Weird push-pull performance

[Ewan Penkey]

Hi,

I wonder if anyone can help me diagnose a problem with my friend's 5F6A (ish) bassman?
It's an American reissue which has been converted to turret board at some point. It looks like a very good job, but there's a couple of problems with the output section.
The symptom is that there's very little power, and very early distortion.
I've tried adjusting the bias, but above a certain point, the output starts reducing rather than increasing, and I get the distorted waveform in the picture. I beginning to think the output transformer maybe damaged, but the measurements I've made on that check out ok so far.

Here's the measurements I've just made:

AC supply measured at fuseholder: 115Vac
HT at the plates: 460V
Screens: 326V
DC Bias at grids: -50V
Cathode current (measured across 1 ohm resistor to chassis) : 9.6mA
AC signal at grids: 6V peak to peak
output volts across 2 ohm dummy load 2V peak to peak.

At these settings I get a clean sinewave out, but if I increase the grid signal, or the bias voltage, I start to  get the distorted waveform pictured.
This is on a new pair of Tung-Sol 6l6GC STR's and a 5U4GB rectifier, but I get the same or similar results with other pairs of older 6l6's or 5881's

The output transformer is correct for a 2 ohm load, and the resistance across the windings seems ok. I've measured the turns ratio, and that checks out close to ideal for push-pull 6l6's so I'm assuming it's ok at present, although I've not meggered it.

Can anyone suggest what else I can measure or point me at why this amp is performing so badly?

Many thanks in advance

Ewan
 

[sluckey]

Has the amp ever performed correctly?

Screen voltage should be about the same as plate voltage. Suspects... wiring, choke, filter cap, maybe 470Ω screen resistors. Measure voltage on each side of choke. Should be about the same.

For your scope pic,,, where are the probes connected and what is the amplitude of the two signals?

[shooter]

AC signal at grids: 6V peak to peak

the distortion might be the PI getting cranky.  you should be able to get something clean at close to bias volts, so at 50vdc I would expect something like 30VAC out of the PI

[Ewan Penkey]

Hi, thanks for responding.
The amp hasn't been used for a couple of years as my friend lost faith in it, and didn't like the tone much.
He did gig with it though, and I don't see how he could have done that with such a low output, so i guess it must have been working.

I replaced the 470 ohm screen grid resistors for slightly higher wattage (3W) jobs. The voltage off the choke is the same as the the plates, so I'm dropping 134V (460V-326V) across the resistors.
Does that sound ok?

On the scope, the top trace is the grid of one of the 6L6's and is at 2V per div. so just under 10V
The bottom trace is across the dummy load and is 1V per div. so just under 3V

Ewan

[Ewan Penkey]

AC signal at grids: 6V peak to peak

the distortion might be the PI getting cranky.  you should be able to get something clean at close to bias volts, so at 50vdc I would expect something like 30VAC out of the PI

The signal at the grids of the 6L6's remains clean as a whistle though. The distortion is only in the output section.

Ewan

[Ewan Penkey]

so I'm dropping 134V (460V-326V) across the resistors.
Does that sound ok?

I just realised they would somehow be dissipating 38 Watts! - Clearly not!
I just measured them in circuit and they are 470K, but marked as 470 ohm!

Ok, let me correct that and report back 

Thanks guys!

Ewan

[sluckey]

That would do it! Let's see a pic of those resistors.

[Ewan Penkey]

Oh man - it's not even marked up wrong. I guess maybe they must have been picked wrong when I bought them.
I can't imagine why I would need a 3W 470K resistor.
They are marked up purple, yellow, black, orange. I guess I didn't notice it was a black line and not a brown one, and ignored the orange line.

A schoolboy error - I won't do that again! 

I had some 560's that I've just put in and now I'm getting sensible output and a decent range of bias on the pot, so happy days!

Cheers

Ewan

[thetragichero]

okay that makes a lot more sense... i plugged 470r and 134v into a power rating calculator and got 38w as well.... certainly didn't seem right to me