Trem switching weirdness HELP PLEASE

[jimboyogi]

Hi all. I am in the final stages of finishing a custom build amp. The chassis is all wired up, and almost everything works perfectly now, except for one one strange thing.

My trem oscillator is the same basic design as this one on the Valve Wizard's site -



The differences are that the cathode is biased with a red LED instead of resistor/capacitor, and I have changed some R and C values to alter the speed range of the oscillator.

The trem works perfectly as is, but when i try to engage the switch to turn it off, instead of stopping oscillations, the trem just oscillates at a fixed frequency. Altering the 'speed' pot does not change the frequency of this 'trem turned off' oscillation. The freq is ~ 50Hz. If I try grounding the first or third feedback network junctions, i also get fixed freq oscillations, but just at different frequency at each junction.

The wire that connects the switch/ground to the feedback signal point runs ~ 15 cm to the edge of the board, where the footswitch wiring attaches. I have not connected the footswitch wiring yet, I have just been using wire jumpers to test first. The wire was initially unshielded, but running beside the chassis for most of it's length.

So far I have tried -
Changing the connecting wire to a shield one, with the shield grounded at one end - no change
Jumpering a 10uF cap across the red LED, in case for some reason it wasn't effective at AC bypassing - no change
Trying grounding the long 'off' jumper wire at various points in the grounding scheme, including at the oscillator PS node - no change
Using a very short jumper wire and grounding right at the trem circuit position - problem solved

Now this short jumper works, and is how I had the circuit set-up when i breadboarded it. But it doesn't help me much now, as i need the wire to go all the way to the footswitch! What it may do is give a clue as to what the problem is. I suspect that it is related to stray capacitance, but any help would be greatly appreciated

[Jazz P Bass]

Try 'lifting' the ground to the 1M pot.
That should stop it dead cold.
I think you are correct in your assumption that your long wire is acting as a capacitor.

[Enzo]

And instead of thinking of 50Hz as oscillation, perhaps think of it as hum.

[jimboyogi]

Thanks for the responses Jazz P Bass and Enzo. I do really appreciate the way that you share your knowledge and advice, this forum has taught me a huge amount, and obviously i'd better keep on learning!

JPB I tried lifting the ground reference from the 'speed' pot, and it did stop the oscillation beautifully. There is a very slight pop when the switch is closed again after being opened.

Enzo, I was just guessing about the 50Hz frequency of oscillation when the 2nd feedback node was grounded. I just fired up my scope and measured the frequencies of the oscillation when each node was grounded via the shielded long wire. I was nowhere near it!

Node 1 - 39 Hz
Node 2 - 15 Hz (this is the node tied to the cathode bias, which I intended to switch)
Node 3 - 1 KHz

I don't know what is going on here, but the simplest answer is to just switch the 'speed' pot ground as you suggested JPB.

Thanks again for your help, and if anyone knows why i am getting the unwanted oscillations, and how to solve, that would be great to learn.

[Steve Conner]

I guess it must be some sort of RF parasitic oscillation problem, as there is no way a low-frequency signal can get past the short. Try a stopper resistor of a few hundred ohms in series with the footswitch wire, right where it connects to the node.

[jimboyogi]

Thanks for your advice Steve. I tried a 100 ohm resistor in series, between the 2nd feedback loop node and the grounding wire. The 15 Hz oscillation was still there. It's a weird one.

[Steve Conner]

Try a larger resistor then, see if you can get the oscillation to stop or at least change frequency.

Why is the second node returned to the cathode rather than grounding it? Is this supposed to generate a transient that kicks the oscillator into action promptly?

[jimboyogi]

Exactly right about the cathode grounding/transient kick start, as described on Merlin's Valve Wizard site.

I'll try some bigger series resistors tomorrow, and see if I get any frequency change also.

[jimboyogi]

OK I have just tried a 470 ohm in series with the switch wire. Attached the resistor to the 2nd feedback node, the switch wire to ground.
When I connect the switch wire to the feedback node side of the resistor, I get the 15Hz oscillation (as per previous problem)
When I connect the wire to the outside of the resistor (in series with the node), the oscillation is gone, and the trem turns off completely.

So you got it right Steve, and it must be HF oscillation, that is quelled by the series resistor! 100 ohm was not a high enough resistance to stop it, but 470 ohm is.

I think that I will use the 470 ohm resistor and the 2nd feedback node as the switch. It has one advantage over the 'speed' pot ground lift method. Using the pot lift, when you lift the ground to stop the trem, the trem carries on for a short while before stopping, maybe one cycle. With the feedback node grounding, the trem stops immediately when switched.