Safe to put V1a coupling capacitor on a switch for variety?

[Enzo]

Hum comes from many sources, it is not generic. You elevation will only reduce hum if that hum was coming from heater/cathode leakage.

[Mr. Bill]

This is neat and I think worth mentioning. With the volume halfway up, I can crank the gain to a sweet spot where hum is quite diminished. I thought it was neat.

Now I have a stupid question. I was probing the AC coming off the coupling caps of each stage with the gain off and the volume way up. When I probe V1b (remember I'm calling the added stage V0) the amp goes haywire with super loud hum. Is this normal for some reason, because no other spot in the circuit does this.

[Justin Thomas]

The fact that the hum goes in and out with various changes and adjustments of gain and volume controls means there is more than one source of hum, they are out of phase with one another, and are phase-cancelling as various levels of the hum are reached or passed. If it steadily increased as the gain or volume was advanced, it would mean there is only one source of hum. As far as how to find them without a scope, I know nothing more than trial and error... :P

Justin

[Mr. Bill]

That's what I figured. It's a new behavior with this particular preamp tube in V2. And you're right, I'll be getting my own scope in the coming weeks. It's difficult carting things to school and back when the lab is open.

The fact that probing between V1b and V2a producing an insane amount of hum when probing any other part of the path ever so slightly reduces the noise seems like a major clue to me, but I could be missing an obvious reason why that would be so. I'm going to retrace my grounding and layouts today and check all my joints just to make sure I didn't miss something silly. It's happ

[Enzo]

If you have an amp that hums most at mid-setting on volume control, start at latest point, say the master, that does this. It it is the only control doing it, fine, if the gain also does it, start at the later one anyway. It hums least at mid, meaning it hums more at zero. SO set it to zero and fix THAT hum. After that, THEN you can turn up that control and deal with hum in the earlier circuits.

[Mr. Bill]

Thanks Enzo. Maybe I wasn't clear on that, I apologize. With gain off, hum steadily increases as volume is raised. No hum at zero volume. However, with volume set to around noon, increasing the gain decreases the hum until about 9 o'clock, in which case it begins to increase again.

[Enzo]

Clear enough, I was speaking in general to expand on Justin's thought.

Wherever in the control settings it nulls, it still points to two separate sources of hum that are out of phase. One needs to cure one at a time.

[Mr. Bill]

I've been reading some of these good grounding threads that Chuck mentioned, and......well........I've always followed the old Marshall style grounding schemes which couldn't be further from what you guys recommend for quiet operation. On a stock 2203 layout, I count 13 or 14 possible paths to ground not including safety earth. Looks like this could be done much cleaner to begin with. So.... after studying a bit more, I may be ripping the whole thing out. On one of those threads I dug up, a guy said he managed to get the output voltage on an amp he was working on with gain off / master wide open....down from 400mV to 4mV just by changing the grounding scheme. That's NUTS! Learning is fun.

[Chuck H]

Rip it all out if you REALLY want to idealize and have the time and $$$ for wasted components, etc. The thing we didn't know from the start was just exactly what this amp was, other than a sort of 800 type build. Which can be done with typical Marshall grounding and a "normal" amount of noise. The 800 models that stack the channels are a little noisier. Then you kept adding switch options for circuit parameters. Now have stacked an additional gain stage. Some switches are possibly grounded for convenience of location. So, yeah, you basically complicated the potential for ground loops, less ideal ground location and amplified the problem by a gain of, what, 40 to 60? Whatever your added stages gain is. In fact, had I known you were building a four stage cascade I might have made some different suggestions earlier. And we can't forget that you had an orientation problem with your OT and it's lams are parallel to the PT. But anyway...

