Model T build questions

[eddie25]

I've been working on a Model T style amp and I decided to go with a dual bias setup. It comes off the HV secondary and goes to two completely separate Marshall style bias circuits from there (which feed each side of the power section).

There's a section in the T schematic where the cathode of the third gain references the bias voltage through a couple 22K resistors and my issue is that when I do this from one of my bias sections it drops the voltage (less negative, from about -60VDC to about -45VDC) of that side. The other side stays normal.

I'm not really sure why this is happening as there really is no direct ground reference beyond that 22K except the single (reversed) E-cap.

Do I assume something is amiss? Does anyone think it's odd to be dropping voltage? I have double checked the wiring several times.

Should I run a third leg sans pot just to give me another separate negative voltage reference?

Should I run resistors from both bias sections so that they drop the same amount?

Any thoughts would be appreciated.

[Leo_Gnardo]

There's a section in the T schematic where the cathode of the third gain references the bias voltage through a couple 22K resistors and my issue is that when I do this from one of my bias sections it drops the voltage (less negative, from about -60VDC to about -45VDC) of that side. The other side stays normal.

woops now the schemo's here, let me put on my thinkin' cap.

Although there's no connection to ground, current is drawn from point E in the bias supply by the cathode of the pre tube you mention and that would upset the balance between your twin bias controls. I'm thinkin' it may be best to provide a separate path for that current, as you said a third leg sans pot.

Wonder what Sunn's designer was trying to accomplish with this feature.

[John_P_WI]

Leo, I'm guessing more headroom? If you can't raise the plate voltage, why not lower the cathode?

Regarding the OP question, my guess is the marshall bias setup is too high of impedance to handle the additional load of the second added circuit. Hard to tell without a schematic what exactly was done....

[Leo_Gnardo]

Leo, I'm guessing more headroom? If you can't raise the plate voltage, why not lower the cathode?

Regarding the OP question, my guess is the marshall bias setup is too high of impedance to handle the additional load of the second added circuit. Hard to tell without a schematic what exactly was done....

Yep, some hi fi designs do it, and some even put a constant current supply in the negative leg. That extra headroom may well be needed driving a couple of big KT88 or 6550 grids.

Eddie25 you may have to experiment with resistor values to satisfy the current draw for that triode as well as keep the output tubes' bias requirements happy. All part of DIY fun. Keep us informed how you make out with your T.

[eddie25]

Thanks guys, that makes sense. I think I'm going to try to see if I can make room for an extra negative voltage leg. Otherwise I guess I can run a single bias setup and use the other for the triode.

Here's how I have the bias set up currently, but I think those 220K's ended up as 330K's, HV is 800VAC.

[JazzGuitarGimp]

That's a very interesting gain stage, and it seems to contridict accepted wisdom in terms of biasing a 12AX7. Nirmally, the cathode is at some positive voltage wrt to the grid. In this design, when the bass control is all the way down, the grid has a 43K DC path to ground, yet the cathode is held at some negative potential. What keeps the triode from red-plating?

[Structo]

I use dual bias pots on my Dumble clones.

But both amps have bias taps on the PT so they aren't high voltage.

I just like to know what each side is doing current wise.

Plus you can swap tubes around for the best pairing.

[sluckey]

yet the cathode is held at some negative potential.

It seem like it, but the cathode will really be slightly positive with respct to the grid. When you consider the entire cathode/plate current path you will see a voltage divider from point E (negative voltage) to point C (positive voltage). The path starts with point E, R56 (22K), R55 (22K), the tube, R57 (120K) and finally point C.

Just assume that point E is -50V (just for round numbers). If 1.2ma flows thru the tube, it would also flow thru 44KΩ (R56 and R55) cathode resistance. Voltage drop across that 44K would be 52.8V which means the cathode would actually be sitting at +2.8V WRT ground.

[eddie25]

That's a very interesting gain stage, and it seems to contridict accepted wisdom in terms of biasing a 12AX7. Nirmally, the cathode is at some positive voltage wrt to the grid. In this design, when the bass control is all the way down, the grid has a 43K DC path to ground, yet the cathode is held at some negative potential. What keeps the triode from red-plating?

Dang, you must be right. I just set up a new negative voltage node at exactly -55 and while it worked great and the amp sounded fantastic, the tube from that gain stage (12AX7) went out after about 5 minutes.

With the tube in the voltage drops from -55 to -19VDC. I'm getting +2.6VDC on the cathode, just as sluckey mentioned. 288VDC on the plate, kinda wondering why the tube went out.

[Smokebreak]

Eddie, offtopic here, but what transformers are you using for this build?

[JazzGuitarGimp]

yet the cathode is held at some negative potential.

It seem like it, but the cathode will really be slightly positive with respct to the grid. When you consider the entire cathode/plate current path you will see a voltage divider from point E (negative voltage) to point C (positive voltage). The path starts with point E, R56 (22K), R55 (22K), the tube, R57 (120K) and finally point C.

Just assume that point E is -50V (just for round numbers). If 1.2ma flows thru the tube, it would also flow thru 44KΩ (R56 and R55) cathode resistance. Voltage drop across that 44K would be 52.8V which means the cathode would actually be sitting at +2.8V WRT ground.

That one slipped right by me. Thanks for setting me straight, sluckey!

[eddie25]

Well, new tube in that spot with no problems so far after playing an hour. Everything sounds great and it's putting out almost 200 watts wide open.

Smokebreak, I used Hammond 1650T, with two hammond 1.5H 200mA chokes and a huge 800VAC PT. B+ is about 520.

[Smokebreak]

Glad you got it all worked out, and thanks for the info. It just so happens someone approached me about this amp the same day you posted.
I'd love to see some pics when you are done.

It looks like one could use the Hammond 159V choke as well, and only 1 of them as it's 500mA.

What did you use for the mid inductor?

[eddie25]

I used a small R. F. choke that AES sells, I think it was the 2.5mH one. The mid switch sounds cool. I'm pretty sure I wired it incorrectly, I ignored the SS pole and used a SPDT. I have no idea how they do it in an original. I used to have some killer pictures of one I repaired but I lost 'em. I'll let you know if I photograph it, hobbling together that third -VDC node sorta killed any vibe of wanting to add it to public record gut-shot wise.