Booster tube in front of V1

[sbirkenstock]

Hi Everybody,

I put a booster tube (12ax7) "in front of" a JMP-Style amp. So just an extra tube (only one side used of course) before the original V1.
Works fine and sounds good (better turn the master volume down...)
I checked with my USB oscilloscope and got some "strange" findings though.
I apply a 440hz sinus wave into the "booster 12ax7", plate resistance is 100K, 1k5 at the cathode with a 22uF bypass cap.
I connect a 0.002 uF cap to the plate, goes directly to a 1Meg volume pot, and the middle lug is connected to the grid of V1.
Now the surprising part:
I measure the voltage at the grid of V1, slowly turning the volume up.
Till about 2,5V (5 V peak to peak) the sinus wave swings around 0 volts just fine.
If I turn the volume higher, the positive peak (at the grid of V1) will not exceed 3 volt, but the negative peak will get very negative, even lower than minus 100 volts when the volume pot (after the booster, in front of V1) is all the way up.
I added a 220k grid stopper between the pot and the grid of V1, did not change anything noticeable.

I also checked on other amps, seems to be everywhere.
Grid positive voltage "ends" between plus 2,5V and 3V.

I assumed "saturation" or "cut off" would happen at the plate, but not at the grid.

Any hints what is going on?

Thanks a lot,
Stephan

[pdf64]

There’s effectively a diode between grid and cathode. When the signal gets big enough (>= bias voltage), it conducts.
See grid current clipping, p19 section 1.14 http://www.valvewizard.co.uk/Common_Gain_Stage.pdf

[Stevem]

Also you should keep this in mind.

When a gain stage amplifies signal what goes in on the grid for example as the positive half of a sine wave comes out as the negative half of the wave on the plate.

In short a gain stage flips phase unless it’s wired as a cathode follower which then has no gain, but no phase inversion either.

[sbirkenstock]

Thank you guys!
@pdf64
perfect answer (with great documentation)
I just figured out that I downloaded this file already in 2014....