Preamp negative grid voltage, oscillation?

[Michael Allen]

I rebuilt an 18W Trex into a trainwreck rocket with a switchable second gain stage. 2xEL84 output, LTPI, cathode follower driven tone stack. It's essentially an 18W rocket right now for all testing purposes.

I'm getting weird issues in the preamp stages, lots of negative voltage on the second stage grid after the volume control. Its around -30V. This is causing the signal to clip on one side and obviously sound wretched. I think it's oscillation causing it. Paralleling the second stage plate resistor with a big cap (0.02u) makes the sine wave at the second stage grid look more normal at full volume, but at much lower voltage output, and if I turn the volume down the signal regains that weird one sided clipping. The voltage on the grid is about -10V with the plate resistor snubbed.

Any ideas on what could cause it? Where should I look for oscillation? The circuit before (Trex) didn't have this issue, so I think the majority of the layout and parts placement should be fine....

[Michael Allen]

I should add, if i disconnect the volume wiper to the second grid, the signal is normal and there is no negative voltage on the wiper or grid. Have changed tubes, added filter caps, added grid stopper...

[Alex R]

The clipping is a product of the voltage sending the tube off-bias, by the sounds of things, ie a symptom not a cause. A few things to try - unplug the tube and thus see whether the tube or something else is producing the voltage, and check the cathode voltage, which if not in the 1-2vDC area would point to issues with the cathode resistor/cap - are they grounded correctly, etc. Is the volume control grounded correctly, and thus giving the grid a ground reference?

[Michael Allen]

Alex, yes I noticed the clipping and tracked it down to the negative voltage being generated on the grid. Google came up with a random snipit that oscillation can cause grid voltage buildup (or builddown as it were.)

I've got reasonable cathode voltages, about 1.2V. Have tried replacing cathode R's and C's... Volume does correctly control the signal level...

Thanks!

[Alex R]

Yes but the stage is going to get its grid ground reference through the volume control, so that should be checked - maybe just measure resistance to ground from the grid to be sure of it. 1.2v on the cathode shows the cathode refers to ground ok. The reason I'm going on about this is that drifting voltages suggest poor ground reference somewhere.

It may or may not be an oscillation issue. A scope would tell you for sure. In the absence of one - ground the stage 1 grid - if that stops it then oscillation looks more likely.

A schematic of this amp would help.

[Wilder Amplification]

Did you use the wrong value cathode resistor on that stage (perhaps mistook a 82K for an 820R or something like that)?

[Chuck H]

If you disconnect the wiper (assuming this is where the following grid is connected) how can you judge the quality of the signal??? Is there -V on the pot when disconnected from the tube grid??? As mentioned, a schematic would help. Not a reference schem like a Rocket or Trex, but an actual schem of your circuit as it is.

[Michael Allen]

Chuck, I meant if I disconnect the wiper from the following grid the signal on the wiper is unclipped. Cathode R values are all checked and replaced.

I'll check to see if grounding the input grid stops the negative voltage buildup on the next stage grid and I'll try and put together a schematic today...

[Steve Conner]

This -30V, is it with the amp sitting idle, or driven by a signal generator?

[Michael Allen]

Hey guys,

Here's the schematic I've drawn for the rocket channel. There is a switchable stage in between stage 1 and 2 as shown but i didn't include it because i have it disconnected until i get this part fixed.

The -30V on the second stage grid/volume wiper is when the amp is driven with a signal generator. If I ground the input grid or mute the signal generator the voltage disappears and grid is at 0V. This indicates an oscillation and not a bias problem or grid reference. The second stage grid measured 500k to ground when volume is full up.



Actually the first stage has a 100k plate resistor

[pdf64]

Try a grid stopper on that 2nd stage grid, 10k-100k, one leg soldered to the tube socket terminal.

[Michael Allen]

Nothing I've been able to do is helping. I've stripped down everything except the first stage, replaced the PT (unrelated), rewired heaters and replaced can cap with stack of caps. I think every wire has been moved or replaced.

I've got the first two stages only in the amp. If I wire up each stage and put in a sine wave, it looks good on the wiper of the volume pot. The second I connect the wiper of the vol pot to the grid of the second stage, negative voltage builds up on the grid/wiper and the signal clips. There isn't negative voltage buildup on the grid of the first stage though so it doesn't seem like it's mis biasing the first stage. So you would think the second stage right? Well if i disconnect the plate of the second stage, the signal on the wiper of the volume pot is still clipped. Only if I disconnect the cathode of the second stage does the signal return to normal. I don't get it. If there's no anode voltage then the tube is off, why would touching the grid of the triode with the volume wiper cause this weird clipping?

I've replaced the tube, added 100k grid stoppers to both stages, tried shielding from vol pot to stage two, wired it so it's using a triode off of two tubes, fed the stages from different B+ nodes, rewired the grounds, given each stage separate returns to the star......

This is the most infuriating project I've ever worked on, but its nearly impossible for me to give up on projects..

[Chuck H]

Are you saying that you can inject a signal at the second stage and all is well, and you can inject a signal into the first stage and it's output is also fine, but when you connect the two stages via a pot you get -V building up on the second stage grid and the pot wiper?

[MWJB]

Why is the voltage at your cathode follower 60vdc? Should be roughly the same as the preceeding plate?

I think we need to see the cicuit board & layout, I'm concerned because the input triode of the cathode follower is called "V3A" & your input tube is "V2", what does V1 do? Your PI is labelled "V6A" & "V7A" how many preamp tubes do you have?

100K Mid pot?

0.1uf for the PI input?

With your signal generator test, how much clean W do you get at the speaker?

[jmaf]

My 2 cents : check the voltage directly between pins on the tube, not between ground and pins, but between cathode and grid, with the positive lead on the grid you should get -1.2 to -1.5 VDC max. Keep the master volume down, or this will 60hz your windows.

Edit: Double-check R5 as Wilder A. said, I've been guilty of that one before, using 82k instead of 820R, especially with carbon film resistors I have difficulty reading the values under fluorescent light.