New Bassman+ build with issues

[tubeswell]

Thanks loudthud. I'll look at the gridstoppers but I want to get to the source of this hiss. FWIW I took the GFB resistor out and the problem persists. there is occasional static intermingled with the hissing and when the static is intense there are blue flashes on the plates, so it seems that maybe something is up with the bias supply. I wonder if I need a 5W limiter resistor in front of the bias diode, or whether that resistor being there is only really necessary if the bias is tapped from the HT (in my case the bias is supplied from its own 50VAC winding). (BTW I have used this electrical switchboard before and never had any problems with it, so I don't think it can be that).

[tubeswell]

Some progress

Okay one thing at a time

I got rid of the power tube hiss by redoing the layout at the power supply end of the main board and moving some of the eyelets further apart (the screen eyelet and the bias supply eyelet were really close together, now they are at least 2 " apart. Now with only the output tubes plugged in, there is no more hiss (- yay). I am not sure why this is so but maybe it is an issue with the board (as loudthud alluded to - but I am not sure why this would be so, because I have never had issues with this type of board before). But the output stage hiss is gone so anyway...

Movin' right along, there is a crackling/random static noise when the PI is plugged in (in addition to the output tubes) with all the controls cut. This happens no matter what tube I use in the PI slot. I have changed the PI plate resistors to no avail. Hmmm....

[MWJB]

Might be an photographic illusion but there doesn't appear to be any solder in the PI eyelet, where the grid, cathode & tail resistors meet?

[tubeswell]

Hi MWJB - there was solder there but just not in the whole eyelet. Anyway I reflowed all the eyelets with my hot iron to clean the crap out of them, and there was a bit of gunk in one of two. But that didn't fix the problem.

Then I thought that one of the plate pins on the PI socket looked a bit loose, so I re-tensioned them all, but that still didn't fix the problem.

Then I wondered if it was the little 47pF cap between the PI plates, so I temporarily lifted it, but that didn't fix the problem either.

As I have two 'power rail' lines on the back of the main board (one for the pre-amp supply from the pre-amp supply node, and one for the RT and LFO supply from the screen node), I was wondering whether they could be interacting in some way to cause random voltage fluctuations on the board?

Maybe removing one of the supply rails from the back of the board would help?

I have used this type of board in three amps already without any issues, but this is the highest-voltage amp yet, and the board this time is from a different 'batch'. Is it possible that this board could be inferior? Hmmm...

[MWJB]

Just a thought, rather than have the voltage supply to RT come from the screen node on the board, why not have it come direct throught the chassis from the cap pan, like on a BF amp? Would this make the routing more direct?

Do you pick up any voltage on the board material itself? Isn't it a non-conductive material?

Checked for DC leakage via tone caps? lift the wires going to the pots from the top of the caps, so the caps are only connected at slope resistor end and check for DC, more than 0.25vdc change the cap.

[tubeswell]

Thanks MWJB - I tested all the coupling caps for DC already. After trial and error elimination I suspect a leaky board as of yesterday (although loudthud did point this out earlier - and I have used this type of board before without issues - however maybe its from a different batch, or maybe the voltages in this amp are higher than the others?? - It might be handy thing from an authentic fender design point of view if this type of board is slightly leaky, but before I get into the merits or otherwise of that I better get this build working properly first). Anyhow I am planning to take the screen node supply off the main board. (I note that Fender amps have these higher voltage wires like this off the main board). We will see.

[loudthud]

The problem with Fender boards is usually attributed to moisture absorbsion and can be cured with a hair dryer or baking the chassis at a moderate temperature. (30 minutes at 50 degrees C or 120 degrees F.) You can usually find leakage on the board that acts like leaky caps but you lift one end of the cap and the leakage is still there. Or you disconnect everything from an eyelet but can still measure of voltage on it.

One thing that affects the PCBs in my day job is flux. Any water soluable flux is acidic and will cause leakage. Any of the so called 'No Clean' fluxes cause leakage, we have to scrub boards with a tooth brush and board cleaner, then bake the boards. Rosin core flux (type RA or RMA) has the lowest leakage and you can clean it off without much trouble. The boards I work on measure leakage so they are more sensitive than guitar amps.

One of the 'Old guys' that tought me how to track down leakage used an old Triplett VOM. The X100K resistance range used a 30V battery. Any deflection of the needle on that range indicated leakage that had to be eliminated. I have a DVM with a 2000 Megohm range that works almost as good.

