The Edcor meets the 6AV5

[sorenj07]

All right, thanks! I have the older GE type so I'll try and see if 320V is OK if I ever end up building one of these. I can upload my implementation later once I scan it/take a steady picture.

[Boris_The_Blade]

If you are getting a custom transformer from edcor for this for cathode feedback anyway, ask for a separate UL winding, that way you can run the screen at ~200V and the plate much higher. But also, if you run enough cathode feedback, you will get a UL effect anyway with a safe fixed low screen voltage.

[sorenj07]

I would have just used the 8 ohm secondaries for CFB, with the cheap 15-watt 5K:16 version which comes with UL taps anyway. Is it expensive to have Edcor make all these extra windings? For maybe 7 watts absolute maximum it seems like the 15 watter should be ok..

[Boris_The_Blade]

For new designs from Edcor, there usually a $20 one-time fee, split between two $20 trannies ain't too too bad. But, they may also have already made something like that, in which case they waive the fee...might wanna just ask what they have that is 'close,' and design around that. They're nice people to talk to on the phone, and usually willing to help out. You can also email them but sometimes it takes a while for them to get back to you.

[jon_010101]

Hi Soren, ... How about running the UL tap to a pentode driver stage's plate resistor, a la E-linear? My mental calculations suggest that this is the same as adding a *boatload* of CFB, without having to couple it through the secondary -- it might even be "too much", in which case you can reduce the feedback with a voltage divider.

You could probably pick a textbook 6AU6 operating point to get plenty of gain and voltage swing to drive the tube to clipping. Check out the Ebb=300V operating points in the data sheet: 6AU6A

60Vrms with a gain ~160 given a 300V supply.

[sorenj07]

Umm, wow. What kind of output impedance would this 6AU6 have? I can imagine this driving a bigger tube, or maybe in push pull...

[jon_010101]

Originally posted by sorenj07
Umm, wow. What kind of output impedance would this 6AU6 have? I can imagine this driving a bigger tube, or maybe in push pull...

Pretty high! ~100k -- but I think it should be sufficient for good bandwidth. "Rs" in the data sheet is less than your likely grid leak resistor value, so that's in the right ballpark. But I have no idea what the input impedance of a 6AV5GA in UL is (probably benign, but someone may correct me).

Consider that the Quad II amplifier uses a pair of EF86s with 180k plate loads to drive 6L6s in CFB-UL, and then puts a global loop around that, and has usable bandwidth to 40kHz IIRC. The 6AU6 should be in a similar or better position.

If there's any doubt of bandwidth, you can go completely nuts and run a 12HL7 driver tube instead. But I'd regard that as extreme overkill, easily capable of driving a very big triode (or serving as a nice little SEP power amp).

BTW, another possibility is to reduce the plate resistor on the 6AU6 and pick a new operating point -- trade some gain for lower output impedance. Lots of ways to run a pentode, but it can take a bit of work to find 'em.

Another BTW: You'd also probably get enough feedback this way to run the 6AV5 as a "real" pentode, with no screen connection. Then just use the UL for "EL" plate feedback, maybe even jack the plate up to 400V, and run low/safe screen voltage from a separate supply. You have a lot of control over the circuit that way, and don't need to watch out for glowing screens. It's also trivial to drive that way!

[sorenj07]

Thanks, a lot of cool ideas to consider. I'm not averse to the idea of using pentodes at all.

[Boris_The_Blade]

Did some experiments the other day and decided to do triode curves of a 6AV5, and now to share it with you all I may regret this later when these tubes all go for mad $.

Pardon the slight sloppyness...I don't own a curve tracer beyond that of a few meters and power supplies.

[sorenj07]

Hey, thanks a ton! The curves don't look very similar to those of a 2A3/6B4 though