Grid stoppers - Oscillation fix?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
So I'm getting a nasty audible oscillation and even it sounds like AM radio discharges if I crank the volumes to max and leave the FB open.

Circuit

An externally hosted image should be here but it was not working when we last tested it.


basically that + another triode for tone recovery after the tone stack, and a SE 6v6 for output.

after the tone stack there's a volume control so when the two volume pots are to the max it oscillates.
I'm wondering of a grid stoppers might help, also I've heard of ferrite beads at the triodes plate leads. Do I need a grid stopper also for the cathode follower grid ?

thanks for the hints.
 
Last edited:
Not sure if this will help but you don't need the cathode follower in there. You have a 100k slope resistor, so the plate of the triode with the 100k resistor will drive your tone stack just fine. This will simplify the circuit. Just couple these stages together with the cap of your choice.Usually, you use a cathode follower when driving lower impedance stacks. The CF is a high to low impedance matching circuit.
The usual setup of this configuration would be to interchange the 56k cathode and 100k slope resistors. And no, there usually is not a grid stopper going to the CF.
 
Last edited:
I would use grid stoppers on the first two stages but the first stage has a 68K in series anyway. So maybe just grid stop the second stage.
Also I would decouple the first two stages separately so there is no path for feedback via the B+.
I had a similar problem on a guitar amp and I found I had to get rid of the cathode bypass capacitors to stop oscillation. You could try just getting rid of one first.
 
Thank you nigel, good tips. I tried decoupling the plates, added a 22k resistor between the 220k and 100k plate resistors with corresponding 10uF electrolitics, improved the hum level, but did not stop the oscillation, it's weird because I've used this same circuit (vox-topboost) in PP and never had this problem, I'm wondering if the fact that this time is SE and there's current flowing all the time the OPT is causing the triodes to resonate somehow, when I put may hand near the opt it acts like a theremin!! I'll take pictures next time.
 
Take a photo or show us the physical layout of this circuit. The components and wiring from each plate to the next grid should be as short as practical, and as far as practical away from the earlier grid circuit components and the later plate circuit components.

It's possible for feedback/oscillation signal to go through the ground or through the Vcc, er, B+ between tube stages, but with the high impedance of tube line-level circuits, the capacitance between signal wires in different stages can be enough for oscillation. IMHO this, perhaps precipitated by layout, is a likely cause.
 
Yes Nigel, I tried all NOS (ECC83), Telefunken, KenRad and others. 5965 (GE). ECC82 (Miniwatt) and of course with less gain the oscillation tends to disappear, as I said the oscillation frequency changes when my hand is near the triode, using the metal shield for the ECC83 does not seem to really help, it just oscillates at a different frequency.

I will check the star ground (I believe I did but I'm not sure) and also post pictures.

Benb, I will take a few pictures of the "draft" layout.

Thanks again guys.
 
Last edited:
I agree. Aside from plain stuff like proper rail decoupling this circuit shouldn't require any modifications. It should work perfectly fine as is .... assuming it's put together properly.

Some of the suggested mods (e.g. get rid of the CF, get rid of the cathode bypass caps) would also change the tone of this circuit radically, which I suppose really isn't the goal in here.

So, hook it up to scope and find which stage oscillates. Further on, if it was left unclear the schematic displays only some of the entire circuit: There's also additional gain stage and a complete SE power amp stage. AFAIK, the oscillation could as well take place in any of those. IMO, it's somewhat pointless to randomly suggest mods while we don't actually even know what the real issue of the amp is and where it manifests itself circuit wise.
 
A few pics.
Don't pay attention to the first octal it's a pentode input and works perfectly fine, the noval it's a 12AX7, and the third socket it's a 6SQ7 and it's the CF in the circuit.
At this point I added a grid stopper (68k) in the second half of the 12AX7 and I have to say it improved a lot, but there's still some audible noise, like an FM radio not tuned. (this happens with the secondary of the opt floating, if I add FB the oscillation disappears.)

An externally hosted image should be here but it was not working when we last tested it.



After the 6SQ7 there's a tone stack (fender in this case) and then a high mu triode that feeds 2x 7C5 in parallel, the pentode in the picture is of course a 6J7 (EF37) that input sounds sweet, the problem is with the 3 triode circuit I described above, tho I remember this has worked for me before.


Another angle.

An externally hosted image should be here but it was not working when we last tested it.





valve layout.

An externally hosted image should be here but it was not working when we last tested it.
 
