chassis earth at the other end returning to the umbilical to the PS chassis.

Any ideas?

Thanks,
Dan

[lt_texan]

a little more accuracy on the attachment

[gootee]

Have you tried separating the small signal grounds and taking them back to the power supply ground with their own separate conductor?

AC and other time-varying ground-return currents induce voltages across the ground-return conductor's inductance and resistance. Those voltages appear back at the non-PSU-ground end of the ground-return conductor. Since that includes the ground reference point for your grid input resistor, those voltages will be arithmetically summed with your input signal!

That is why star grounding concepts should be used.

[lt_texan]

Swapped in my 2A3 monoblocks and....no hum.

Well, with my ear up against the speakers, a bit, but not noticeable.

Could that be just less power?

Oh, and no lift in the 2A3's.

[lt_texan]

Originally Posted by gootee

Have you tried separating the small signal grounds and taking them back to the power supply ground with their own separate conductor?

AC and other time-varying ground-return currents induce voltages across the ground-return conductor's inductance and resistance. Those voltages appear back at the non-PSU-ground end of the ground-return conductor. Since that includes the ground reference point for your grid input resistor, those voltages will be arithmetically summed with your input signal!

That is why star grounding concepts should be used.

I've always had good luck with using a bussbar approach (read in MJ 2nd edition, I think, don't see it in my 3rd ed) where you start at one end with the low current signals (grid leak) and work your way over to the powersupply filter caps and chassis ground.

[gootee]

Originally Posted by lt_texan

I've always had good luck with using a bussbar approach (read in MJ 2nd edition, I think, don't see it in my 3rd ed) where you start at one end with the low current signals (grid leak) and work your way over to the powersupply filter caps and chassis ground.

That would minimize the problem with the shared conductor length. But was that ever with a much longer shared ground-return conductor going back to an external PSU?

Anyway, why degrade performance when it's avoidable? Sometimes it might not be noticeable but it will always be creating a problem to some extent.

[lt_texan]

Originally Posted by gootee

That would minimize the problem with the shared conductor length. But was that ever with a much longer shared ground-return conductor going back to an external PSU?

great point!!!!!

[lt_texan]

Well changed the grounds around separating the later stage filter caps from the signal grounds in the preamp chassis. No joy.

When I wave the preamp chassis around the power supply chassis, I get lots of nice hum. Especially when the pre output transformers are close to the power transformers.

I get a nice reduction of hum by putting 100R across the output transformer out terminals.

Back to the drawing board.

[gootee]

Can you post some photos of the layout and wiring? Sounds like it must be a problem with enclosed loop area.

Did you separate the input (grid-to-ground) resistor's ground return from everything else, or just from the filter caps?

[lt_texan]

Got it.

The Q on my preamp power supply was all messed up.

The first stage filter is choke input and a cap. 20H and 100uF. The internal resistance of the choke is 182R. That all calculated out to Q=2.5

I saw on older notes I had 500R additional resistance in the snubber circuit that wasn't there now.

So I put in a 500R resistor before the choke and after the 220nF cap. After the choke, have another 220nF cap.

I also increased the capacitance to 200uF for good measure.

This brought Q down to 0.46, under the 0.5 limit that I understand.

No more hum!

Always something other than you think.

Thanks!