1970s Ampeg V4 Strange Input Stage Behavior

[Helmholtz]

Not sure what is going on with the photo numbering and pinouts, but 2nd stage of V1 has around 2V at the cathode. That means at around 2V pk at the grid (pin2) it will start clipping. Your photo # 4 shows far more than that.

It is not so easy to drive a tube with unbypassed cathode resistor into grid conduction, because the cathode voltage follows the grid voltage as with a CF.

E.g., if the gain from grid to cathode is 0.5 (chosen for convenience) and the cathode voltage is 2V, a 2V increase in grid signal will cause the cathode voltage to rise to 3V.
So there's still a 1V margin to grid conduction and it would take an 8Vpp signal at the grid to hit Vgk = 0 at positive signal peaks.
In reality the "CF" gain will be above 0.8, so a clean grid signal of 20Vpp should be possible.

[dubulup]

What is the voltage drop across (between the ends of) R13 (68k)?
That voltage determines the positive clipping level of the plate signal (V1 pin 1).

drop is measured at 81.2V
326V E voltage to 243V at V1-1
C7 shows some insignificant oscillation 0.001 to 0.005V...nothing worse than the replaced coupling caps show

[dubulup]

My thoughts:

I notice V1,1 plate is directly connected to V2,1 plate without any mixing resistors. This means V1,1 plate has to drive the output impedance of V2,1 plate (and vice versa) making clipping very asymmetrical as you can see from the scope screen shots. It could be working “as designed” like Enzo said.

The vol pots are linear taper which makes the volume come up quickly with pot rotation and there may not be much volume increase after 12 noon compared to a log pot. I'd set the pre-amp clean and check the power output at the onset of clipping into a dummy load.

yes, I've noticed that the V1-1 plate signal shows up on V2-1. I thought this could be the smoking gun since both channels share this network...no change if V2 is pulled. Parts for dummy load show up today, so I'll be able to verify (or not) clean pre-amp can be driven clean to load; the sentence is loaded to "verify" as everything else I'm seeing with the power section, PI included, seems appropriate...even with hot bias, zener'd screens, and lower B+.

[dubulup]

I haven't given up on my phase inverter notion. I wasn't suggesting the 12AU7 was bad, but more that the circuit around it was funny. And half a phase inverter might still work reasonably well up to some moderate level before going wonkers.

I'll dig more into this today...and report findings

[Helmholtz]

I think I now understand what's going on and it was Dave H who helped me.

With only one triode connected to the plate resistor the positive headroom would be around 80V.
But as the other triode always draws its idle current dropping 40V across the plate resistor, the positive voltage swing is limited to 50% of that 80V.

Consequently positive plate signal peaks will clip at 40V, well corresponding with scope pics.

So, as Dave said, the behaviour is normal by design.

[dubulup]

That is interesting, I need to recheck with V2 pulled...because wouldn't release the limit from 50% to 100% giving the player more head room in the pre-amp section?

still doesn't explain the abrupt change...unless something shook loose/into place and the second channel returned to normal operation, haha!

I still think something upstream in his signal changed and it just hotter now.

[Helmholtz]

That is interesting, I need to recheck with V2 pulled...because wouldn't release the limit from 50% to 100% giving the player more head room in the pre-amp section?

That won't help, as it halves the voltage drop across the plate resistor, leaving the headroom unchanged.
You would have to double the value of the plate resistor when you pull V2 to get the 80V headroom.

While not completely impossible it is very unlikely that some fault significantly increases amp gain.

Let's see if you get full output power.

[dubulup]

https://share.icloud.com/photos/05a9...jLkotHxFPECwWw

I will try to post this 15 second video using iCloud (Vimeo is down at the moment)...if a picture is worth a 1000 words, a 15s video must be worth at least 360k words. (edit: looks like you have to download it...don't have to sign in or anything, just download if you feel like you'd like to see it. I'll upload to a hosting site once Vimeo returns)

V2 is vacant and R13 shows a 40V as expected
input = 0.3V 400hz
plot = V1-3 cathode with 10 o'clock volume

initial plot is -9db...then -6dB...and finally 0dB; which is distorted at 0db
then switched back down to -9dB and volume rotated from 10 o'clock to 2 o'clock
and finally back down to 10 o'clock

[Helmholtz]

I will try to post this 15 second video using iCloud (Vimeo is down at the moment)...if a picture is worth a 1000 words, ....

