Debugging a 1972 Fender Bassman

[vigilante397]

Sorry I always post about amps here. I know this is a stompbox forum, but I like you guys better than all the amp forums, and frankly I've always been able to get the help I've needed from the brilliant people here.

So I'm working on a 1972 Silverface Bassman. It turns on, it makes noise, but the noise is awful. It breaks up way too soon; it is remarkably muddy, to the extent that unless the bass is turned all the way down and the bright switch on the amp is unusable; and it is also much quieter than a 50W amp should be.

I went through with a capacitance meter and checked all the caps, and pretty well everything except the ceramic disc caps were WAYYYYYY out of spec. So I replaced all the signal caps with Mallory 150s and the electrolytics with F&Ts, and there is a slight improvement, but not a big enough one to think that was the problem. It has a brand new set of Sovtek tubes in it, but just for the sake of science I swapped all of them out and there was still no difference. Running a multimeter around it seems like all the pots still read to spec so I don't expect they are the problem. Also, the internal bias control seems to do very little to the tone.

I'm pulling my hair out because I'm running out of ideas and things to check. I know the common overreaction especially among beginners is to assume that the output transformer is bad, but could this possibly be a problem with the transformer? That's the only thing I haven't really poked at because I generally assume the output transformer is never the problem

Thoughts on what to look at next? I didn't think to write down exact voltages because I remembered checking that the plates were getting proper B+ and the heaters were all getting ~6.3, but of course I can get exact readings tomorrow if they would be helpful.

Another confusing bit, the inside of the cabinet says it's an AB165, but comparing the layout it looks a bit different, even though the owner swears he bought it brand new and it has never been modded. It looks like it follows the AA371 layout (http://www.thevintagesound.com/ffg/schem/bassman_aa371_layout.gif)But for general reference, here is the AB165 schematic:

[Phoenix]

I'm sure you've probably already done this, but have you checked the obvious like shorted or open screen grid resistors? That's an exceptionally common failure and can cause a large bass emphasis which it sounds like this amp is suffering from.

You're right that it's almost never the output transformer, but if there was a problem with the screen grids that could have cause collateral damage to the output transformer, such as shorted turns, which can also cause bass emphasis and low volume. You know how to check it for shorted turns? I'm doubtful that this will be the problem, but it can be good to check, even if it's just for the sake of your sanity.
Disconnect all output transformer leads from circuit, and connect a low voltage transformer like a filament transformer, or a variac set to low voltage (be careful if using a variac in this manner, ideally use an isolation transformer between the wall and variac) and measure the voltage on the secondary, and compare it to the voltage on the primary. This will give you the turns ratio, and if you square it, it will give you the impedance ratio. The primary impedance should be around 4k. Anything in this ballpark will indicate the transformer is fine.

As to the noise issues, I'd be replacing the ceramic caps, as they can be noise prone especially at this age. I also make a habit of changing the location of any anode bypass caps from across the plate resistor, to instead from plate to ground/cathode (depending which one is more convenient to the layout). This maintains the same frequency response but doesn't couple any residual noise that is present on the B+ into the signal path.

Also check for any resistors that are looking discoloured, if you find any they could probably do with replacement. Old carbon comps can get progressively noisier with age, and even more so if they've been overloaded.

Maybe do go through and check all the voltages again (cathode voltages, plate voltages, B+ nodes, screen voltages, bias voltage), and write them down. Not so much for our benefit, but just doing it again can sometimes help you see things you missed before, especially if you write them down.

Good luck with this, let us know what you find.

[wavley]

Everything Phoenix said.

Also check the plate resistors on your phase inverter, I've recently seen two amps this past week where half the phase inverter had an open resistor and the signal was quiet and like a bad gatey fuzz.

[thermionix]

^^All of the above, plus:

The bias control in those is just a bias balance, not a control to set the overall bias.  What current is the bias set for now?  I would highly recommend rewiring the bias balance as a bias adjuster, and using matched output tube pairs.  It's extremely simple to convert in most cases.  30mA per tube should get you going.

