Oscillations problem help

[vintagekiki]

It is in 150 Mhz range

Question
How you hear that frequency, or how it manifests on sound.
What instrument did you use to measure exactly 150 MHz.

[catalin gramada]

Is relative to have an ideea in what range of freq is. The noise at the both ends of pot track is more like a buzz/ floor noise. It apear suddenly close to the ends position and is consistent on those fraction of track till the phisical limits of the whiper, like a switch...Is very strange as it act similar at both ends. I unmounted the pot and checked and nothing is wrong with it...the case is well bolted to chassis, the ground return goes directly to cathode and nothing else linked to it, the signal wire well shielded with the shield tied to ground filter node...

[Helmholtz]

According to the ‘grid stopper value >= 7/gm’ guideline

Where does this rule come from?

I agree that 100k may be overkill. I would start with 1.5k directly at the tube socket.

Another method to stop the oscillation would be a small value anode bypass cap.

But an audio circuit oscillating at 150MHz is nuts - why on earth does it have any gain up there?

Well it's a high gm VHF tube. Using RF tubes might require using RF layout rules.

[vintagekiki]

QUOTE=catalin gramada]It is in 150 Mhz range[/QUOTE]
I forgot to ask
On which speakers do you hear those 150MHz and
which signal frequency (Hz) is connected to the amplifier input.

You say "The noise is more like a buzz / floor noise" It is audio spectrum, and the speaker can reproduce.
The 150MHz frequency you mention is RF wave (electromagnetic spectrum), it is registered via antenna and radio receiver.
I am vitally interested on which magic and on which speakers you hear the 150MHz (VHF band), unless you have connected the input of your amplifier to the output of the some radio receiver.

Useful links
https://en.wikipedia.org/wiki/Audio_frequency
https://en.wikipedia.org/wiki/Radio_frequency
https://en.wikipedia.org/wiki/Hearing_range

[catalin gramada]

QUOTE=catalin gramada]It is in 150 Mhz range

I forgot to ask
On which speakers do you hear those 150MHz and
which signal frequency (Hz) is connected to the amplifier input.

You say "The noise is more like a buzz / floor noise" It is audio spectrum, and the speaker can reproduce.
The 150MHz frequency you mention is RF wave (electromagnetic spectrum), it is registered via antenna and radio receiver.
I am vitally interested on which magic and on which speakers you hear the 150MHz (VHF band), unless you have connected the input of your amplifier to the output of the some radio receiver.

Useful links
https://en.wikipedia.org/wiki/Audio_frequency
https://en.wikipedia.org/wiki/Radio_frequency
https://en.wikipedia.org/wiki/Hearing_range





[/QUOTE]

There is nothing more to add except to mention the input of first triode was shunted to the ground directly to cathode ground point. Same thing with 8 ohm speaker or 8 ohm resistor connected to the output. The signal was scoped into the second triode plate.

[Helmholtz]

In his book "Designing High-Fidelity Valve Preamps" Merlin elaborates some more on spontaneous RF oscillation.
For mitigation he recommends to use a grid stopper of at least 2/gm and a ferrite ring/bead on the cathode lead.
He also mentions that high gm tubes are more prone to oscillations.

[vintagekiki]

There is nothing more to add except to mention the input of first triode was shunted to the ground directly to cathode ground point. Same thing with 8 ohm speaker or 8 ohm resistor connected to the output.

If I understood correctly on amp is a short circuit input. You hear some undefined noise on the speaker.
This can be a problem with the ground loop.

I don't understand your graph in post #34 and the mention of 150 MHz in post #21
OT cannot transfer RF (150 MHz). LC circuits are used for these frequencies.

Useful link
http://www.geofex.com/article_folders/stargnd/stargnd.htm
Star Grounding in Tube Amplifiers

[catalin gramada]

If I understood correctly on amp is a short circuit input. You hear some undefined noise on the speaker.
This can be a problem with the ground loop.

I don't understand your graph in post #34 and the mention of 150 MHz in post #21
OT cannot transfer RF (150 MHz). LC circuits are used for these frequencies.

