Negative Feedback Problem with JCM800 EL84 Build

[kenny1547]

Hey guys!

Ive built a JCM800 Clone with the 18W EL84 Marshall Poweramp but have a little Problem...

It has a switch to go from no negative feedback to low and high negative feedback. It works flawlessly with no negative feedback, but either one of the other two settings results in nasty oscillation, and it doesnt even depend on phase, swapping the output transformer secondarys didnt do anything. The only way to get it to be stable with NFB is to turn the presence up to max? Maybe someone has an idea what could be wrong with it

Could it be too much NFB? It definitely is worse on one of the NFB resistors, but i loosley followed layouts from robrobinettes site and he has a 6v6 jcm800 with a 100k@8ohm NFB resistor in there, so the 75k should be okay at least, right?  Or a phase problem somewhere else?

NFB resistors are 27k and 75k on the 4 ohm tap of the OT, presence control is the normal JCM800 presence with a 25k pot, 100nf cap and 4,7k resistor, but ill attach my layout too, maybe something obvious i didnt see.

It also doesnt depend on volume or gain, stays at the same loudness pretty much even with both on 0

All help appreciated!

[sluckey]

It's not easy to swap OT secondary leads when you have multiple taps. Better to swap the primary plate leads (pin 7). Or the grid leads (pin 2). Don't swap plates and grids, only swap one, whichever is easier. Give that a try.

[kenny1547]

Oh, sorry, i meant primary of course, the secondarys stayed the same, i swapped the leads on both pin 7s. I also did get it to squeal just how it does when you have positive feedback. It now is the other way round, still have that nasty farty oscillation (dont even know if it is oscillation to be honest, might be something different) sound though with NFB on

[VMS]

Where is the nfb switch located in actual amp?


And have you tried if moving nfb wires has any affect on the oscillation?

[kenny1547]

Its next to the outputs about where its located in the layout too, the wire from the switch to the PI input is isolated too, wiggling it around doesnt do much

I once had the same problem( or similar at least ) with a Friedman Shirley Clone, switching the NFB and presence circuit to a Bassman style circuit helped there ( 25k pot to 5k and have it be the 5k resistor to ground for the PI etc ), but i really dont want to mess with the JCM presence now, and i already built a regular 2204 too so the only thing i could think of that might be wrong is the amount of NFB, is there a more or less easy way to calculate or measure if that is the case?

[sluckey]

Since you suspect the amount of NFB just change one of the resistors on the switch to 100K like the 2204. Does that cure the problem?

[kenny1547]

Okay well, for anyone finding this topic through search or Google in the Future...

Remember to ground your outputs for Amps with NFB.... 

Thanks for all your suggestions though!

[vampwizzard]

Hey, glad you figured it out!