Weird transformer...?!

[6267]

A friend dropped off a "Musicale Coronet" 1953 record player. I removed all of the very rusty and drifted internal circuitry (it's a tight fit, makes the filmosound I did last month look easy) to build a nice little blues amp out of it. It used a 6x5 recto, which sounded like bad news in a music amp that gets banged around, so I redid the circuit with a solid state full wave into 100uf. The amp powered up with the new design ok (redid power and PI so far, replacing everything) but with "objectionable" hum (120hz). It plays music well, but during silent passages or when the input is grounded, it hums. Same with PI pulled, stops when power tubes pulled.

So I started poking around in the power supply, and I noticed that the resistance between B+ AC leads and ground is very, very low. There is no center tap, but they were using the 6x5 full wave. IS the transformer itself being used as the center tap? It's not shorted (I believe) because I'm getting around 400v out, which seems reasonable. I also measure an AC offset between chassis and myself (24v) and a 7.5v between chassis/transformer and ground. What's the deal? Is the transformer itself the B+ center tap? Both B+ leads went to 6x5 plates in the original, but no center tap was present. Weird. Any ideas out there?

I should note that the resistance is still low with transformer out of circuit. Thanks for any help you folks might have!

[Enzo]

That transformer was made specifically for that application, as opposed to some Hammond transformer on a shelf for general purpose applications. I would not be surprised if there was an internal center tap grounded to the frame. Sometimes you can see them - a little enamelled wire sneaking out from the inside and welded to the frame outside.

besides, it works, so accept the magic.


24v between chassis and yourself? Well that is a new one on me. You mean like just holding a meter probe you get a reading? Sorta like touching a guitar cord plug makes hum.

That recto tube was sufficient when it was playing records, why won;t it work with guitar? Replacing it with diodes will raise your B+. Was it at 400v before the change?


You have not described the amp. Push pull or single ended? SIngle ended amps are WAAAYYY more likely to hum, while the natural cancellation in PP stages reduced hum there. Oh wait, phase inverter, never mind. If this hum is there without the PI tube, then it is coming from the power stage. Seriously out of balance PP stages will hum. Any unwanted DC on power tube grids - as in leaky coupling caps? Got same B+ voltages on but sides on plates AND on screens?

[6267]

Ok, so the circuit is essentially a williamson 6sn7 dc-coupled cathodyne with half of the sn7 driving. There's more to the story, but not built yet . All of the caps, PS and coupling, are brand new nichicons. The same tubes do not hum in other amps with similar output stages.

The amp *does* have cathode feedback, which is a new one for me. The cathode winding is center tapped, with a rheostat currently on the center tap to ground. Each leg of the cathode winding goes to the cathodes of the output tube. I haven't tried ignoring the cathode feedback circuit yet, so I may try that, or reversing phase. The amp sounds really good other than the hum, really good 50s clean sounds, i.e. warm transients and good mids.

I have other PTs, if I have to use them, but I'd like to avoid that. It's actually about 315 loaded, which is ideal for what I'm trying to do. I've had SS rectified amps on the bench though that hummed a good deal less with less filtering, so I figure there's something more to the story here. If my williamson splitter was say not getting enough b+ (about 289 before anode loads) would the amp end up "single ended" enough to hum badly?

I guess my worry is that either the PT is leaking B+AC through a large resistance to the chassis, or that whatever is on the chassis may be causing ground to be a variable thing, inducing the hum. The amp is star grounded (entirely, except the PS caps, ship in a bottle indeed) to the chassis at the input.

If it doesn't sound like it'll be deadly, I'll be ok with it so long as the chassis is at ~ 0v compared to the guitarist.

[Enzo]

The tubes and caps may be OK, but is the circuit?

HOW much ripple is on B+?

Are the grids where they need to be?

Are both sides pulling similar currents?


A dual diode B+ supply must have a center tapped transformer winding. Without one there would be no voltage to ground. A simple winding would require four diodes in bridge.

I don;t follow how reversing phase on the NFB would cure HUM. SPin the CT pot back and forth, does it affect the hum? I'd predict not.


I don;t follow how low voltage to the phase splitter would make the thing single ended. Besides, the hum remains with the PI missing, right?

If no DC haunts a grid, with the PI out, ground each power tube grid to see if it makes a difference.

How is bias created? If this is cathode biased, where does the NFB winding go? If ther is fixed bias, then how clean is that?

[6267]

Hopefully I'm measuring ripple correctly--I set my meter to AC volts, highest setting. I put the + probe at the OPT center tap, and the - lead at the common grounding point. That reads 0, so I reduce the setting to 20v AC, and get between 1 and .5v.

With it wired up as it was, I could read DC 389v to ground at this point. Since there are only two diodes, I assume that means the transformer is its own center tap.

It's cathode biased, but the shared cathode resistor comes off the center tap of the NFB winding (i.e. the rheostat). Each leg of the winding goes to the cathodes of the power tubes. I was assuming that by reversing it I'd be reintroducing common mode stuff like whatever magnetic coupling is present between PT and OT and amplifying it. Maybe not, only one way to find out

I'm not sure what I was thinking about the PI, guess that was eliminated. I guess I'll try removing the CFB winding and grounding out each and both grids to see what I can get. Thanks Enzo!

[Enzo]

I am still confused, forgive me.

The tube cathodes each connect to one end of the NFB winding, and its center tap then connects to the cahode resistor which itself is grounded? Yet in post #3 you mention there is a "rheostat on the center tap to ground." or are you saying this rheostat IS the cathode resistor?

[6267]

Yes, rheostat is the cathode resistor. Figured it'd help determining the best value. So I increased the power supply capacitance with two 680uF/200v caps in series (for the voltage rating). The hum decreased, but the largest decrease came from grounding one of the power tube grids. I think I just have a stray coupling issue. I checked that the grid leaks were grounded.

Hum is much better now, but still not dead quiet (I have similar amps with much less capacitance in the psu that hum less...at least this one stays on for 50 seconds after you pull the plug ). I think my filmosound redesign has something like 50uf after 5y3 and it is studio quiet even at full bore.

The OPT has poor low frequency response (perfect! It's a guitar amp, anyway) and so I think 60hz comes across mostly as the harmonics and not the fundamental. I guess it's time to just go all out and bias the heaters positive, experiment with lead dress and bypass the screens better until this thing is quiet as a churchmouse. Thanks for the help Enzo, I hadn't thought to ground out a grid, so I would have put a whole farad on the B+ before I figured out that it was somewhere else

[Enzo]

Was ther any DC on that grid?