Marshall JTM/1962 , of 1998, amp clipping

[N_Cook]

Output ok up to 1 watt out then output clips both sides of sine. All valves
test fine for gain and internal leakage. Signal is fine up to the wiper of
the treble, ie before the splitter/driver, that signal stays sine well into
clipping of the output. Changing the splitter is the same. Running at
clipping and dropping the HT via variac makes no difference , changing
output bias makes no difference to the clipping. The splitter valve cathodes
measure about 34V and the grids about 23V at point of clipping. All the
surrounding Rs measure ok. Output Tx primary measures 139R red to white.
Capacitor problem? amp was subjected to a knock that dropped the output of
the rectifier valve , since changed, a session before this problem emerged.

 

[N_Cook]

v3p1 anode is 220V and v3p6 anode is 168V so leaky .1uF letting through some
of the neg bias from the output bottles , I assume

 

[Phil Allison]

"Nutcase Kook"

v3p1 anode is 220V and v3p6 anode is 168V so leaky .1uF letting through
some
of the neg bias from the output bottles , I assume

** ROTFL...

 

[Phil Allison]

"Nutcase Kook"

Output ok up to 1 watt out then output clips both sides of sine. All
valves
test fine for gain and internal leakage. Signal is fine up to the wiper of
the treble, ie before the splitter/driver, that signal stays sine well
into
clipping of the output. Changing the splitter is the same. Running at
clipping and dropping the HT via variac makes no difference , changing
output bias makes no difference to the clipping. The splitter valve
cathodes
measure about 34V and the grids about 23V at point of clipping. All the
surrounding Rs measure ok. Output Tx primary measures 139R red to white.

** The output tranny may have shorted turns.

I test them by driving the speaker winding with an audio generator set at
400Hz and checking the AC at the anode of each EL34 / 5881 with a scope.

The amp is powered off for this, of course.


..... Phil

 

[N_Cook]

I test them by driving the speaker winding with an audio generator set at
400Hz and checking the AC at the anode of each EL34 / 5881 with a scope.

The amp is powered off for this, of course.


.... Phil

I forgot to say swapping output bottles did not swap the problem side.
Trying a 600 ohm sig gen backwards into "8 ohm" so about 50mV pk-pk there,
0.36V on the (from low V3 anodes) problem side "white anode" and 0.5V on the
red other . I assume they should be much closer than that. I suppose the
conclusive test is swapping red and white wires and see of the low splitter
V3 anode voltage swaps halves.

 

[Phil Allison]

"Nutcase Kook"

Trying a 600 ohm sig gen backwards into "8 ohm" so about 50mV pk-pk there,
0.36V on the (from low V3 anodes) problem side "white anode" and 0.5V on
the
red other . I assume they should be much closer than that.

** The OT is fucked.




.... Phil

 

[N_Cook]

"Nutcase Kook"



** The OT is fucked.




... Phil

Looks more and more like that
I forgot about the feedback and simply swapping white and red creates an
ultrasonic oscillator via the phase change. Changed that .1uF cap and same
clipping problem.
So nowhere else to go , I may try disconnecting the feedback , take some
readings with red and white in correct positions and then swap white and red
and see if the problem side changes.
Any other ideas to try with the Tx in the amp and wiring in place?

 

[N_Cook]

Can't really follow that.

If you pull the phase splitter, are its grids getting an undistorted signal
measured at the empty socket?

Similarly, if you pull the output valves and put back the phase splitter,
are they getting a good grid signal?



Gareth.

I may as well try removing the output valves and trying on 80 percent mains
or so ,and check their cathode signals, but looks more and more like a turn
or few , shorted turns in the Tx

 

[Phil Allison]

"Nutcase Kook"

Any other ideas to try with the Tx in the amp and wiring in place?

** In my workshop, I constantly monitor the AC current draw of any piece of
gear on the test bench.

Faulty OTs are immediately obvious as the power coming out of the amp, when
under load and drive, is SFA - while the AC current draw is full on.

Also, the resistances of each half of the OT primary differ by more than
10%.



..... Phil

 

[N_Cook]

"Nutcase Kook"


** In my workshop, I constantly monitor the AC current draw of any piece of
gear on the test bench.

Faulty OTs are immediately obvious as the power coming out of the amp, when
under load and drive, is SFA - while the AC current draw is full on.

Also, the resistances of each half of the OT primary differ by more than
10%.



