Help with Plexi clone hum

[framos]

Hi,

about an year ago I completed my first build, a plexi 50w clone (Ceriatone kit).

Since then, I have been using it almost exclusively (99% of the time) plugged to my Suhr Reactive load (SRL), for bedroom use.

Last weekend, after several months, I plugged into a real cabinet - probably for the third or fourth time ever - and noticed this hum. I really don´t remember hearing it before. On the SRL it´s barely audible.

The hum is present without any input, and is not affected by the volume control. It´s not too loud - in the video you can even hear my hands on the phone. It also does not get any louder when playing.

Any help would be extremely welcome!

Mybuild3.jpg

Ceriatone-69-Plexi50-lead.jpg

Thanks in advance,
Fabiano

[Phil_S]

It's 120 Hz, which means it is DC. Turn the amp on plugged into a real speaker cabinet and poke around with something non-conductive like a chopstick or plastic rod. You may find a wire that needs to move away from the output jacks and speaker selector, or maybe something near one of the tubes. It is hard to say what might help. I can't see anything obvious in the picture.

[martin manning]

You can also remove the tubes beginning at the input to help isolate the source. As Phil points out, the 120 Hz frequency says it’s coming from the HT ripple voltage, after the rectifier. Since the volume control doesn’t affect it, it must be entering after the stage fed by the volume control.

[pdf64]

It seems quite a buzzy hum, does the presence setting affect it?
That implementation of multiple 0V chassis points seems non-ideal but presumably it's been proven good.
So in the 1st instance, it may be a good idea to re-establish all the chassis 0V connections, ie slacken the fasteners off, rotate things a few degrees, and re-tighten.

[framos]

Thanks for the replies!

I am trying to recall any modifications I made to the amp since I built it (assuming the hum was not there when I built the amp). AFAIR I only replaced the bright channel coupling cap (0.022 TAD ->0.0022 Sozo) and the mains filter cap (32/32 -> 50/50).

Is it possible that they're the source of such 120hz hum?

[framos]

It seems quite a buzzy hum, does the presence setting affect it?

Just tested it and yes, turning presence all the way to zero seems to improve things. What does it mean??

So in the 1st instance, it may be a good idea to re-establish all the chassis 0V connections, ie slacken the fasteners off, rotate things a few degrees, and re-tighten.

Will try that!

[framos]

This morning I tested removing the preamp tubes (nothing connected to the amp).

With v1 (only) removed, the hum continues.

With v2 (only) removed, the hum stops. I tested 5 different spare preamp tubes I had (Mullards, chinese, sovtek...) and the hum continues, so I pretty positive it's not the tube.

[martin manning]

So now you can look for ways that the hum could be getting into the signal path, most likely via the grid leads going to V2. There is a gain stage and a DC coupled cathode follower there. You could try grounding the grid of the gain stage (pin 2) to see if it is getting in there. Does the bright volume affect the hum?

[framos]

Does the bright volume affect the hum?

Not really. It adds a touch of hiss when past 7 (which I think is fairly normal), but the fundamental hum remains unchanged.

[pdf64]

It seems quite a buzzy hum, does the presence setting affect it?

Just tested it and yes, turning presence all the way to zero seems to improve things. What does it mean??
...

Buzz always makes me think that that heaters are involved.
So look to the heater 0V reference, also eliminate it just being noisy tubes, eg maybe several have had their heater-cathode insulation compromised?

If not heaters, the the HT path with the nastiest harmonics is that between the reservoir - terminal and the HT CT, so check that is all good.

[R.G.]

+1 for pdf64's comment.

There is a subtle trap in the wiring for a power supply. The negative-side lead of the power rectifiers carries BIG pulses of current, big enough to cause audible hum in your amp if you don't connect that wire absolutely correctly. It appears that the wiring diagram you posted (thank you!) has this done in a way that will cause a buzzy hum.

