LTP PI - Is this normal?

[Daedalus]

Aiken guided calculations

I just went through the exercise of following Aiken's "designing LTPs - load line approach" using the 325 volt source voltage that I currently have. Using the -2v bias line for a 12ax7, I calculated (approx) 1K and 56K tail resistors needed.
Using the -2v bias line for a 12at7wa, I got 500ohm and 22k tail resistors.
So I put the at7 back in and changed the tail back to 22k; I'll take some new scope pics.
The one thing I noticed right off is that the output signals are a little less balanced on the scope.

[Daedalus]

This is the at7 with 22k tail. I turned the gain up till the cathode signal started to clip (first pic, top: grid, bottom: cathode). Second pic shows the plates at the same level.
Hope that's what you meant...

[loudthud]

The waveforms you are seeing on the cathodes is the result of not driving or connecting feedback to the lower tube. When the driven side tries to go negative, the waveform clips at the common cathode point because the undriven tube's cathode won't allow that point to go negative much below the static bias point. All the current from the tail resistor is going through the lower tube and the upper tube isn't conducting so you could drive the grid to -100V and it's cathode will stay at the same voltage because the other tube is holding it there. Does that make any sense?

[Daedalus]

I actually have tried connecting the feedback. When I do, I see in-phase signal at the second grid, but the clipping on the cathode doesn't change. I was starting to rationalize that this must be either: a normal artifact of this design or my bias point is off (cathode is at 72v and grid at +45v with no signal).
I also put a resistor sub box from the PI end of the 220K to ground to attenuate the input signal. That didn't change the clipping at the cathode either until I got down to almost no signal. It's all very confusing; I wish I knew more.
Thanks for your input.

[loudthud]

You can't reliably measure the DC grid voltage on an LTPI to ground with a normal multimeter. The 1meg grid leak resistors mean the bias point will shift when you connect the meter probe to the grid. To measure the bias point, measure across the bias resistor (most commonly 470 ohms). If you suspect grid current or a leaky cap, measure across the 1meg, you should see zero volts. If you had two meters, you'ld see that the 72V doesn't stay at 72 when you probe the grid. The same goes with scope probes. Unless your scope is AC coupled attaching a probe to the grid will shift the operating point.

[Daedalus]

Very interesting. Here's what I see:
With no input signal: 1.5V across 470R, .01V across pin2 1M, and .002v across pin7 1M.
Max input signal: 1.28v across 470R, -30vdc across pin2 1M and -4.6vdc across pin7 1M.
As I turn up the input signal the voltage across pin2 1M raises to -0.5v (at about volume 3 where clipping starts) and continues to rise as the volume is increased.

Again, I don't have a clue as to whether this dc rise across the grid resistors is normal or indicating leakage or some other problem. Can anyone comment on that?

It's probably not related, but I also noticed that at max signal, if I turn the master volume all the way down (shorting the signal outputs) the voltage across the 470 rises to 1.7V (from 1.28v)

Also, when I calculated the tail resistor values following Aiken's example, I was targeting -2v grid bias and +90v on the cathode. I'm thinking of tweaking both the 470R and 22K up to reach those values. Does that seem reasonable?

Thanks to everyone looking at this...

[ETR]

I just went through the exercise of following Aiken's "designing LTPs - load line approach" using the 325 volt source voltage that I currently have. Using the -2v bias line for a 12ax7, I calculated (approx) 1K and 56K tail resistors needed.
Using the -2v bias line for a 12at7wa, I got 500ohm and 22k tail resistors.
So I put the at7 back in and changed the tail back to 22k; I'll take some new scope pics.
The one thing I noticed right off is that the output signals are a little less balanced on the scope.

It would be great to see some shots with the gain turned down.

[Daedalus]

Here you go. Sorry I misunderstood your first request.
These are both driven by a clean sine wave input.

The first two pics are traces of the PI plates, the third is the grid (pin2, top trace) and cathode (pin3).

[MWJB]

"My voltages are quite a bit lower than that. My PT has a higher voltage output available that could get me up there, however, I have a pair of 6V6GTs in this, don't know if they could take the 460v plate voltage.
Here's the whole thing... " I meant just see if the proportions of the AC voltages around the PI correspond to 65 Super Reverb, or 65 Deluxe reverb, which have a broadly similar PI.

You are putting 75v on the amp's input (Normal #1), surely not? Measure the input signal voltage with a meter.

[Daedalus]

"My voltages are quite a bit lower than that. My PT has a higher voltage output available that could get me up there, however, I have a pair of 6V6GTs in this, don't know if they could take the 460v plate voltage.
Here's the whole thing... " I meant just see if the proportions of the AC voltages around the PI correspond to 65 Super Reverb, or 65 Deluxe reverb, which have a broadly similar PI.

You are putting 75v on the amp's input (Normal #1), surely not? Measure the input signal voltage with a meter.

My apologies: I went off on a tangent there. I looked at the wrong AB763 schematic which has much higher voltages throughout. Today I looked at the 20W ab763 and it has voltages exactly like mine.

On the input signal voltage, I need to point up my basic ignorance. The scope trace reads 75V P-P. That should correspond to 26.5V RMS (p-p / 2.83), but my Fluke (true RMS) reads something quite a bit less than that (I don't have my notes here). The scope hasn't been calibrated for the last few years, but it's probe calibration output reads correctly on the display.

Attached is a pic of the 12at7wa plate characteristics with a load line added (blue) corresponding to the 325v source (80v on the cathode) and 82K plate load of the 20w ab763 design. It seems like a kind of a really low operating point.(?)
Correction: the left end point of the load line should be down one division (325-80)/82K => 2.9mA

[MWJB]

I'm not asking what the scope trace reads, but what signal voltage are you applying to the amp input (V1 pin 2). This will typically be <100mV. Measure signal applied & output with meters.

In the AB763 at Deluxe style voltages the 12AT7 PI tube will dissipate 1.5 to 2mA per triode, yours will be a little less as you have an unusually low B+.

[ETR]

Here you go. Sorry I misunderstood your first request.
These are both driven by a clean sine wave input.

The first two pics are traces of the PI plates, the third is the grid (pin2, top trace) and cathode (pin3).

Those are some ugly looking waveforms if that's supposed to be a sinewave, are you sure your scope is grounded properly?

[Daedalus]

Those are some ugly looking waveforms if that's supposed to be a sinewave, are you sure your scope is grounded properly?

Yes, back at the input to the preamp, the scope shows it as a perfect sinewave. That trace is from two channels in parallel (for what it's worth

[MWJB]

Your 2 channels are out of phase with each other, running them in parallel will degrade the sound.

Put 100mV 1KHz (confirmed with a meter) on one input at a time (tone controls fully clockwise), scope the output at the dummy load, what kind of W do you get with a clean sinewave?

[Daedalus]

For real: the scope reads 75v p-p and the fluke reads 35vac (across the 1m) and 43vac from pin 2 to gnd (at max signal).
On another post in this thread I show a scope trace of the input signal at pin2 which shows the signal at 12v p-p (12 divisions at 1v/dev). There is also a pic of the max signal input at 75v p-p in an earlier post.