Monkeymatic Acetone Princeton

[xtian]

Thx for the replies, guys. I applied 29.98 vac to the OT primary, and saw 1.38 vac on the secondary. This is 21.72:1. Squared, 471.8:1. With 8 ohm secondary load, 3774:8. And this matches closely the marking on the transformer, 3500:8.

I see that a pair of 6v6's want 8.5k-10k load. So, am I halving my power because of an OT mismatch?

[cbass]

OK, I'm mostly stupid, but not all the way. I put in 5K6 load resistors instead of 56K. My space shuttle would have blown up on launch...

But now that I've fixed that, the amp sounds almost exactly the same! Not loud enough! Voltages look better on the cathodyne, but not perfect (259 plate, 29 grid, 82 cathode).
[/b]

SHouldn't the grid be just a few volts lower than the cathode?

Have you tried a different tube?

[JazzGuitarGimp]

I think you nailed it. The original transformer for that model uses an 8K primary. If you're using an 8 ohm speaker, you're presenting a 3.7K load to the 6V6 plates. If you have a 16 Ohm speaker you can try, that will get the load up to 7.4K, which is be close enough to see if that's the issue. AA964 schematic calls for an HT of 420V. Where is yours?

[xtian]

@Lou: My HT is 440. I don't have a 16ohm speaker, but I have a dummy load I can try.

@Cbass: I don't know! Merlin's page has great info on the cathodyne, but doesn't list voltages, and the AA964 schemo doesn't list the grid voltage.

[JazzGuitarGimp]

Getting to the PI grid voltage is just a matter of math. Let V represent the measured voltage at the cathode, R1 is the top resistor (in the AA964 schematic, it's 1K) and R2 is the bottom resistor (56K in the AA964 schematic).

Then:

Vg = (V / [R1 + R2]) * R2

Using the values on the AA964 schematic, we have:

Vg = (6.5 / 57,000) * 56,000
Vg = 0.00011403509 * 56,000
Vg = 6.3859 V

[xtian]

Thanks, Lou. The schematic actually shows 65v at the cathode. Your math shows the predicted voltage at the junction of the 1K and 56K cathode resistors (63.8v on the schematic). But you have to cross the 1M resistor to get to the grid. Would you expect the voltage to drop across the 1M?

I guess there is a tiny amount of current flowing thru the grid. To check, I lifted one leg of the coupling cap (shown with red X). I measure 81.5v at the junction of the two cathode resistors, and 29v at the grid, a drop of 52.5v, or 0.05mA.

I guess what I'm asking is, are my cathodyne voltages OK, and I think the answer is yes. Can you confirm?

Second, I want to add a 470K grid stopper as Merlin suggests; does that go at the red X? Keep the coupling cap? (Merlin's schematic does not show a coupling cap.)

[JazzGuitarGimp]

Not an expert here, but here are my thoughts:

First, I'd really be surprised if lifting one end of the coupling cap has an effect on the idle, DC voltage at the grid. Unless the cap is leaky, of course.

I am surprised by the voltage drop across the 1M resistor. I didn't expect this. The input impedance for a 12AX7 is supposed to be really high, so I expected a much smaller drop across the 1M resistor. I was actually expecting no drop across the 1M resistor - someone please correct me if I am wrong as I'd like to learn more about this.

If this grid circuit is operating correctly (big drop across the 1M being correct and expected), then I'd think adding a 470K grid stopper between the 1M resistor and the grid would significantly change the operating point. Better to place the grid stopper where you've indicated. But remember that for a grid stopper to do its job well, it has to be soldered with as short a lead as possible right at the tube socket terminal. I'd even go so far as to bring the 1M resistor down to the tube socket, keeping its lead as short as possible as well.

Have you tried measuring the grid voltage with the tube removed? If so, do you still have a big drop across the 1M resistor?

[Tillydog]

Thanks, Lou. The schematic actually shows 65v at the cathode. Your math shows the predicted voltage at the junction of the 1K and 56K cathode resistors (63.8v on the schematic). But you have to cross the 1M resistor to get to the grid. Would you expect the voltage to drop across the 1M?

...

There's negligible voltage drop across the 1M resistor - no grid current - the grid will be at the voltage at the junction of the 3 resistors (1k, 56k, 1M) which looks about right.

You can't measure grid voltage directly - you just get nonsense, as you've found because even DMMs have too low an input impedance.

65V P-P should be enough to drive your 6V6s fully, I think, but haven't done the maths for your B+, etc. What's your bias voltage?

