Bandmaster AB763 low output w/ weird distortion (scope shots attached)

[waspclothes]

I have a BF Bandmaster here that is only outputting about 8W and shows some really weird distortion. For the test, I'm using new tubes and the bias is set correctly. Here's a scope shot:



Blue - output into 8 ohm dummy load.
Cyan and yellow - outputs from P.I.

[mozz]

Voltages on all tubes ok? Ripple? How much are you driving the amp with?

[The Dude]

^^^^^^ Yes. That. Check your voltages and for ripple. It looks like B+ might be "giving up the ghost". My best guess would be failing filter caps.

[waspclothes]

Voltages on all tubes ok? Ripple? How much are you driving the amp with?

Hi Mozz, all voltages seem to check out. The ripple on the B+ is about 10Vpp and I'm using a .2VRMS sine wave as a test signal.


Here's the voltages:

[waspclothes]

^^^^^^ Yes. That. Check your voltages and for ripple. It looks like B+ might be "giving up the ghost". My best guess would be failing filter caps.

Hey Dude that's what I thought too, but the voltages seem to check out (see schematic I posted). The PT has been replaced by a Mercury Magnetics model (not by me) and the B+ is actually spot-on - first time I've ever seen that in a fender amp, I'm guessing that Mercury compensates for the higher line voltages we have now.

[Jazz P Bass]

How about trying it at 1K instead of 440Hz. with a 100mv input signal.

Also, it appears that the amp should be run into a 4 ohm load.

http://www.thevintagesound.com/ffg/s...b763_schem.gif

[Randall]

I have a question. Where does the 100V on the PI cathodes come from?

[Leo_Gnardo]

I have a question. Where does the 100V on the PI cathodes come from?

Some disconnection between PI cathodes and ground, my best guess. Either the ground wire from circus board to ground bus has come loose or that pesky feedback node resistor (100 ohms usually) has come loose, as they do.

[Justin Thomas]

I don't know how it works, but I know the LTPI just does that... No mistake or busted connection. There's a huge writeup on Randall Aiken's site about how that PI works, but it's over my meager abilities; I still don't grasp impedance and all that... But the schematic says it belongs that way!

Justin

[waspclothes]

Some disconnection between PI cathodes and ground, my best guess. Either the ground wire from circus board to ground bus has come loose or that pesky feedback node resistor (100 ohms usually) has come loose, as they do.

The 98.5V I've indicated on the schematic is taken from that middle node where the 4 resistors meet (1M, 1M, 470 and 22k) that should be 108.5V according to the schematic. My mistake, I put the text in a bad spot there.

[Tom Phillips]

waspclothes,
You did an excellent job of documenting the DC voltage readings. Marking up the schematic is the best method to convey the information.

I suggest that you do the same with AC voltages with the following setup:
Connect a 4 Ohm load to the output
Set all the volume & tone controls to max.
Make sure the tremolo is off
Inject a 10mV, 1 kHz signal into the #1 input of the normal channel.
Mark the resulting AC voltages on the schematic.
Also verify that the readings down line from the pre-amps are relatively the same for each channel.

With that data, we should be able to see where things are going wrong.
If all is good up to the 6L6 grid pins then we may need to substitute in another OT. BUT FIRST - measure the 6L6 screen DC Voltage right at pin 4 of the tube sockets.

[J M Fahey]

I see nothing wrong in the scope images, both PI plates are strong and clean, with jutthe top end *arely* reaching clipping, and output signal into speaker somewhat rough but nothing to write home about.
By the way, you are not using the proper load impedance either.

1) use 4 ohms load

2) rise volume until it *really* clips, signal becomes visibly squared, take mental note of the peak it reaches, both sides, then slowly lower signal until most of it goes below that peak level you just found, letting positive and negative peaks still smash against them.

A true squarewave would be 50% of the wave width (meaning 50% of the time) flat against the top limit; other 50% (width/time) against the bottom one.
Lower signal until, say, 10/15% peak widths are clamped against the limit.

Now take RMS measurement again and repeat the Math; it should have improved a lot, although I doubt it will reach 40 or 50W.

3) if you wish, play with bias voltage, vary it between, say, -40 and -50V , see if power output as defined above improves by 5 or 10W .
Couple watts gain is not worth it.
Designers in general go for maximum output power , and colder bias; musicians go for "tone" and have no clue about power (they never *measure* it, so .... ), couldnīt care less about dissipation, tube life, etc.

[J M Fahey]

I have a question. Where does the 100V on the PI cathodes come from?

From the 22k shared cathode resistor instead of usuall 820 ohms or so.

[waspclothes]

I see nothing wrong in the scope images, both PI plates are strong and clean, with jutthe top end *arely* reaching clipping, and output signal into speaker somewhat rough but nothing to write home about.
By the way, you are not using the proper load impedance either.

1) use 4 ohms load

2) rise volume until it *really* clips, signal becomes visibly squared, take mental note of the peak it reaches, both sides, then slowly lower signal until most of it goes below that peak level you just found, letting positive and negative peaks still smash against them.

A true squarewave would be 50% of the wave width (meaning 50% of the time) flat against the top limit; other 50% (width/time) against the bottom one.
Lower signal until, say, 10/15% peak widths are clamped against the limit.

Now take RMS measurement again and repeat the Math; it should have improved a lot, although I doubt it will reach 40 or 50W.

3) if you wish, play with bias voltage, vary it between, say, -40 and -50V , see if power output as defined above improves by 5 or 10W .
Couple watts gain is not worth it.
Designers in general go for maximum output power , and colder bias; musicians go for "tone" and have no clue about power (they never *measure* it, so .... ), couldnīt care less about dissipation, tube life, etc.

Thanks for the info Juan Manuel, I thought an 8 ohm load would be fine but I'll make a 4ohm one and re-test. When I set the maximum output volume (as shown in my scope picture), I turn it up as you describe to see the heaviest clipping, and lower it slowly until the sine wave isn't clipping at all. That's when I take my RMS measurement to calculate the output wattage, not at 10-15% clipping as you describe. I thought output power was measured with an undistorted sine wave?

[waspclothes]

waspclothes,
You did an excellent job of documenting the DC voltage readings. Marking up the schematic is the best method to convey the information.

I suggest that you do the same with AC voltages with the following setup:
Connect a 4 Ohm load to the output
Set all the volume & tone controls to max.
Make sure the tremolo is off
Inject a 10mV, 1 kHz signal into the #1 input of the normal channel.
Mark the resulting AC voltages on the schematic.
Also verify that the readings down line from the pre-amps are relatively the same for each channel.

With that data, we should be able to see where things are going wrong.
If all is good up to the 6L6 grid pins then we may need to substitute in another OT. BUT FIRST - measure the 6L6 screen DC Voltage right at pin 4 of the tube sockets.

Hi Tom, I will try these things. Is that input signal 10mV RMS or p-p? Voltages at pin 4 on the 6L6's are basically the bias voltage and symmetrical - I don't have measurements in front of me but I saw nothing wrong with them. I can't help thinking there's problem with the OT, although I don't have enough experience to have ever come across one that "works, but barely". The P.I. is outputting strong 100Vpp sine waves, and the output is weak with new tubes, yet symmetrical so I'm assuming both tubes are working and all voltages seem right. Just very hesitant to purchase a new OT still.