Each stage seems to have its input and output wiring/components very close together - even criss-crossing each other. This guarantees capacitive feedback. For example, the orange wire (input to CF?) wraps all around the CF valve socket. You need to think about socket orientation, so connections are shorter and more direct.

Keep component leads short at the valveholder end, to reduce stray capacitance.

The screened cable coming from the volume pot is grounded where? It should be grounded somewhere near the 12AX7 valveholder, where the valve gets its signal reference, but it doesn't appear to be. At this point you appear to be using the chassis as a signal ground. OK if you are happy with some hum.
 
Each stage seems to have its input and output wiring/components very close together - even criss-crossing each other. This guarantees capacitive feedback. For example, the orange wire (input to CF?) wraps all around the CF valve socket. You need to think about socket orientation, so connections are shorter and more direct.

Understood, yes, I only orient the sockets based on where the heaters wiring will be, but then I found this other problems..

Keep component leads short at the valveholder end, to reduce stray capacitance.

will try that today

The screened cable coming from the volume pot is grounded where? It should be grounded somewhere near the 12AX7 valveholder, where the valve gets its signal reference, but it doesn't appear to be.
it's grounded at the volume pot, I will invert the screened cable.

At this point you appear to be using the chassis as a signal ground. OK if you are happy with some hum.

sort of, the output stage and power supply ground is floating and "star" connected, that point is also chassis grounded and the preamp has its own ground point connected to chassis, I can disconnect that and wire it to the power supply ground, I have to say I've done this before and couldn't tell the difference, but I think that's me :p

Thanks for the replies, so much stuff to learn that is not in the books...
 
I agree. Aside from plain stuff like proper rail decoupling this circuit shouldn't require any modifications. It should work perfectly fine as is .... assuming it's put together properly.

Some of the suggested mods (e.g. get rid of the CF, get rid of the cathode bypass caps) would also change the tone of this circuit radically, which I suppose really isn't the goal in here.

So, hook it up to scope and find which stage oscillates. Further on, if it was left unclear the schematic displays only some of the entire circuit: There's also additional gain stage and a complete SE power amp stage. AFAIK, the oscillation could as well take place in any of those. IMO, it's somewhat pointless to randomly suggest mods while we don't actually even know what the real issue of the amp is and where it manifests itself circuit wise.
Why would the CF need to be kept.? Its purpose is to drive lower input impedance circuits and with that 100k sloe resistor, the CF it not needed. The
normal way is to swap the 56k cathode resistor and the 100k slope resistor. This is a different set up. and could cause problems. The design of this circuit would change the tone, for the better since the CF is not needed in this situation. Agree that the oscillation could be in the part of the circuit not shown.
 
You would be correct if the CF was preceded by a coupling capacitor, or never ever overdriven.
The DIRECT COUPLED cathode follower is a unique circuit with its very own tone contribution regardless of the following stage.
Mr Merlin valve wizard explains this clearly in his pre-amp book.

The other side of the argument from Mr Kevin O'C TUT is that FMV tone stacks sound better and have greater adjustability if not driven by a CF.
 
VMUNIX: Have you looked at the "oscillation in tube amps" thread (sticky) over on the Tubes/Valves forum? If you don't want to read the whole thread go straight to Post 20.
To the list of high probability of oscillating examples he gives I would add 12AX7's with an anode resistor greater than 100k.
 
I have had a similar problem with a 12ax7 front end.
I noticed if I removed the bypass capacitor on the first stage the oscillation stopped.
This does 2 things, 1 it reduces the gain of the first stage but also disconnects the ac coupling from b- to the cathode.
It turned out my b- line routing was poor. I turned the b- line into a star ground and that improved things no end. The b- line was being modulated by stages further along in the amplifier.

I have also noticed with 12ax7 that you can get hum into the first stage from the heater. To get around this I also use the second triode in series with the first. The hum from the first two stages then cancel each other out.
So care has to be taken which triode is used in what position. In a 12ax7 there is a quiet stage and a noisy stage so its best to put the quiet stage first. Sorry I cant remember offhand which is which.

There are numerous routes for positive feedback.
1/ B+
2/ B-.
3/ Interelectrode.
 
Last edited:
VMUNIX: Have you looked at the "oscillation in tube amps" thread (sticky) over on the Tubes/Valves forum? If you don't want to read the whole thread go straight to Post 20.
To the list of high probability of oscillating examples he gives I would add 12AX7's with an anode resistor greater than 100k.

Now I did. Thank you I will implement a few changes now and post the results.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.