[vigilante397]

Update: I had a bit more time to play with it today. First thing I checked was the screen grid resistors, neither was shorting and both were well within tolerance (marked as 470 ohm, measured 465 and 481). I was really hoping that was it so I could finally get rid of this thing with a simple fix

I swear I remembered checking the plate resistors before, but checked them again today (because why not?) and two of them were outside of tolerance (rated 100k, measure 76k and 80k). There were a couple others in the signal path that were outside of tolerance (220k measuring 124k, 470k measuring 168k), so I'll probably change those as well, but I'm still doubtful that any of these could really be the problem. But as far as I could see nothing is shorting or open

Unfortunately today was a relatively busy day with customers so I didn't have the time I wanted to get voltages, or to check the output transformer, or to look at the bias. I don't have the value caps I need (at least at that voltage rating) to replace the ceramics on hand, so that will have to wait as well. Tomorrow is my day off (it's finals week at uni so I'll be taking exams all day instead ) but Thursday I should be able to dig a little deeper.

Thanks for the suggestions, and I will continue to update this as I have more information If I remember on Thursday I'll even take some pictures to see if more sets of eyes spot something mine are missing

[thermionix]

220k measuring 124k, 470k measuring 168k

Those are REALLY off!  They were out of the circuit when you measured them, right?

A quick, half-assed way to check the OT is to measure DC resistance from each plate wire (blue and brown, pin 3) to the center tap (red).  VOLTAGE DRAINED FIRST!!!  You should see similar, but not the same, resistance on each half of the primary.  In a Bassman head, I would expect about 30 ohms on each side.  This method won't pick up all potential issues, but if you see 15 and 30 ohms, or 30 and 100 ohms, you definitely have a bad OT.  If you see 25 and 28 ohms, something like that, I would start looking at other possibilities before spending more time testing the OT.  Good thing is you can read those without unsoldering anything.

Also, the phrase "at uni" doesn't completely jive with "Idaho, USA."

[vigilante397]

Thanks, I'll definitely give that a try with the OT. And yes, that was with the resistors out of the circuit.

Also, the phrase "at uni" doesn't completely jive with "Idaho, USA."

I could have said "at university," or just "at school," but isn't using as few letters as possible the real American way? My last job was working as a translator for a few years, and the English speakers I dealt with were mostly English and Australian, so I talk a bit weird now On a side note, "cheers" is one of the most versatile words in the English language.

[petemoore]

  The chop stick method is a good thing to try, any totally dry wood stick can be used to gently push and lightly tap components, sometimes to quickly narrow down the search, big bad filter caps sometimes will crackle, 'bark' or behave quietly when whacked somewhat sharply, cold solders generally respond to movement of conductors.
   Ground the first grid, if the same noise continues it has to be getting injected to the ''right'' of the first stage input. Similar to using an audio probe, but since you already have and amp output, and source [the unwanted noise source], a safety danger [grid grounding] approach can be used to isolate the locations [before the output section] that aren't providing the unwanted noise source. It's safe to ground a grid, but dangerous high voltage is right there on the plate pins, it's better if you can use the wire that goes to the grid instead of probing at the socket pins.

[vigilante397]

I guess I should have worded my first post a little differently. When I said "it makes noise, but the noise is awful," I wasn't referring to the fact that the amp is noisy. Swapping the two-prong cord to add a proper ground did miracles, then swapping most of the caps helped as well, so there's almost no "noise," I was just referring to the fact that the overall tone of the amp is unpleasant.

[thermionix]

I think Pete was mentally mixing up two different threads.  Yours and a noisy pedal issue where I was recommending chopstick poking.

[PRR]

If you have drifty composition resistors, all sorts of strange things will happen, and go on happening, until you get them out. Sounds like all the "hot" (plate) resistors are dubious. "Cool" (grid, cathode, tone) resistors drift slower, but 1972 was a very long time ago.

You need to know the power tubes' idle currents. Exact balance is not critical (except for hum), but a dead (no-current) power tube makes an unhappy amplifier.