Phase Inverter Clipping

[trem]

Finally got back into this amp, and there is noise it never had before.
I put a 1k 200mV signal into the amp. Followed it through, and it all looked fine until coming out of the PI.
The preamp and PI is 6SL7...and it is your basic Fender Vol, Tre, Bass, tonestack, and a LTP phase inverter. I did change to a new PI tube, but no difference.
When i put the scope on pin 2 of the PI (after the coupling cap) the positive sine wave is clipped. I changed tubes and it is still the same. I have never really looked at a PI before.
But i assume one side of the Sine should be Off/Clipped, and then just the opposite on the other wave.?
This does not look right does it...the photo with the clipped positive wave.?
What is causing this.?
Thank You
http://i.imgur.com/MAhx1oS.jpg
http://i.imgur.com/qQpsWqc.jpg

[Dave H]

Where are you probing for the two plots? I'm not sure about the first one but the second one looks like the PI is way overdriven and flat at the top is where the signal is clipped by the power tube at its grid as the PI tries to drive it positive.

[loudthud]

Waveforms don't look too unusual for a PI that is clipping hard. There might be a bias issue, but you won't see it with a signal that big. Back off on the drive signal and the plate signals should start to look more equal but out of phase.

Post some DC Voltages with no signal.

[trem]

Wow...OK.
I do not THINK i have the volume up much at all, but i will 2x check.
I can gather quiescent numbers for the pre and PI tubes, and get some bias numbers for the grids of the power tubes.
Thanks

[Chuck H]

When i put the scope on pin 2 of the PI (after the coupling cap) the positive sine wave is clipped.

That is the PI input. I think you're seeing the output of the stage feeding the PI. It could be due to your 200mV input signal. IMHE a 100mV signal is a better simulation for what a guitar will do to the amp. Try backing the input signal off to 100mV and set the amp controls to what you know as "max clean volume" when a guitar with the volume up full is plugged in. You may get a clean signal. Advancing the amps volume control a little from there should get you to a clipped signal. And the power tubes should be well into clipping by then. If not there may be some error in the gain structuring of the design.

[trem]

That is the PI input. I think you're seeing the output of the stage feeding the PI. It could be due to your 200mV input signal. IMHE a 100mV signal is a better simulation for what a guitar will do to the amp. Try backing the input signal off to 100mV and set the amp controls to what you know as "max clean volume" when a guitar with the volume up full is plugged in. You may get a clean signal. Advancing the amps volume control a little from there should get you to a clipped signal. And the power tubes should be well into clipping by then. If not there may be some error in the gain structuring of the design.

The input.?
Do i have my terms mixed up.?
The two outputs are on Pins 2 and 5.....and then i have my scope probe after the coupling caps (to the power tube grids) to avoid any "High" DC that this scope may not like.
Anyway.....just going out to my shed now to check all those Quiescent/DC and Power Tube bias numbers.

Hey Chuck...are you thinking i am using a 12A X or T7.? This is a 6SL7.
Anyway, let me get those numbers and try what you suggested.
Thanks Again

[Chuck H]

Ho! Yes, I was thinking of a 12a** tube. My bad. What I said abut the input voltage still applies

[trem]

OK...a couple of revelations. Though i do not think me being an idiot is much of a revelation.
I was Not Inputing 200mV PP...it was almost 500mV.

Here are the Numbers at idle.
V1a
cat = 1.42 VDC
plate = 157
V1b
cat = 1.45
plate = 154

Phase Inverter
cat = 48 VDC
plates are both about 170

Power Tubes are 6K6 (not 6V6 or 6L6).
I should have checked again, because it was about a month ago that i set the bias. I think they are at about 5-6 Watts. I will have to 2x check.
Negative grid is 28 and 30.
Plates are 302

OK.....in my own defense. I started this almost 2 years ago. I had One, new production Tun-Sol 6SL7 and i put that in the PI...so i needed one more.
A reputable dealer had the (NOS) JAN Philips 6SL7 at a decent price. I bought 5 of those. When hunting this noise, i went from one, two, three, of these NOS tubes in V1. The amp sounded the same with all 3, so i did not think it was a tube. I DID give them a light tap with a stick, and they Did ring a bit, but it did not seem too bad for V1...plus The All did it.
Just now, i was tapping on the front panel of the amp, and it was ringing the tubes pretty bad. Only V1 was sensitive to tapping. But now i have another New Production Tung-Sol. I slipped that into V1 and the noise is 80% better. I tried the remaining two JAN tubes, and they are all Sensitive/Microphonic. What can i say.?
So (until i get back to the trem) i do not think i have a noise issue.

But...back to the PI on the scope.
Now that i have 100mV going into the front of the amp, The PI output looks much better, but both outputs are clipping the positive half of the sine. Is that normal.?
(photo links below)
Thank You

http://i.imgur.com/D1hUI1e.jpg
http://i.imgur.com/i97DZk4.jpg

[Chuck H]

Yes, that's normal-ish. The PI may be biased a bit hot. As long as you have enough positive swing to get the power tubes to clip before the PI it's NBD.

EDIT: With the power tubes installed you will see power tube grid clamping at the bias voltage that won't affect the PI's negative swing. So that may be what you're seeing. To get a real feel for what the PI is doing you'll need to scope it with the power tubes removed.

[trem]

Aahhhhh...OK.
Will not get to that tonight, But i WILL give it a shot tomorrow.
Its not like i have all kinds of experience with a scope, but with this "problem" i realized i had never seen the PI output before. Its been interesting.
Thanks Again.....i really do appreciate all the time and effort you guys donate.

[pdf64]

Just to note that the waveforms inside a negative feedback loop will include an error signal component, so can be expected to look weird. Best to open the loop when scope eg the LTP plates.

[Chuck H]

I'll guess trem isn't using NFB on this amp. I don't see the tell tale pointy nipple on the negative swing at all. Interested to see if I got this one

[trem]

Oh Yeah.....NFB. There is So Little room in this chassis, that i did not install any. I am going to find the room.
1. You said the PI bias seemed a bit hot. What kind of cathode voltage drop should i be looking for with a 6SL7.?
2. Any idea on Resistor Value for the NFB resistor.?
My OT secondary (a single 8 Ohm tap) is not grounded. Does that ever come into play for some reason.?
Thank You

[Chuck H]

You said the PI bias seemed a bit hot.

I said it might be. I actually rethought that when I wrote the edit. Grid clamping is most likely what you are seeing.

Note your settings when you start to see the clipping. Then remove the power tubes and retest with the same settings and I'll bet you'll see a nice round positive swing

Technically (like "I" have any business ) there will be more voltage on the PI with the power tubes removed and this would increase headroom. But it's still a good way to note the affect of grid clamping on the PI waveform.

So as to specific bias resistor value... It depends on your PI plate voltages and your circuit intentions. The Valve Wizard site as well as a couple of others have write-ups better than I can cover here for how to design LTP's

As to a good feedback resistor value... Who says you need one? If you like the amp without NFB then don't use any. If you want to add NFB then the circuit values depend on a lot of things (including the subjectivity of tonal preferences) and I'd need to see a schematic.

[Malcolm Irving]

...My OT secondary (a single 8 Ohm tap) is not grounded. Does that ever come into play for some reason.? ...

I think it's only essential when you have NFB, but it's (probably) a good idea to ground one side anyway. That's for safety reasons, in the sense that any conductor can somehow become live due to a fault (probably a very unlikely one with a speaker cable - but nevertheless). If it is grounded it can only become live until a fuse blows or a breaker trips.