Help!! problem with Non HRM with FX loop

[shane]

Hi Folks
I'm Building a 100W 4x KT88 non HRM with the Dumbelator loop built in.
I've kinda followed the schematic Heisthl posted (the PV conversion one).

The problem I have is I have a major oscillation problem even on the clean channel the osc is about 10khz with everything full up and falls to about 7.5khz when I turn the master down.

Now if I remove the FX return stage or even bypass it with a jumper the osc stops!
I have tried to make my layout follow the various posted ones and have tried shielding most sensitve wires moved stuff about but no joy.....

I've tried 12ax7,at7 and au7 in the loop but same thing

Its getting into hair tearing out territory now
I'd be grateful for any suggestions...

Cheers
Shane

[mlp-mx6]

Do you have a scope? Can you see the oscillation?

[Fischerman]

Post close-up pics if possible...those will make it 1000x easier for people to see any possible culprits.

What are the voltages on those two triodes? How is your power supply and power dropping resistors configured...i.e. how/where are you getting the supply voltage for the FX loop tube?

Also, when referring to the 'Master' try to be clear in no uncertain terms exactly what you mean. Clean master? OD Master? Or, for example, in my amp I labeled the Return Level control the 'Master Volume' because that's how it behaves to the user.

[mlp-mx6]

Especially if you can see the oscillation on a scope - here is how Gil Ayan described the debugging procedure in another thread that I used successfully.

Get a .001 cap and clip it across each plate resistor, one at a time, starting at the PI and working backwards to the input. The furthest back resistor where the cap removes the oscillation is the section you need to focus on.

Just FYI - on another ODS amp I had an oscillation in the PI that was solved by rerouting the feedback wire (the one from the speaker out back to the PI tail).

[shane]

Hi Guys
Thanks for the replies.
I will try and post some pics tommorrow if I can.
I have configured it like fischermans in that what is normally the master in the non HRM is labelled preamp out put and the return level is called master turning either of thes down reduced the osc.

With a scope i can see the osc right back to the output of the first stage!

the supply for the loop taps of the PI supply with a 15k resistor 22u cap for the return stage followed by a 100k/ 22u cap for the send.

I'll try the cap across the plates trick tomorrow and report back
Cheers
Shane

[Fischerman]

You might consider changing the loop supply from the PI node up to the screen node but I'm not saying this is causing the oscillation. Stuff can couple through the power supply and even ground in addition to just through proximity to other signals. Actually that coupling is always there...it's just a matter of degree.

[dogears]

Tap the loop B+ off of the screen supply. Add an extra cap to decouple it one extra level, 33uf is good. If you have around 450V, I'd use something like a 27K to feed the cap, then use another 22K to 27K (up from your 15K) to feed the recovery amp cap.

You should hear an improvement in the overall amp response and tone.

[shane]

Thanks Guys
I'll try changing the tap to the screens, They are at about 550V since the KT88's are running up at 560V or so but I can work out the resistor values.
Cheers
Shane

[dogears]

Shane, try adding that extra filter node. Then feed the loop from it.... So, you have a total of 3 caps for the loop.

Thanks Guys
I'll try changing the tap to the screens, They are at about 550V since the KT88's are running up at 560V or so but I can work out the resistor values.
Cheers
Shane

[shane]

OK.
So I changed to the screen tap and added the extra filtering no joy

Tried the 1n cap trick and it has most effect on the plate resistor of the recovery stage.

checked all my wiring cant see anything obviously wrong or too close.
tried rerouting NFB wire nope!

jumpering out the return stage kills oscillation dead so I guess its gotta be something to do with that stage.

one funny thing though, instead of relying on the switching jacks in the send and return to pass the signal when no effects are use I installed a switch which connects the return jack to the top of the send level pot bypassing the jacks.
Now when i have the switch in the position where the jacks are in circuit but with nothing plugged into them you'd think I'd get no signal BUT i do get a very reduced one and guess what? .... it dont oscillate!

Now I'm wondering if thats connected with the problem. I dont see how but it might be the sign of some other coupling issue. What do you think?

Anyway as soon as I get chance I'll get some photos up so you can see inside it.

Lord my hairs getting thinner by the day!!!
Cheers
Shane

[dogears]

Shane,

Sorry if I implied my filtering scheme would fix your issue. I was merely suggesting it as a better method once you fix the oscillation.

[heisthl]

Shane - I've never had oscillation problems with an internal loop but I have experienced a lot of hum issues with loop input/output and send/return lead routing. If your bypass switch is any distance away from the jacks it will have to be shielded. If you have drawn up a schematic and/or layout specific to your build and/or have taken pictures, I'd be happy to look at them. I've also gotten into the habit of putting a resistor before the loop input (10k to 100k) and one on the output (10k-100k) But you've probably done that if you were loosely following a schematic I posted.

[JimiB]

I've also gotten into the habit of putting a resistor before the loop input (10k to 100k) and one on the output (10k-100k) But you've probably done that if you were loosely following a schematic I posted

what purpose does this serve?

[Fischerman]

Now when i have the switch in the position where the jacks are in circuit but with nothing plugged into them you'd think I'd get no signal BUT i do get a very reduced one and guess what? .... it dont oscillate!

That's what I was talking about when I wrote:

Actually that coupling is always there...it's just a matter of degree.

You're probably just hearing that coupling. For example, if you disconnect the wire/connection from the PI input cap and have a medium to high gain preamp and play through the amp you'll still hear a faint, thin, signal because of that coupling that is always going on. It's very possible that you just need to get this coupling under control.

Generally speaking, the grid wires are the most sensitive for 'picking up' those stray signals and the plate wires are the worst 'transmitters' of these stray signals. The earlier in the preamp the more sensitive to 'picking up' stray signals (because they will just get reamplified so much) and the later in the preamp are the biggest transmitters (because the signal is so big there). Some of the PI wires are ginormous transmitters...even the PI input grid wire is a big transmitter.

Just some ideas to work with. If it ends up being a huge buzz-kill and you just want to play the amp for a while...just put the snubber cap in there and use the smallest value you can get away with. Revisit later...after your hair grows back.

[shane]

Yeah saves on barbers fees I guess
Heisthl - i put a 100k before and after the loop, I hadnt anticipated any issus with it but ah well.. I guess it will be something stupid I 've overlooked.
I'll get the camera out and post some pics.

Dogears - I didnt think the changes you mentioned
would cure it but I'll keep them in 'cos it does sound better when I play it with the changes you mentioned -thanks mate.

you're all REALLY helpful and I APPRECIATE IT LOTS!!
Thanks guys
I'll keep you informed
Cheers
Shane