Amp Diagnosis Help

[_ej_]

I've got an 18watt kit I built that has an issue I can't quite figure out. There is a static noise that is audible under the notes. The noise more apparent as notes decay. It is present across multiple guitars and cabs. Its happens on both channels.

The kit was originally a Mojotone TMB kit but it had unusable amounts of gain so I modified the TMB channel with a cathode follower to reduce the gain. I'm got a modified schematic below (modifications in blue). I've confirmed with a scope that it is indeed unity gain across that stage.


I've gone through the wiring a few times and I can't find any issues. I've checked the voltages and there are a few that are a low in the preamp but I lifted the components going to those nodes and they all check out. I also put a 100hz and 1000hz sin wav into the input and checked the different gain stages with a scope. The waves looked clean throughout the whole preamp.

I've rearranged the preamp tubes as well as swapped a known good tube into all positions. It did not affect the noise.

I also tried new power tubes and it didn't help.

The output after the power section looks pretty distorted but its a Marshall so that's not surprising. Yellow is the input signal, pink is the output. I have no idea what this should look like but the core tone of the amp is good, its just this static noise that's an issue.


Since the issue is present in both channels I have to assume its a problem with the PI or the power section. I'm going to go through everything again to make sure I didn't miss something. I'm also considering ordering a new rectifier tube. Any help you guys could offer would be appreciated. I'm stumped and just want to enjoy the amp.

Here are some audio clips. I'm using a hotplate but it happens without it.
https://youtube.com/shorts/iij_tn--AI8?feature=share

This is with the gain at noon then I turn it down to 25%. Doesn't affect the noise. I also use the bridge and neck pickup.
https://youtube.com/shorts/IAg7YENqHE4?feature=share

This is an older clip with an entirely different guitar.
https://youtube.com/shorts/44pZbBPrDTY?feature=share

[martin manning]

I think you are right aboutw the issue being in the PI of power amp. The smears on the pink trace are high frequency oscillations, which can sound like static when they come and go. Could be caused by a lead dress problem.

[Stevem]

Could you provide numerous really clear gut shots of the amp because it does sound like a layout issue.

Just of kicks try removing the cathode bypass cap from the output section, does that help any?

How is your ground layout ?

You could be coupling thru the filters.

The grounds from each gain stage should go back to the filters those stages are powered from.
All the those filter grounds should go back really close to where the ground for the power transformer and the D.C gets its ground, so none of the filter grounds should be daisy chained .

Also to back track here the the D.C. Ground off of your diodes should go first right to the negative of the first power supply node that feeds the OT, then a wire from there should go to ground by the PT.

[_ej_]

I think you are right about the issue being in the PI of power amp. The smears on the pink trace are high frequency oscillations, which can sound like static when they come and go. Could be caused by a lead dress problem.

I will say the signal coming off both plates of the PI was clean and free of the high frequency noise. Not sure if that points towards the power section a little more or not.

Does the general shape of the output waveform look right?

I'll post some gut shots later today

[TUBEDUDE]

I'd touch up all the joints in the output section especially the grounds. It way be lead dress also, move some wires if possible. Just for fun, I'd lift one end of that protection zener also.

[_ej_]

Here are some gut shots. I did change the grounding a bit from the Mojotone layout. I have all the preamp grounds going to a star near the channel input jacks. How do you go about check of a ground is the issue?

Also ignore the "passive loop", I installed it to help with debugging. It's not a permanent addition.

If y'all want to see anything closer just let me know.

[lonote]

It may be lead dress also, move some wires if possible.

This would be where I would probably start.

Looking at the pics, there seems to a good bit of extra wire length with many runs & lots of random layers of wires crossing or even weaving to their destinations. Trying to keep things run efficiently, away from one another & sorted by signal/high voltage has helped me a great deal in the past.

I have had weird, seemingly random issues or interference that came down to lead dress &/or proximity to higher voltage leads.

[Helmholtz]

Just to make sure, did you add global NFB?

[_ej_]

Just to make sure, did you add global NFB?

I don't think so, what is that and what would it accomplish?

[Helmholtz]

I don't think so, what is that and what would it accomplish?
[/quote]

I meant NFB from the output to the PI tail.
Just confirm that the power amp circuit conforms to the Mojo schematic (except for the zener).

[_ej_]

Yeah the only mods are in blue. The power section matches the schematic beyond the addition of the zener

[_ej_]

I made a proper recording of the issue. This was into the clean channel with the MV and tone at noon. I am using an attenuator but I have verified that it doesn't cause this issue with other amps.

https://soundcloud.com/ejendres/marshall-issue

[martin manning]

Did you try lifting one end of the Zener at the cathode? You could also try larger grid stoppers. Marshal's 1974x used 8k2, and they are best located on the sockets.

[Phil_S]

This likely does not directly address the problem, however, the comments on poor lead dress generally relate to various noise issues. For example, the Switchcraft input jacks may or may not be wired correctly -- it is hard to see. The wires fly all over, with long lengths of wire. I'm posting two items that show how this is often done cleanly and neatly -- very typical amp construction. The jack wiring in your pictures is just the tip of the iceberg. There is a considerable amount of cleanup that could be done.

For a new builder, it can be difficult to make the leap between schematic and layout. Schematics are representations of the various component connections. Schematics do not tell you very much about layout. A good exercise might be to find a Fender schematic, layout, and photos of a particular amp. Often the schematic and layout can be found in a single PDF. Focus on how to get from the schematic to layout.

The jack wiring diagrams are a good example of the disconnect between schematic and conventional layout. It's easy to think that long black wire from the sleeve is the appropriate lead dress when you only have the schematic. Generally speaking, it's just not done that way. I'll note that some layouts place the 68K resistors on a tag board and others put them directly on the grid pins. These are all OK but make the wire lengths short.

You've already come a long way. Yet, there's room for improvement. We all start somewhere. It's a journey.

[_ej_]

This likely does not directly address the problem, however, the comments on poor lead dress generally relate to various noise issues. For example, the Switchcraft input jacks may or may not be wired correctly -- it is hard to see. The wires fly all over, with long lengths of wire. I'm posting two items that show how this is often done cleanly and neatly -- very typical amp construction. The jack wiring in your pictures is just the tip of the iceberg. There is a considerable amount of cleanup that could be done.

For a new builder, it can be difficult to make the leap between schematic and layout. Schematics are representations of the various component connections. Schematics do not tell you very much about layout. A good exercise might be to find a Fender schematic, layout, and photos of a particular amp. Often the schematic and layout can be found in a single PDF. Focus on how to get from the schematic to layout.

The jack wiring diagrams are a good example of the disconnect between schematic and conventional layout. It's easy to think that long black wire from the sleeve is the appropriate lead dress when you only have the schematic. Generally speaking, it's just not done that way. I'll note that some layouts place the 68K resistors on a tag board and others put them directly on the grid pins. These are all OK but make the wire lengths short.

You've already come a long way. Yet, there's room for improvement. We all start somewhere. It's a journey.

Hey Phil, thanks. My plan is to try and clean up the build quite a bit when it's actually working. I've already been looking at high end builds of similar amps to get some inspiration.

I am going to move some of the leads around the pi and try and get some distance between high voltage and signal lines. I still feel like this isn't a lead dress issue, it's not a constant noise/hum, it sounds like something is being overloaded.