In the end this could be a good thing. If you do rebuild it you'll have the option to omit some of the switches you probably find you aren't going to use. I imagine you must have a favorite position that, in practical use, you would never change for your variable preamp cathode circuits. I mean. really? You have some hundred possible combinations? You're probably not really going to use them all. In fact I'll wager there are two of those preamp cathodes you already have a favorite position for and only switch them for the novelty.?. Of course, the holes are there now so you'll want to use them because a plug is aesthetically displeasing on a build you went to some cosmetic trouble with. It's a very ambitious build for anyone. It would be for me and I've designed several amps and mods. FWIW you've already done better, faster than I expected.

You can still change a lot about the grounding, likely enough, without ripping everything out. You could probably even change to DC filaments without undoing most of the filament wiring. I'd probably start with these two things before a tear down.

[Enzo]

I am not going to go find the 2203 drawings, but on many models, GROUND, as in chassis, is separate from circuit common. Circuit common, what we usually call ground, is separated from chassis by a small grounding network, typically a pair of diodes, a cap and a 10 ohm resistor.

The number of paths to ground isn't the issue, the issue is sharing ground paths where it doesn't belong. MArshall put a lot more care into that than it may seem.

[Mr. Bill]

Can someone confirm something for me. I redid that lamp battery test, which ends up slightly under powering the heaters by a volt or so. Should you be able to play with the heaters on a 6 volt battery? Would the tone be thin with reduced output if the heaters were only showing 5 volts or 4.8, etc?

Edit: I guess I just answered my own question by using a different battery that was fully charged. I think I'm doing this test more correctly now, which is leaving the guitar unplugged to rule out other interference outside the amp. Gain off, volume max, and measuring actual voltage off of the speaker output.

Leaving filaments AC I'm getting 185 mV off the speaker jack. Still don't have it on a scope, but the fluke set to measure frequency is showing it as 60 Hz predominant. My best guess is since the cab can't reproduce that very well, the mic is showing the strong harmonic.

Switching out to the lantern battery (hooked up right this time) is giving me almost dead silence in the room. Voltage on speaker jack is 11 mV. SPL at 1 inch from the speaker is 48 dBA.

[Chuck H]

Well then... I hate to say it, but I may have led you wrong in suggesting that you could follow the standard Marshall filament lead dress. With the extra gain in your circuit and the extra lead wiring it could be that a tightly twisted pair right to the pins is what you need. Also, make certain that all your grid leads meet at right angles to any filament leads and are elevated high above those leads and even the chassis they sit near. This has worked for other builders using AC filaments in high gain amps before DC filaments were all the rage. A PITA I know, but not undo-able.

But we also can't neglect that there is probably another hum issue. as evidenced by your previous observations with phase cancellation at specific settings. Still... Solve for one first as identified, then tackle the other.

Going to DC filaments could still be an option and you wouldn't need to redo your filament leads. This would tax your winding even a bit more than it is now, but I think it would be ok. I think the winding could handle it and the tax on the voltage may not be enough to drop voltage below spec. The only issue might be heat. You could build the DC circuit on a separate board and just zip tie it in somewhere if it works out, or skip it if it doesn't work out.

[Enzo]

So steal the circuit MArshall did use, where they ran the first couple tubes on DC. REctifier, couple caps.

[Mr. Bill]

I have a theory that the noise is there despite the twisting because I've located all of those cathode rotary switches on the rear panels right above the twisted heater wires.

At any rate, for the sake of learning, I'd like to reattempt elevation to see if that helps, and possible the hum pot idea. Then whether I use it or not would like to try and build a DC supply.

Can you point me in a direction of how to actually do the elevation correctly? Ditto for DC, I have found a bit of conflicting information as to the necessity of regulation, etc.

[Chuck H]

The elevation circuit is easy. See schematic below. The first figure would simply be connected to the screen supply. The second figure incorporates the screen supply balance resistor on the totem so there is no additional load on the HV rail and uses one less resistor (total count for screen circuit and elevation) R2 and R3 should have a series value close to R1 and create a roughly 10:1 voltage divider for the elevation circuit. I don't think either is better, but I like the eloquence of figure 2.