If you can't find leakage with a DVM another thing to try is touching a .1uF cap from the signal path to ground. Move up and down the signal path and try find where the noise originates.

[tubeswell]

Thanks loudthud

I re-did the layout today by removing the filter board and mounting the whole power supply in the doghouse on tag boards. I also took the screen/RT/LFO supply off the main board. Now the power amp (without any pre-amp tubes or the PI) is beautifully quiet, but when I add only the PI plugged in, the amp still roars like a distant thunder storm (even tho' all the controls are cut).

Having read your post just now, I wonder if it is some leaky flux somewhere?

What's that voltmeter trick? I have a DMM with a voltmeter scale that goes up to 200M (but it is powered by a 9V dry cell battery). Will that work? And can you please explain it in a bit more detail?

Cheers

[loudthud]

Using an ohm meter to find leakage usually involves removing all connections to an eyelet and them ohming it to ground or any near by eyelet. Not a great way to troubleshoot but helps confirm things once you find the problem area. I would look for voltage where the voltage should be zero. Caps will leak some, but say at the tone stack, the DC voltage when the amp is powered up should be tens of millivolts, not volts. If you think you have found a point with a suspicious voltage, reverse the DVM leads and see of the voltage reverses on the meter. This lets you know that you are not chasing a ghost.

[tubeswell]

Thanks for the tip about reversing the meter leads loudthud. I have already checked the coupling caps and tone stack caps for DC leakage and couldn't find anything significant (there was a piddly few mV on one of the PI to PA coupling caps). I think I will put the PI circuit on tag terminals elevated above the main board. I am guessing it may be a flaw in this particular piece of board (but I can't be bothered at this stage making a completely new board - or can I? - time will tell).

(I could say that troubleshooting is a fun part of building an amp but I don't want to tempt fate, so I will think about it instead ).

[loudthud]

Before I built a completely new board, I'd try substituting every component one at a time to verify that it's not a component problem. There is no direct evidence that the board is bad. In fact you looked for evidence and didn't find any. You have to eliminate all the other possible causes. One thing to try is just touch the DVM probe to points on the board between eyelets and components. If there is leakage on the board you might be able to detect small voltages where there should be none. Here's where reversing the DVM leads confirms that voltage is actually there.

Since the suspect area is the PI, I would disconnect the input or solder a wire to ground. Noise still there? Check every component by substitution. Just remember that the board is a component.

[tubeswell]

I haven't taken the pre-amp DC rail off the main board yet (I have only taken the Screen/RT/LFO power rail off), so it could still be that. I have treid subbing the PI plate resistors and the PA grid load resistors, and no difference. I'll try subbing the PI-PA coupling caps next, then lifting the whole PI circuit off the board.

[tubeswell]

I was looking around the PI tail today and I noticed that with all the pre-amp tubes in, when I clip the Voltmeter on to the grid-side of the treble-wiper to PI-grid-input (22nF) coupling cap the staticky noise disappears. So it can't be happening after the PI grid can it? It must be something to do with the signal before the grid. The treble pot wiper has shielded cable to that coupling cap. I wonder if the shield has an intermittent connection to the signal path? I'll check it out tomorrow.

Also the voltages on the PI are:

18V at pin 2, 19-20V at pin 7, 59V at the cathode and 58V at the cathode resistor/tail node. The plate voltages are 243 (on the 82k side) and 230 (on the 100k side). These are probably a bit higher than what was originally envisaged. My copy of Merlin's book hasn't arrived here yet, so I'm not sure if its biased right. The grid load resistors are 1M each, the tail resistor is 10k, and the cathode resistor is 470R. I had CCs in the tail, and wondering whether they were duds, I swapped them for CFs (after loudthud's suggestion to try subbing in parts), but no difference. Not sure if these voltages are too far out of ballpark or not.

[MWJB]

"My copy of Merlin's book hasn't arrived here yet, so I'm not sure if its biased right." 1v/470 = 2mA, or 1mA per triode...seems "right" enough for a 12AX7 to me. Original schem suggests 1.5mA per triode.

Voltage at PI tail seems high though, nearly twice what the schem suggests?

[tubeswell]

Voltage at PI tail seems high though, nearly twice what the schem suggests?

Yep well I triple-checked the tail resistor and presence pot and they are the correct values. I don't get much difference with different tubes either. What might cause that? (I don't have the 27k GFB resistor connected at the moment)