.... Phil

I forget whether measured cold or warm but 61.3R and 66.4R, which seems
normal enough, will try again cold and also with a 1KHz RLC for "R" and L
of both

 

[N_Cook]

a corroded break rather than shorted turn/s? now taking Tx readings
DVM White 61.8R , Red 75.3R was 61.3/66.4
RLC(L) 20.9mH, 34.0mH
RLC(R) 215 imp, 338 imp
Things certainly not right but what is wrong inside? tugging the tails makes
no difference

 

[Phil Allison]

"Nutcase Kook"

a corroded break rather than shorted turn/s? now taking Tx readings
DVM White 61.8R , Red 75.3R was 61.3/66.4
RLC(L) 20.9mH, 34.0mH
RLC(R) 215 imp, 338 imp
Things certainly not right but what is wrong inside?

** Valve OTs do not get hot enough to melt the enamel insulation on the
wires, as often happens with power trannys. What brings about their demise
is the very high voltages that exist during operation into a loudspeaker,
particularly when driven into clipping.

Voltage peaks can then occur on the anodes of the output tubes that range up
to thousands of volts - resulting in insulation breakdown and arcing
between adjacent layers of the primary winding.

Sometimes an OT will test OK with a multimeter or at low power only to arc
internally as soon as full power operating conditions are applied. This is
due to carbonised material in the area of the previous arcing reigniting in
a new arc.

Marshall amps using EL43s are the prime example of this kind of OT failure.

Some amps, notably Music Man and some Boogie models, have high voltage
diodes fitted from output valve anodes to ground to prevent the sort of high
voltages that damage OTs.


..... Phil

 

[[email protected]]

You either got raised resistance of the cathode resistor(s) of the splitter, low voltage to an element of the outputs or shorted something on the output.

Do you know how to use a scope in the DC mode ? Whatever the cathode voltage is on the outputs, remember that and set the trace to a line somewhere, over it. If the grid drive gets up near that level then the problem is after.. Has to be in the load or the bias of the output stage then.

If however you see that the output grid drive clips before getting near thecathodes' DC level (of the outputs of course) then look to the splitter stage. The way to tell if the cathode resistor value(s) has gone high is thatthe plate voltage will be high. Usually.

J

 

[Phil Allison]

<[email protected]>

You either got raised resistance of the cathode resistor(s) of the splitter,
low voltage to an element of the outputs or shorted something on the output.

Do you know how to use a scope in the DC mode ? Whatever the cathode voltage
is on the outputs, remember that and set the trace to a line somewhere, over
it. If the grid drive gets up near that level then the problem is after. Has
to be in the load or the bias of the output stage then.

If however you see that the output grid drive clips before getting near the
cathodes' DC level (of the outputs of course) then look to the splitter
stage. The way to tell if the cathode resistor value(s) has gone high is
that the plate voltage will be high. Usually.


** OTOH - the amp might just be hexed.

Don't you think ?



..... Phil

 

[N_Cook]

You either got raised resistance of the cathode resistor(s) of the splitter,
low voltage to an element of the outputs or shorted something on the output.

Do you know how to use a scope in the DC mode ? Whatever the cathode voltage
is on the outputs, remember that and set the trace to a line somewhere, over
it. If the grid drive gets up near that level then the problem is after. Has
to be in the load or the bias of the output stage then.

If however you see that the output grid drive clips before getting near the
cathodes' DC level (of the outputs of course) then look to the splitter
stage. The way to tell if the cathode resistor value(s) has gone high is
that the plate voltage will be high. Usually.

J

+++++

The owner reckons he knows someone across the country with a spare Tx and
ability to fit it.

 

[[email protected]]

So what now, he ships the thing there ? I think it would be better to ship the Tx than the whole amp. Of course some people are nuts.

Really though, if this thing runs class AB, if the Tx is shorted (turns) causing this, I would imagine the output tubes would redplate if you keep on pushing it no ?

Maybe you should plug the primary of that transformer into the mains and see if it smokes.

HEY, that is not stupid as it may sound really, but I would definitely remove it from the cabinet first. Think of it, you put 220/110 at 50/60Hx to itat the primary ? That shouldn't bother it at all. If it does, it's bad anyway. Difference is, this is a source with some balls (current) behind it. The impedance SHOULD keep it from smoking.

It's called follaw the smoke, I've done it meself sometimes.

J