Another power supply puzzler

[stokes]

This time its a Gibson white panel Skylark GA5. Its not the Champ circuit, its got 2 6BQ5's for output in PP.One 12AU7 ( soon to be 12AX7) and a transformer for the PI. Looking at the power supply there are 4- 20uf caps. The schem shows the interstage transformer and the plate of the preamp tube connected to paralleled 20uf's. But in the amp the paralleled 20ufs are connected to the screen on the other side of the 3.3k resistor. Which seems more correct? I got this amp from the client not working and he never plugged it in. It came to him with one 6BQ5 and 2 12AU7's in the other 2 sockets so its best he didnt try it. I'm more inclined to follow the schem and change what appears to be the way it came from the factory.What say you guys?

[nosaj]

There are many multiples of schematics for this amp. Did this schem come with the amp? make sure you have the right one.
nosaj

[Chuck H]

Well, being as it's a diode rectifier I would be inclined to use the parallel 20uf caps on the OT CT (plate supply) and a single 20uf on each node after. Or, if the caps are being replace then I'd just use a single 47uf on the plate supply and a 22uf on the following nodes. With the design at hand I can't understand why there is only 20uf on the power tube plate supply and 40uf on the preamp.?. What is the reason for that decision?

"It came to him with one 6BQ5 and 2 12AU7's in the other 2 sockets"

Well hopefully no one else fired it up that way either. Interested to see what your bench tests show.

[stokes]

Well, being as it's a diode rectifier I would be inclined to use the parallel 20uf caps on the OT CT (plate supply) and a single 20uf on each node after. Or, if the caps are being replace then I'd just use a single 47uf on the plate supply and a 22uf on the following nodes. With the design at hand I can't understand why there is only 20uf on the power tube plate supply and 40uf on the preamp.?. What is the reason for that decision?

"It came to him with one 6BQ5 and 2 12AU7's in the other 2 sockets"

Well hopefully no one else fired it up that way either. Interested to see what your bench tests show.

To me, neither scenario is correct. With the schem its got 40 on the preamp and what I see in the amp has the 40 on the screen with only 20 on the power tube plate. I have a lot of latitude in what I do to this amp. It comes from a new studio in NYC where my son was doing some recording recently. The owner says he doesnt care about keeping it stock, he just wants a good sounding low power amp for use in his studio. Told him Gibson wouldnt be my first choice but we'll get it working.My son tells me he's got about a dozen non working vintage tube amps in storage that he wants looked at. Right now I'm planning to go with my gut and just change the B+ rail to what I'm familiar with, 40 at the input and 20 for each node down the line, as you pointed out Chuck.Unless somebody can point out a reason not to.
As for the tubes that were in the amp, the chassis is also marked with pencil that matches what was there 6BQ5, 12AU7 and 12AU7 instead of 6BQ5,6BQ5,12AU7. Nothing looks fried inside so I dont know if anybody actually ever fired it up with that tube lineup, but they thought it was right enough to mis mark the sockets on the chassis.
Got a feeling this is going to turn into a complete or near complete rebuild. I already rewired the heater circuit, it had the old one side grounded wiring, I did the twisted pair with 100ohm resistors as a center tap wired to the power tube cathode and am changing the preamp cathode resistor to 1.5k and that 470k grid resistor for 1m ala Fender I will then try 12AX7, 12AY7 and 5751 and see what works best. But first new electros and a new 3 prong AC cord so I can fire it up and see what else it needs.

[J M Fahey]

This time its a Gibson white panel Skylark GA5. Its not the Champ circuit, its got 2 6BQ5's for output in PP.One 12AU7 ( soon to be 12AX7)

WHY would you do that?
12AX7 is the worst tube to use in a high current application (as in a transformer driver).

and a transformer for the PI. Looking at the power supply there are 4- 20uf caps. The schem shows the interstage transformer and the plate of the preamp tube connected to paralleled 20uf's. But in the amp the paralleled 20ufs are connected to the screen on the other side of the 3.3k resistor. Which seems more correct?[/QUOTE]

I would use 2 parallel 20uF caps on the +315V rail, since they are the main ones, just after rectifiers and feeding plates, a single one feeding screens on the +295V rail, and the remaining one feeding the 12AU7

[Chuck H]

Juan has a good point about the 12au7 being a better choice with the PI transformer. So I wanted to mention that if more gain is desired (probably) then a 12dw7 could be used with the higher gain triode as the first gain stage.

[stokes]

Juan has a good point about the 12au7 being a better choice with the PI transformer. So I wanted to mention that if more gain is desired (probably) then a 12dw7 could be used with the higher gain triode as the first gain stage.

Was actually thinking the same thing.We'll cross that bridge when we get there.

[Helmholtz]

JMF is right, a 12AX7 is not suited as driver here.

From circuit voltages the driver triode current calculates as close to 10mA and plate dissipation is 2.3W.
12AX7 limits are 8mA and 1W, also its very high internal plate impedance doesn't allow to effectively drive a low/medium impedance load.

A 12AT7 (15mA/2.5W, more gain than 12AU7) might work but probably requires adjusting bias (cathode resistors).

12DW7 is a good idea. Cathode resistor of input tube should be adjusted.

[stokes]

Ok, so I went with 40uf on the main, 20uf on the screen and 20uf to the preamp tube and the driver. I rewired the tube socket for a 12DW7 so the first stage is the 12AX7 half and the driver for the trans is the 12AU7 half.The 1k and 3.3k in the rail were both open, changed them. Rewired the input like a Fender, took that .0047 cap out and changed the grid leak resistor to 1m.I have the owners go ahead to do whatever it takes to get it sounding good so I should be able to power it up tomorrow and see what else needs changing.

[Chuck H]

Ok, so I went with 40uf on the main, 20uf on the screen and 20uf to the preamp tube and the driver.

Some brand was offering a "reverse" dw7 type tube for a while. With the lower mu triode on pins 6, 7and 8 instead of 1, 2 and 3. I remember getting one for a project. I'll be damned if I can remember the model name but I thought it was offered by EH. Anyway, it saved me having to rewire a socket in something or other. Since some searching didn't turn up the model or find a current manufacturer I guess I'll be hearing from whomever owns that amp if that tube fails before the amp gets trashed

[g1]

Some brand was offering a "reverse" dw7 type tube for a while. With the lower mu triode on pins 6, 7and 8 instead of 1, 2 and 3. I remember getting one for a project. I'll be damned if I can remember the model name

JJ ECC823

[Chuck H]

Thank you! I'm surprised it didn't pop up when I was searching around. And it seems to still be a valid model. Probably a better move on stokes part to rewire the socket so the grandfathered in model can be plugged in. I admit to taking the easy way out when I subscribed to that tube. And I'm glad to see it's still available.

[stokes]

I have a 12DW7 on hand so I rewired the socket.