Twice you have mentioned the bias current of the 6V6s - once as 0.023mA, once as 0.023mV across 1R - are they typos? Maybe double check that your 1R resistors are really 1R and that your bias current is ~~25mA (I can't see the resistor values in the photos) - Might explain your high B+.

(From the voltage readings, I guess you did change the 5k6 in the PI cathode, as well as the plate side).

I suspect the main problem might be the OT mismatch - also, it looks like you've added 1.5K screen resistors to the 6V6s which I guess might reduce the output some (no idea how much, though). Trying a 16R load would tell you.

[xtian]

First, I'd really be surprised if lifting one end of the coupling cap has an effect on the idle, DC voltage at the grid. Unless the cap is leaky, of course.

You're right, no change in grid voltage with coupling cap removed.

I am surprised by the voltage drop across the 1M resistor.

I was, too, until I saw that this is only 0.06mA. Really small.

But remember that for a grid stopper to do its job well, it has to be soldered with as short a lead as possible right at the tube socket terminal.

I put in the 470K grid stopper, but not right on the tube. Isn't this advice just to prevent external radio interference?

Have you tried measuring the grid voltage with the tube removed? If so, do you still have a big drop across the 1M resistor?

Voltage at the grid with tube pulled is zero. There's nowhere for the current to go with the tube pulled!

[xtian]

What's your bias voltage?

Bias is -40.7vdc on the grids. Yes, typo earlier! My voltage across the 1R resistors at the cathodes reads 0.024vdc, meaning 24mA.

I suspect the main problem might be the OT mismatch - also, it looks like you've added 1.5K screen resistors to the 6V6s which I guess might reduce the output some (no idea how much, though). Trying a 16R load would tell you.

Will do. Thanks, 'dog!

[JazzGuitarGimp]

Yes, I forgot about the meter's input impedance. Thanks Tillydog.

And, yes, of course pulling the tube breaks the connection from B+ to the cathode divider network. Forehead Slap!

I guess we all need to buy a VTVM to measure grid voltages accurately. Oy, the thought of going back to a needle pointer!

[xtian]

I hooked up a 16ohm resistive load, and I observed 15.8 watts output just before clipping. Seems right on target, doesn't it?

Any idea how that 15.8 watts into a resistive load compares to pushing a real speaker?

Anyway, next test is to jury rig two 8ohm cabs in series for a real world test.

Hmm, I notice the two speakers I've tested have been WGS Reaper (sens. 99.6dB) and Celestion G12P-80 (sens. 98dB). Certainly I could do with a more efficient speaker...

[JazzGuitarGimp]

I hooked up a 16ohm resistive load, and I observed 15.8 watts output just before clipping. Seems right on target, doesn't it?

Any idea how that 15.8 watts into a resistive load compares to pushing a real speaker?

Anyway, next test is to jury rig two 8ohm cabs in series for a real world test.

Hmm, I notice the two speakers I've tested have been WGS Reaper (sens. 99.6dB) and Celestion G12P-80 (sens. 98dB). Certainly I could do with a more efficient speaker...

15.8 W looks good for cathode bias. Is it fixed bias or cathode biased?

Do guitar speakers get all that much more efficient that about 100dB? I think I've seen 101, maybe 102 in my travels...

[xtian]

Speaker test went OK. Definitely louder. Perhaps 3dB. Still, for comparison, not NEARLY as loud as my 15 watt Tweaker, 15 watt Super Champ XD, 15 watt Gibson GA20, 15 watt Traynor YBA. Hell, this amp isn't as loud as my old 5 watt Black Heart.

Lou, this is a fixed (adjustable) bias amp.

Just want to make sure my power section is functioning to spec: I was seeing my 1kHz sine wave 60-70vac p-p, clean, out of the PI into the power tubes. Power tubes have 440 plates, 437 screens, -40 grids.

Guess I'll order an OT.

[cbass]

Doh forgot about the impedence loading down the meter.You should see hardly any drop acroos that 1 meg.Sounds like the pi is right now.

Your power output sounds about right.I would expect closer to 20 watts with those voltages and fixed bias.Guess it all depends on how much distortion you allow.

I wouldn't expect a mismatched primary to reduce power that much.At least not to less than 5 watts
but what do i know?

I've used febder hotrod deluxe OT with 6v6's with good results.It;s a beefy tranfo for 2 6v6's.
JUst use the taps at 8 and 16 ohms instead of 4 and 8.CE has them for like 30 bucks.
Maynot be what your after if you want stock princeton sound