Marshall 2203 (JCM800) oscillation with mods

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
It is often reported that the 100nF causes oscillation and the series resistor is reportedly the cure.
I don't contest it. I just say that the 100nF is there to increase HF gain, and inserting a resistor in series goes against it. I believe it's a quite convoluted way of doing things. Two steps forward, one step back.
 
Which jcm800 are you starting with?

Most 800's don't adhere to strict Buss Grounding rules. For instance, the currents flowing through the Midrange pot and the master volume ground connection may be more effectively connected to the Gain control ground. Whereas in strict Bus Grounding, the midrange ground and Master volume ground should be connected to the ground of the phase inverter. And the Gain ground should be directly connected to the ground connection of the 10k cathode resistor. I don't know what you would call it, but the random ground scheme of jcm800's limits how much gain you can get out of them.

If you are going for a Slash mod, I would recommend looking at some threads on ampgarage or sloclone. This is a known issue. Strangely enough, 2k7||680nF doesn't have the same problem that the 10k||100n has.
 
I don't contest it. I just say that the 100nF is there to increase HF gain, and inserting a resistor in series goes against it. I believe it's a quite convoluted way of doing things. Two steps forward, one step back.
Yes! I agree with this and your other comment. Matador hasn't said, but this sounds like a Slash mod. The mod is known to be on the edge of harshness, but people do love it and supposedly it records well. It may be a convoluted way of doing things, but at this point it's like a historical fact. In the end the mod can seem so harsh that it sounds harsher than the album, and the series resistor takes the harshness down enough that it may even sound closer to the album once the mixing and mastering has been done. This is just me relaying info from forum threads on Slash jcm800 mods as far as I have understood what is going on.
 
Sounds like there's coupling in that preamp section due to board layout. You could go partial CJ and rewire the problematic preamp point-to-point or on tag strips. 3D and air are your friends in these cases.
 
Matador hasn't said, but this sounds like a Slash mod.
It could be: I'm not sure...I'm not an expert on Marshall mods. Some of the other changes in your schematic I can't see on this amp, so it must be some sort of hybrid (or half modded).

Most 800's don't adhere to strict Buss Grounding rules.
I would describe as "hierarchal grounding rules", however indeed as you say, it doesn't follow much of any scheme that I can tell, with myriad connections strapped to the chassis (especially filter caps).

With 10K and a 0.1uF, the gain turn is roughly around 150Hz, whereas 2K7 and 680nF turns an octave lower at around 80Hz. Maybe 2k7 and 330nF would be closer to the same voicing?
 
Matador: So it is someone else's mod. My guess is hat the inspiration is Slash. I would recommend the series resistor, but the source of the oscillation may always be a mystery. Another option is to do "Larry Grounding" which moves all the chassis grounds all over the place and is a known mod that people do to Marshalls to tame this stuff.

Cold_Clipper_Meme.jpg
 
The turnaround frequency may appear to be similar, but the overall response is very different.
I think I was arguing that indeed they would be different. :) I was curious if 2K7 || 330n might give a similar frequency response, albeit at lower overall gain. The 1/gm factor would be roughly equivalent between the two cases, right?

Also, doesn't bypassing the cathode render some of the "cold clipping" effect as moot, given that the cathode appears to be at AC ground (at least at "high" frequencies)?

Another option is to do "Larry Grounding" which moves all the chassis grounds all over the place and is a known mod that people do to Marshalls to tame this stuff.
I can't seem to find a clear diagram of this...did you have something specific you were thinking about?
 
Also, doesn't bypassing the cathode render some of the "cold clipping" effect as moot, given that the cathode appears to be at AC ground (at least at "high" frequencies)?
So-called "cold clipping" describes the operation at low current of the triode, about 0.3mA, when the other stages, with a 2.7k cath res operate ar 0.7mA, and a Fender preamp operates at a little over 1mA.
A direct consequence is that the anode sits quite close to the B+ rail, which means that the positive peaks are clipped earlier than the negative ones, which results in significant 2nd-order harmonics.
Whether the cathode is bypassed or not does not change the fact that the output max level is changed in the same proportion.
To resume, the clipped output of tubes with a 10k clip is one third than those with a 1.5k and half those with a 2.7k.
Now the gain of the stage with an unbypassed 10k is about 9dB lower than that with an unbypassed 2.7k, so clipping will appear at about the same input level, but with a much lower output for the "cold clipper".
Now bypassing the cathode will increase gain, so clipping wiil appear earlier than those unbypassed.
Here is the simulation of the three opions,
In green, the unbypassed 10k cath res
In Red, bypassed with a 100nF.
In brown, 2.7k bypassed with 330nF
It's easy to see that the responses are very different and that clipping is almost non-existant for the unbypassed stage.
At this particular setting of frequency and level, the 100nF bypassed cold clipper has much more second harmonics than the 2.7k//330nF.
Actually it takes about 3.5 times higher input to produce the same amount of clipping.
12AX7 response.jpg

12AX7 clipping.jpg
 
Might be able to identify the coupling path for the osc if you post some pictures of the wiring and pcb of this particular Marshall.
 
Back
Top