Stereo SE EL84 amp hum

[jkey04]

I have been working through a new amp design that incorporates two single-ended EL84 amps in one amp chassis (stereo).

The amp is complete and sounds really great, but it has a slight 60HZ hum when the standby switch in switched into play mode. The hum comes from both SE amps, but only if the power tubes are installed. Pull both power tubes = no hum. If the power and standby switches are switched on at the same time, it takes several seconds for the hum the come in as the amp warms. Maybe this proves a power supply issue? I know SE amps inherently hum more so than a push-pull amp. I have built several 5F2 style amps that were almost dead quiet compared to this EL84 based SE amp pair.

I have restructured my grounding scheme to no avail.
I have rotated one of the OTs slightly to no avail.

In my troubleshooting, I noticed there is 2.5VAC/60HZ on the first filter cap. Which translates to about .25VAC/60HZ on the secondary side of the OT (speaker jack). This seems to be the hum issue. The question is: how to fix this? I tried 3 different GZ34 tubes with no luck.

Two ideas I am planning on trying next:
1 - Put 1N4007 diodes in series between the PT and the GZ34.
2 - Flip phase of heaters on one of the SE amps.

[Colossal]

There is no common mode rejection in SE, so you need a very clean power supply and good grounding. To that end, I employ a CLC filter BEFORE B+1. Meaning, rectifier, then CLC, then your OT gets fed from that last C. Also, elevate your heaters to DC.

[Stevem]

X2 with the above!

How much AC ripple is there on the plates of the outputs and the screen power supply node?

[didit]

Hello -

Proposed "pi" filter at the head of power rail works. A good solid-state voltage regulator design could work even better. Search out HiFi SE power supplies.

Best .. Ian

[jkey04]

Thanks for the replies!

So the power supply i’m using is essentially taken from the Trainwreck rocket schematic with slightly less filtering. CLC (60uF/Choke/30uF) with the B+1 on the 30uF cap.

Also, I have tried elevating the heater center tap to the cathode of one of the EL84s. That didn’t help any..

Stevem - I’m not sure how much ripple is on the plates. I’ll have to check. It might be a day or two before I can get this amp back on the bench. I set it aside to move to something else and try to clear my head from it before I start in on it again.

I’ll look into a SS voltage regulator. Thanks!

[Colossal]

So the power supply i’m using is essentially taken from the Trainwreck rocket schematic with slightly less filtering. CLC (60uF/Choke/30uF) with the B+1 on the 30uF cap.

If you are now feeding the OT center tap off that 30uF, the ripple and subsequent hum should be GREATLY reduced from that 2.5VAC on the 60uF cap. By comparison, I had a SE amp, SS rectified into a 16uF - 10H - 8uF CLC filter. The ripple post-rectifier was 9.5VAC. After the choke and on the second cap, the ripple was 330mVAC, a 29x reduction. The preamp was a single 6SJ7 into a SE KT66 power amp. That amp was very, very quiet. The ground system was a single bus with all preamp nodes (only one, really) returning to its B+ cap and only one connection made to the chassis at the input. The rectifier loop was kept physically small and on one side of the chassis.

If you made the change to B+1 coming off that 30uF and you still have a noisy amp, consider the path of your rectifier and AC wiring. You want that loop small and as far from any signal leads as possible. Church and State: SEPARATED.

[jkey04]

That makes total sense! I’ll give it a try and report back. Thanks

[jkey04]

So Colossal’s suggestion worked. Moving the B+1 node post choke killed the hum!

My issue now is that my plate voltage on the EL84s is down around 250VDC and my output power and volume is lower than I want it to be. I need to put my scope on it now and see where it is and monkey around with ways to increase power with the existing PT I am using.

Thanks for the help!

[wpaulvogel]

What spec power transformer are you using? I’m surprised that the b+ dropped but I didn’t read what it was initially.

[jkey04]

I’m using a Mercury Mag Princeton Reverb PT with lower B+. (250-0-250)

The choke is a MM Vox AC15 choke. Maybe it has too much DC impedance? It’s dropping from 274vdc to 246vdc through the choke from B+ node 1 to B+ node 2

[Stevem]

That's a big drop in voltage!
A max of 10 volts is more like it.

[Colossal]

So Colossal’s suggestion worked. Moving the B+1 node post choke killed the hum!

My issue now is that my plate voltage on the EL84s is down around 250VDC and my output power and volume is lower than I want it to be. I need to put my scope on it now and see where it is and monkey around with ways to increase power with the existing PT I am using.

Thanks for the help!

I agree with Steve; that voltage drop is a bit bigger than expected. What choke are you using (inductance and DC resistance please)? Also, have you measured the current across your bias resistor? What is the primary on your OT? 250VDC and 250VDC plates into a 5k2 primary biased at 135Ω (~48mA) is right off the datasheet for a SE EL84 and will put out about 5.5W.

[TUBEDUDE]

The choke is now carrying much more current than it was supplying for the screen and the preamp tubes. A lower resistance choke should get Ep back in the ballpark.

[jkey04]

The choke I was using was a Vox AC15 replacement. I swapped it for a Fender vibrolux choke and now I have 270vdc pre-choke and 260vdc post-choke. With the plates and screens fed from same node (post-choke) separated by 100R 5W resistor.

I’m using two SE Mercury Magnetic OT’s that are 9k on the primary.

[didit]

Hello -

Some specific goals for this amp? Did you select output primary roughly 2x spec for a reason?

Vibrolux choke underspec'ed at ~50mA, whereas that duo of EL84 SE will idle at around 100mA. Hammond's 158Q or similar?

Best .. Ian