Better Low Z driver stage design based on our Bootstrapped Gain Stage thread

[nickb]

A battery is fine so long as it's within say 5% of required heater voltage. I used a bench PSU and tied one side of the heater to the cathode.

BTW the PP cap won't leak significantly. I had just been working on something else where a 1uf 450V electrolytic cap with 80V was leaking enough to throw things off so this risk was in the front of my mind.

[SoulFetish]

I built your gm tester into the test rig board today. I went with the first suggestion. and conveniently tapped the fixed cathode to Plate voltage from the amp power supply. I'm using a 9V battery for my negative supply with a 250k pot for adjustment. I'm going to do some tests tonight (or tomorrow morning) stay tuned....

[nickb]

Just a data point. I tested a 12AU7 at 250V and Vg= -8.5V in my tester and got a gm of 2.

PS: I found out how I managed to get a gm of 3.5 earlier. I...err..ahem... applied a +3.5V grid bias instead of the - 3.5 I had intended.

[SoulFetish]

Just a data point. I tested a 12AU7 at 250V and Vg= -8.5V in my tester and got a gm of 2.

PS: I found out how I manged to get a gm of 3.5 earlier. I...err..ahem... applied a +3.5V grid bias instead of the - 3.5 I had intended.

haha! yeah, dude, I set up the test rig and ran some tests....
I mean, it couldn't be a simpler setup, but why can't any of this just be easy.
So, to start with, I installed another 9-pin socket on the board with the other tube circuit we've been testing to make it easier to just share power supply and heater sources. I was basically making point to point connections because of limited space, and it started getting really tight. So it was one of those deals where I was making the last few solder connections and trying to solder 3 wires to a single point and as I'm balancing my soldering iron so as to not burn the other wires pvc, the heat caused one of the wires to retreat out of the hole. So it too longer to build than it should have.
Anyway, I couldn't find a decent location on my amp to tap of 250V while it was lightly loaded, so I used +200V to begin with.
+200V, with a -9V grid bias = measured a idle current of 2mA.
So, I increased it a volt (per your instructions), and it stayed right at 2mA. Then I increased the supply to +300V, which is where the good stuff happens for this setup, because when I went to double check the voltages at the tube pin, the supply was +425V.
Yup, I triple checked it. My supply unloaded doesn't even touch +370. But there it was. Ugh. all the flying leads must have caused some solid oscillating. Plus I'm an idiot and I put the sense resistors on the plate rather than the cathodes.
I just went to an electronics supply shop today, and picked a few ferrite toroids to suppress some HF when wiring to the socket terminals. This is where I wish I had a good HV adjustable power supply with a heater tap.

[nickb]

For a variable HT supply I use a surplus power transformer run off a variac. For the heaters, a DC power supply. It's not that critical - I just picked the values i suggested off the data sheet as a known reference.

[Helmholtz]

Tube parameters vs. operating conditions (S = gm):

[SoulFetish]

For a variable HT supply I use a surplus power transformer run off a variac. For the heaters, a DC power supply. It's not that critical - I just picked the values i suggested off the data sheet as a known reference.

dude, I literally started searching for some info on adjustable HV supplies shortly after my last post and one of the first circuits I find was this, on Max Robinson's site:



It's so simple, I feel a bit foolish that I hadn't already thought of it. But, it's certainly something I can already set up without having to wait for parts. Earlier, I had considered using the triac control circuit you designed for that burned out... Bassman (?.. or maybe Twin, I can't remember). But, I like the variac idea a lot.

[nickb]

Yes, that basically. As the purpose is rough and ready for testing I didn't bother with the suppressor cap or an LC filter. A fuse or a polyswitch on the HT out is a worthwhile addition.

[Helmholtz]

Anyway, I couldn't find a decent location on my amp to tap of 250V while it was lightly loaded, so I used +200V to begin with.
+200V, with a -9V grid bias = measured a idle current of 2mA.

Why can't you use the +/- 127V supplies of your amp?
Also you could just use your amp circuit and vary R1 for a change of grid to cathode voltage of 1V and monitor change of cathode current.

BTW, in your amp the CF triodes operate at around 135V/7mA.

I'm happy to own a couple of these: https://www.radiomuseum.org/r/rohde_...?language_id=2

[SoulFetish]

Why can't you use the +/- 127V supplies of your amp?

Who wants to know??
(just kidding)

BTW, in your amp the CF triodes operate at around 135V/7mA

In fact they do not; not quiet that hot anyway. They usually operate around 123V across the load resistor (or -4V with respect to 0V ground), with a quiescent current around 6.8mA or so.
I noticed that I made a mistake copying some of the info onto the schematic for the cathode biased test setup. The cathode resistor R1 should've been 470Ω to match R7.
I've changed it in the original post to show the correct schematic. The 330Ω was a minimum value so as to limit the cathodes to around +4V, with respect to 0V ground. My mistake.

I've been having to break down my personal shop over the past week, but I was able to set up a transconductance test rig tonight and actually test several of the tubes.




But with a HV of +250V, and V_g -9V, the first subject had very well matched triode sections with I_a = 7mA.
at V_g -8V, I_a = 10mA
(textbook 3mA per Volt)

Most of the others measure 2-3mA, with some mismatched idle currents between triode sections.
A few tubes measuring outside those parameters were discarded

[nickb]

Let me just say this. My test jig wasn't anywhere as neat as that....

OK. Good results The junkers have been tossed. Onward and upward!

[Helmholtz]

They usually operate around 123V across the load resistor (or -4V with respect to 0V ground), with a quiescent current around 6.8mA

Means around 130V across the tube, pretty close to what I guessed from your original schematic.

What I meant to say is if you measure gm with a plate to cathode voltage of 250V instead of 130V, gm results might be 25% lower than in actual circuit.
See dependencies in the charts I posted.
OTOH, 250V/10mA makes sense for verifying datasheet gm of 2.2+/-0.4 mA/V.