Useful link
http://www.geofex.com/article_folders/stargnd/stargnd.htm
Star Grounding in Tube Amplifiers

No, the picture from the scope is not from the amp output but from second triode plate. I posted a drawing of the circuit who involved my issue in my first post, did you notice any possible ground loop in please ?

[catalin gramada]

I noticed another thing. As I mentioned the oscillations at the ends of the pot track didn't apear if I plugged a guitar I to the input but just if the input is shunted by jack. With a guitar pluged in I can open and close the volume pot without problem...but...now I noticed with a guitar pluged in and volume pot closed this osscilation apear at the end of the track of the volume pot of the guitar...which basically meant the same thing as a shunted input.

[catalin gramada]

Real layout

Note: 47k to V1 socket
all signal wires to grid and pot are shielded with the shields tied to a rail tied at - pole of the filter cap.

[J M Fahey]

Question
How you hear that frequency, or how it manifests on sound.
What instrument did you use to measure exactly 150 MHz.

Dear VK: obviously nobody hears 150MHz by themselves, but that strong oscillation messes with tube operating point, intermodulates with expected signal, etc.

What we hear , and are well within the Audio range, are the "artifacts" of that oscillation.

For me itīs very common to have an amp brought in for repair (or even detected on stage, without even opening it) with this symptom (basically same what Catalin mentions): rising/lowering Volume or Treble control or activating Bright switch , there is a point where you her a "ffoooompppp!" sound, looks all the way like a dirty pot, to boot it "always happens in the same spot" , doubly reinforcing the dirty pot theory, .... and itīs actually the point where oscillation starts/stops.

And once it starts sound gets dirty/grainy, specially "sandy/gritty" .
Sorry but those words describe it best, once you hear and recognize it youīll also use them.

As of "a 150 mHz oscillator would require specific elements" (basically L and C), they are there, hidden in plain sight!
PARASITIC L and C that is.

And thatīs why we find 150MHz and not 15kHz (or lower) , parasitic LC elements are tiny.
And I 100% suspect layout/grounding, not "a bad tube".

Also, once generated, those "artifacts" *are* Audio, and travel all the path to the speaker.

[catalin gramada]

One more fact I found measuring dc leakage "current" to the grid I get 85mV with the 1M pot completly open but with pot completely closed I get -0.7 mV. Somehow sweeping the pot it seems I get 0V in the same position where osscilation apear. Are this related or pure coincidence please ?

[vintagekiki]

Maybe I'm wrong
12AT7 is a driver tube, has less gain and higher anode current than 12AX7.
Personally, I think that by choosing the operating voltage (470VDC) and the peripherals (Ra and Rk) that is suitable for the 12AX7 tubes, you have brought the 12AT7 to an operating point that is not intended for it.
Before any intervention, simply install 12AX7 instead of 12AT7.
The ground loop definitely exists.

[Helmholtz]

One more fact I found measuring dc leakage "current" to the grid I get 85mV with the 1M pot completly open but with pot completely closed I get -0.7 mV. Somehow sweeping the pot it seems I get 0V in the same position where osscilation apear. Are this related or pure coincidence please ?

If the DCV at the 1M pot is negative wrt ground, it's just the voltage drop produced by normal grid leak current.
Normal grid leak current is a negative current flowing out of the grid. It's caused by some electrons landing on the grid.

If the voltage is positive it might be caused by some leakage of the coupling cap.

I like the ground wiring proposed by VK better, because it uses separate grounding wires for the shields.
Reason is that the shields act as capacitors and the wires have inductance (a straight wire has an inductance of around 20nH/cm).
So if you use a common ground wire for the shields, you will get some HF coupling between the shielded wires.

[catalin gramada]

VK Thank you for the grounding layout. I.ll give it a chance but same time wanna understand. There are two pics of my actual layout, second just to figure how my shields was related to the ground. I will start removing completely the shielded wire from the circuit and replacing with normal wire. Then basically you drawing show to move the star ground to the first cathode. Just wonder if it made a difference as there is phisical one point for both. Please let me know if it need to leave some lengh of wire in between? Second, I don't understand from a grounding perspective why 1M grid leak resistor is need it as time 47k input is shunted to ground please ?