Parallel 12AT7 for 2A3 driver?

[SHiFTY]

Just wondering if anyone has tried paralleling the two halves of a 12AT7 for use as a single-stage driver for 2A3?

It seems that with the high-ish mu, the decent current handling and the noise reduction of a pair in parallel, it could make quite a reasonable driver, at around 250V and 10mA.

Maybe a bit of 2H, but that would possibly cancel in the output tube.

Any comments or suggestions?

[SY]

I've found the linearity of 12AT7 to be quite good at lowish currents (2mA). You might think about running the pair as a voltage amp @ 2mA direct coupled to a CF, then running the CF at the higher current. That will give you great drive capability while retaining the tube's linearity.

Doing this trick will require elevation of the heater supply to some value midway between the plate voltage of the first section and ground.

[hacknet]

hey SY,

i was looking at the plate curves and they dont look terribly linear, i've got a big bunch of these babies and it got me interested.

so whats a good voltage to run these guys at 2ma?

[SY]

To swing enough for the 2A3, I'd want 150V or so. The CF gives you the opportunity to use a CCS on the voltage amp, which really helps linearization. If you do that, make sure that the cathode is either bypassed to ground or that you use a low impedance reference to bias it.

Yeah, the curves don't look promising, which is why my measurements surprised the heck out of me. Which 12AT7s do you have? I might have some data on them.

[hacknet]

i picked up quite a couple of sylvania blackplates from the scrapyard awhile back for a buck a pieces.

im thinking a ring of two ccs on the cathode with bc550c transistors. which make did you get sweet figures with at 2ma?

[SHiFTY]

I managed to score a couple of Telefunken ECC801S new in boxes... just lookin for something to do with mine.

So I am guessing a simple single stage driver is out? I like the idea of not having to use a CF- although they do drive very well.

Thanks for the tube data

[SY]

Measurements taken with CCS set to 2mA, LED as cathode bias. Plate voltages ran between 130-135V. Output adjusted to 8VRMS. 1kHz excitation. NOS tubes baked for 24 hours to minimize distortion from gas.

Mullard ECC81 (average of 4 sections)
0.06% THD
2nd -60.5
3rd -80.0
4th -88.0
5th -97

GE 6201 (average of 4 sections)
0.08% THD
2nd -57.5
3rd -85.0
4th below residual
5th -85.0
(noted that higher harmonics still visible above baseline)

JJ (average of 8 sections)
0.03% THD
2nd -66.0
3rd -91.5
4th below residual
5th below residual

Genelex B739 (average of 6 sections)
0.06% THD
2nd -59.1
3rd -79.9
4th below residual
5th below residual

Sylvania (average of 6 sections)
0.07% THD
2nd -65.6
3rd -64.9
4th -74.1
5th -89.5

RSD ECC81 (8 sections tested)
I didn't record this data because for all sections tested, the high harmonics did not decay. Horrible tubes.


Clearly, the JJ is the star of this show.

[hacknet]

whoa, thats cool.

what kind of plate loading ccs are you using? a ring of two?

[SY]

Bipolar cascode.

[Yvesm]

Originally posted by SHiFTY
Just wondering if anyone has tried paralleling the two halves of a 12AT7 for use as a single-stage driver for 2A3?

It seems that with the high-ish mu, the decent current handling and the noise reduction of a pair in parallel, it could make quite a reasonable driver, at around 250V and 10mA.

Maybe a bit of 2H, but that would possibly cancel in the output tube.

Any comments or suggestions?

Looks feasible

A bit more different: stacking the 250V supply above the bias one, you have 350V to feed the driver and this allows for an SRPP configuration as here:
http://www.dissident-audio.com/SE_6B4/Page.html

Yves.