5U4GB voltage doesn't seem right

[Phil_S]

I just finished a 5B6. I'm posting here because I think it is a generic question. I tried two different rectifiers with the same result. With all tubes in and amp fully built, it is 758VAC in, 362VDC at B+1. What is sucking so much voltage? I expected something north of 400 and was actually concerned that 450V rated caps might not do the job.

Here are some fresh, as built pictures. Did I do something wrong with the bleed resistors?

More info on this build in the Fender section including voltages posted yesterday -- near the end of the thread.

Thanks for any thoughts.

[Reeltarded]

Expectation? So when you say 758 and I divide that by 2 and then I see around a 17v drop on that first node?

Is that not correct?

[JazzGuitarGimp]

Expectation? So when you say 758 and I divide that by 2 and then I see around a 17v drop on that first node?

Is that not correct?

Don't forget the cap charges to the peak voltage. 768 should be the RMS voltage, so:

((758 / 2) - 17) x 1.414 = 512VDC. Seems like something is amiss...

[LeftyStrat]

Under load it would be closer to:

(758/2 * 1.2) - 17 = 438

The 5u4 should drop it another ~40.

Still doesn't work out. I'd check the value of the bleeders and for leaky caps.

[tsutt]

Looking at the schematic vs your rec socket. it looks like your wired to pin 3 instead of pin 4. i dont know anything about a 5u 4. or its an optical illusion.

[Reeltarded]

Wow. It sure does..

[Deric]

Agreed....rectifier wiring looks suspect.... could be the pics.....

However....

...bleeders..... Are they green, blue, yellow or green, blue, orange?

Green , blue, yellow would give a total of 280K which should be fine.

Green, blue, orange would give 28K....not fine.....


Pull the bleeders till all else is worrked out. Double check rectifier socket wiring.

[Firestorm]

My money's on the socket wiring; it sure looks like pin 3 is connected. Running a 5U4 configured half-wave should produce a pretty big current on the single charging pulse per cycle, probably enough to drop the output voltage down to the full load number.

[Phil_S]

Ding! Ding! Ding! Someone wins the bet! Wow, this is embarrassing, soldering to the wrong pin, especially when I know what to do. Well, it is good that pictures tell the story. In my experience, it is so much easier to spot someone else's error than your own. Thank you to everyone who spotted it.

To the question about the bleeders. That's a pair of 560K in parallel for a total of 280K. They are fine.

Resulting voltages, for those who want to know:
756VAC
B+1 408
B+2 345
B+3 326

6L6 plates 396, 399
screens 344, 345
cathode 27 (bad news, that 25V cap has to go)

6SL7 plates 189, 189
cathodes 2.38, 2.34

6SC7, parallel plates 136.2
cathode 1.68

Calcs suggest power tubes are at 108mA for the pair (includes screens and cathodes) and about 40W static plate dissipation for the pair.

For a 5U4GB, I would have expected (756/2 * 1.2) - 50 = 404. The result at 408 is nearly spot on now. The reason I did not give a figure previously is because I was uncertain the VAC input was under a full load and thought it might drop. I think this PT is not even breaking a sweat running this amp, as there was no meaningful change.

[tsutt]

your welcome, glad i was able to help someone in return.

[Structo]

What is that black blob under the first filter cap?

[Phil_S]

What is that black blob under the first filter cap?

I insulated the the retaining nut with a bit of heatshrink as it is close to B+1. It is the other of the pair. The ground side one is in the open. I managed to cut the pre-made board one or two turrets short. I am going to see if I can get a nylon nut to use. I think Ace Hardware has them.

I did that because I observed some arcing in that area. It turned out to be sloppy work on the second turret holding the other B+1 cap, but I don't like how close it is.