What is this meant to demonstrate?

[g1]

My thought is that the preamp is a red-herring. The relatively early pre-amp clipping is normal for this model. From the voltages given in svc.man., 1.4v pk-pk coming out of the volume control will already get you 3/4 of full power at speaker. Input sensitivity is listed as 11mV RMS. I can easily get preamp distortion from my V4 at volumes above 10 o'clock, depending on sensitivity switch and guitar pickups.
If the owner noticed a big change overnight, power amp problems would be a likely cause.
You can use the ext.amp jacks as 'preamp out', or 'power amp in' for testing.
Without a load you can still pull the power tubes and check that you are getting proper signal at both sides of PI tube.

[dubulup]

What is this meant to demonstrate?

just to note/show what is happening at the input

My thought is that the preamp is a red-herring. The relatively early pre-amp clipping is normal for this model. From the voltages given in svc.man., 1.4v pk-pk coming out of the volume control will already get you 3/4 of full power at speaker. Input sensitivity is listed as 11mV RMS. I can easily get preamp distortion from my V4 at volumes above 10 o'clock, depending on sensitivity switch and guitar pickups.
If the owner noticed a big change overnight, power amp problems would be a likely cause.
You can use the ext.amp jacks as 'preamp out', or 'power amp in' for testing.
Without a load you can still pull the power tubes and check that you are getting proper signal at both sides of PI tube.

I think you are right...moving on. When I dial the input channel knob to match the service manual input stage values the 6K11 is spot on with measurements.

The 12DW7 measured low, even at MAX master volume (mid in the 300hz position and all tone controls at noon).

time to dig into the circuit that feeds the cathode follower

service manual / measured
12DW7 PIN
2 - 0.32V / 0.20V
3 - 0.3V / 0.19V
6 - 3.5V / 1.7V
7 - 0.28V / 0.17V

I'll dig more into this today...and report findings

All components around the PI triodes measure ok; with MAX master volume I'm get about half (service manual values) at the grids yielding half at the plates...seems this is operational fine. I'm thinking once the input into the 12DW7 is within spec, the PI will follow...then the power stage.

I measured the power tubes and PIN-3 should be 250 and I only get 163 and 165 at MAX master volume. PIN 5 should be 38 and I'm only reaching 14 and 15.8.

thanks for probing me to move on; seems I get stuck in loops easily.

[dubulup]

first I should apologize...as I thought this thing had a master volume OEM...as I'm stating all previous notations. So sorry for that...this is clearly just another modification; now that I'm intimately familiar with this board.

Anyhow, all the circuits look/measure good for the most part...the low readings on the 12DW7 tube go away with the slightest crank up on the channel volume and are within service manual values. Same for the power tubes. Hell with the hot bias I get close to 300 on PIN-3...and that is with the channel volume at maybe 10:30 / 11:00.

I'll get the dummy load built, hopefully tonight, and scope the output and see what's happening.

I did get some feedback on the input chain gear, older single coil pick ups, moogerfoogers...but the most notable change was using a fuzz pedal.

edit
RE: MoogerFooger
All Moogerfoogers output a line level signal.

[g1]

Measuring those plate AC voltages at the power tubes, you have a speaker connected instead of load box?

[dubulup]

Measuring those plate AC voltages at the power tubes, you have a speaker connected instead of load box?

Yes, there was a speaker connected…is there an issue with that? Didn’t get far on the channel volume as it gets too loud very quickly.

dummy load parts delayed.

[g1]

Yes, there was a speaker connected…is there an issue with that?

No issue. Just wanted to make sure you had a load connected. 250VAC at plates is shown for 3/4 power so 350VAC at plates would be pretty loud.