If I were wiring that amp, I would move the wire from the "0" lug of the 325-0-325 winding on the power transformer from the chassis ground lug to the negative of the 50+50 capacitor. This one change will probably reduce the hum you're getting, as the wiring specified falls into the trap I mentioned above.

There are other details of the grounding in this amp that might cause problems, but this change will help a lot. Try it. It may be all you need.

[martin manning]

Or leave the 50-50 reservoir ground where it is, and ground the 32-32 screen/PI filter directly to the chassis, at the transformer bolt*, or perhaps better yet with the power tube cathodes.

*it would be better to have a dedicated bolt for this ground.

[tubeswell]

What R.G. and Martin said. And maybe reflow some of your soldering joints (see red circles). Can't tell from the pic if these have properly soldered connections or not, but they seem to have too little.

[framos]

You have a good eye, thanks!

I had already fixed the presence pot while doing a round of checking - and was beating myself enough for letting it pass..... now on to the treble.

[R.G.]

Or leave the 50-50 reservoir ground where it is, and ground the 32-32 screen/PI filter directly to the chassis, at the transformer bolt*, or perhaps better yet with the power tube cathodes.

That might work. You could probably tell that I wanted to tell him to move that negative lead on the 50+50 cap. But I restrained myself.

The theoretically correct thing to do is to isolate the charging pulses to the first filter cap on exactly one (1) wire from the transformer center tap to the negative terminal of the first filter cap. That is the wire from the "0" in the 325-0-325 winding. That absolutely must go to the negative terminal of the 50-50 cap to get the lowest noise from the charging pulses. If a wire goes first to a chassis ground as shown, then to the 50-50 negative, the charging pulses run to the chassis, then up to the 50-50. So the wire to the 50-50 cap negative has the same pulses running through it, meaning that its negative is separated from the chassis ground by the pulse currents times the wire resistance.

If you assume that there are pulses of maybe 1A height charging the 50-50 cap, those pulses come out of the negative side of the cap and must end up back at the HT center tap. As the wiring shows, those go first to the chassis ground from the 50-50 negative, then they head from chassis ground to the winding center tap. So the bottom end of the filter cap is sucked below chassis ground 120 times per second. A simple rearrangement of the wire going from the CT directly to the 50-50 cap negative means that the CT and the 50-50 negative are still pulled apart by the charging pulses times the wiring resistance, but since the 50-50 negative is tied to the chassis ground, none of the charging pulse induced voltage appears on ground. You can see this easily enough if you replace the wires (mentally) with 1 ohm resistors. Now the (presumed) 1A charging pulses cause 1V voltage pulses on each wire. If ground is in the middle of two "1 ohm" wires, you get 1V pulses on ground relative to the 50-50 cap, and this gets passed around the amp on the chassis ground. If there is a 1 ohm wire from CT to the 50-50 negative, there are still 1V pulses between the cap and the CT, but now ground is connected at the filter cap, so the pulses do not move the ground around, as that wire from the 50-50 cap to ground carries no pulse currents. From the rectifiers at least.

For lowest noise, the 325-0-325 center tap has to go to the negative of the 50-50 cap.

As for moving the negative sides of the 50-50 cap and 32-32 cap to ground somewhere, I have my normal quibbles about using chassis as a ground. I know that tying "ground" to chassis at various places produces changes to grounding hum and noise, and that some versions are OK-fine. But you can't find that out until you build the amp and start moving ground references on the chassis. I personally would make the chassis an RF shield and an electrical safety shield by tying a signal ground to the chassis in one and only one point. I would also not use the transformer bolt for any grounding at all, purely on safety standards practice. I would use one chassis bolt for a wire from signal ground, which would be otherwise insulated from the chassis. AC mains safety ground would have its own chassis ground bolt, used for nothing else; a safety lab would fail an amp using a transformer bolt. It's a PITA, but true.

Likewise, I would not take the 32-32 cap negative to the chassis; I'd put it on the signal ground point that collects return ground current and references, from places like the output tube cathodes, which I would not tie to chassis.

But that's just how I'd do it, and it's not my amp.