tone stack / LTP impedance problem??

[PRNDL]

I've narrowed down the problem I'm having with my tube amp to the volume pot, which is after the tone stack and before the long-tail-pair phase inverter.

The problem is scratchy distortion with notes played when the volume is between 8 and 9 (of 10). It's fine on 10 and at lower volumes.

I was able to narrow down the problem using the scope. The problem disappears when the probe is put anywhere after the volume stack. I narrowed it down to the volume pot itself.

One solution would be to put the probe on permanently, but the scope would be rather heavy to cart around. The other option would be to put some pF caps to ground with values similar to the probe (10 pF?).

I'm wondering if this is a design issue with the impedance of the tone stack into a long-tail-pair phase inverter.

The Soldano Atomic 16 has a voltage divider before the volume and a 120 pF cap between the LTP anodes. The Vox AC15 has a top cut pot and cap there. The Matchless Spitfire has the volume pot before the tone pot.

This makes me wonder if this is due to the same problems I'm having.

Currently I have a 12AX7 for the phase inverter. Perhaps a 12AT7 would be better due to its lower gain and higher transconductance.

Am I on the right track?

Does anyone know the correct values for the top cut pot and cap?
I have a 2M and 150K available.

[Wakculloch]

I would suspect coupling through the HT line. I know the PI should cancel out, but as the signal voltage here is so much larger than the input stage, any imbalance could lead to problems.
Easy enough to put in another node.
Maybe!

[Enzo]

If it were an impedance problem, then ALL the amps with this circuit would have this problem.

Have you tried a different pot?


Your scope probe not only has capacitance, it also has a high resistance.

And try scoping the signal elsewhere so yu can look at it without the scope affecting the symptom. What does it look like, could it be parasitics? The problem could be layout related, and hanging the scope probe on there is enough to damp them away.

And if nothing else, a 5 or 10pf cap stuck on there might work fine.

[Steve Conner]

It's probably parasitic oscillations, and the capacitance of the scope probe is damping them. A capacitor to ground in the same place as the probe should damp them just as well.

[PRNDL]

I tried several different pots with the same results.

Placing the scope after the volume pot (on the PI) also stops the problem, but not before it (on the preamp).

I was having parasitics, but they were fixed with lead dress. It may be that it needs a starter note, but the hard part is not being able to see it with the scope.

Today I'm going to separate the B+ to the PI and preamp. It's not easy since the wire is beneath the board, but it's pretty much the only thing I haven't tried.

[Enzo]

By definition a parasitic needs a "starter note." If an amp just sits there oscillating, it is an oscillation, not a parasitic oscillation. They call them parasitic since they ride in on the signal waverform. They occur when an amp is stable enough on its own, but when signal swing gets to a certain threshold, this litle oscillation on the peaks - or wherever - occurs.

You can't watch it on the scope where it happens, but you can watch it on a scope later in the circuit. Watch it on the grid or even plate of the power tubes. You can see the nature of the spurious signal.

[Steve Conner]

Here's a thought, have you tried a grid stopper resistor on the input to the PI? Do you have grid stopper resistors on the power tubes, for that matter? They're called stopper resistors because they stop parasitics.

Some parasitics can be so high frequency that they don't show on low-end scopes.

[PRNDL]

try scoping the signal elsewhere so you can look at it without the scope affecting the symptom. What does it look like, could it be parasitics? The problem could be layout related, and hanging the scope probe on there is enough to damp them away.

I added the extra PS node to separate the preamp and PI supplies, but the problem remains.

For grounding, all the PS caps and the PI are grounded together with a wire to the PT center tap, OT/speaker and power cord ground. The preamp cathode cap is connected to the volume / bass pots and the input jack.

I was able to view the parasitic oscillation, which was on the top of the sine wave of one side of the LTP PI output (or power tube input). It was a blur the clean sine wave (I have a 60 MHz scope). (By "on top", I mean that it occurred only at the top peak of the sine wave).

Here's a photo of the chassis.

[loudthud]

Don't put your scope probe on the plate of the output tubes, it will damage many probes. Your amp is right on the edge of oscillation. This is the type of problem that builders without scopes have the most trouble with. Are those 100 ohm resistors as grid stoppers for the output tubes? Try increasing them to something between 2.2K and 10K. The wire from the master volume pot, keep it as far as possible from the PI parts. I'd route it against the front panel in an S shape below the volume and tone pots and then a right angle to where it solders to the board.

[Wakculloch]

Here are some quotes from the "Handbook of line communication".
"High frequency oscillations sometimes occur in amplifier circuits, due to stray capacity and inductance; these are known as parasitic oscillations. They are usually the result of a tuned-anode tuned-grid circuit being set up......
The tuned capacities are the inter electrode capacities, and they resonate with the stray inductance of the wiring......
If the parallel resonant frequency of the anode circuit is slightly higher than that of the grid circuit....producing high frequency oscillations....
There are two main methods of combatting this effect. The first is the insertion in the grid lead of a resistance sufficiently high to neutralise the negative resistance component of the valves impedance....
It is important to make sure that it is connected as close to the electrode as possible."
Apologies for the over enthusiastic mis-diagnosis.
Try a 1K soldered directly to the valve base at the input to the splitter.

[PRNDL]

There are 1 ohm resistors with black cloth that connect to the power tube cathodes. These are used to match the output tubes and check bias.

The output tube grid stoppers are next to the PIO coupling caps. Currently they are 10K, which was originally 1.5 K.

Moving the PI input wire (yellow to pin 2) did seem to change the oscillation, but didn't remove it. Wakculloch - are you recommending that I replace this wire with a 1K grid stopper?

Loudthud - I'll try moving the volume pot wire as you recommend.

This is a PITA!!!

[Wakculloch]

Put it in series with the input wire,at the valve base end.

[PRNDL]

fixed!!!

It turned out to be the wire for the grid of one of the output tubes, which was next to the coupling cap at the input of the phase inverter (to the volume pot).

Moving both the wire and coupling cap away from each other minimized the parasitic oscillation to pretty much zero.

I put a coax wire over the wire to the grid ... removed the inner wire, threaded it over the existing wire, and grounded the outer braid on one end. Overkill ... maybe.

Now that I think of it, this wire goes underneath the board in the Matchless Spitfire layout.

One minor change!