Farfisa Tube Amp High Pitch Oscillation RVRB Problem

[Rob Strand]

As far as the original problem goes it's still not clear what was causing the arcing.    It could be DC issues, heater breakdown, or something promoted by oscillation.

I agree with PRR.  When the failed cap is open there is more feedback and that can definitely increase the risk of oscillations.

It's not clear if the cap fault caused the arcing.

If you changed the output transformer then that can also cause oscillations to occur.    The specifics of the transformer does have an impact on amplifier stability.

If you look at this schematic you will see a series resistor + cap at the first stage and a cap across the output transformer.   Instead of a cap across the transformer as resistor in series with a cap is usually better.  You might be able to use the input network alone or the output network alone.
https://i.imgur.com/i9nGemc.jpg

The thing is determining these components isn't a simple task.  It's more of a redesign.  And you would only expect that such a redesign is required if the transformer change has affected the amplifier stability.   We  don't know if that's true in your case.  Also we don't know if it caused arcing.

As far as getting the unit up and running.  Measure the DC voltages on all the tubes.    You should also check the DC resistances of the transformer - it could have shorted, not likely but possible.    The cathode voltage on the output tube will tell you if that tube passing current, you could check the current looks OK as well.

Who knows, maybe you new transformer had a fault from day one and was breaking down and now it has failed.

[sarakisof]

PRR i see, cathode bypass cap can cause that.

Arcing comes alongside with pop for a second usually during power up warming, but sometimes also later.

ECL86 voltages:
Anode pin6 320V. (exactly as in scheme)
Anode pin9 133V.

By signal tracing (1khz to input then diy tube tracing cable to xtrnal cheap amp) i got signal through hole path until rvrb trafo out (tank in). Then i got pop/arcing again and immediately switched off. It's scary 🤣
So i guess ecl86 is fine.
I don't get splash sound by tapping springs. Tank is ok, so no recovery stage.

So my next step is to check recovery stage. Could bad AC151 germanium cause all this?

Edit: I know see Rob's reply. I'm reading it now.

[sarakisof]

Tube voltages are fine, didn't measured everything but when i have done back then were all fine. Remeasured B+ and anodes at most tubes again, they follow schemo's volts.

Trafo's res: 580ohms input, 3ohms output(tank in). It's ok, it was the first thing i checked when i got pop arcing.

EDIT: Trafo seems fine but i could order a new Hammond 1750AX  just to be sure.
https://www.tube-town.de/ttforum/index.php/topic,24358.msg247927.html#msg247927

The cathode voltage on the output tube will tell you if that tube passing current, you could check the current looks OK as well.

You mean ECL86 cathode voltage? I have a UNI-T 139C multimeter with a μA/mA scale. How do i measure cathode current there and what I'm looking for?

EDIT: About EL34s dissipation they are running at about 72-80% for V7 and 80-88% for V8.
https://www.diystompboxes.com/smfforum/index.php?topic=126741.msg1213275#msg1213275

[Rob Strand]

You mean ECL86 cathode voltage? I have a UNI-T 139C multimeter with a μA/mA scale. How do i measure cathode current there and what I'm looking for?

Measuring current like that is very inconvenient because the meter needs to go in series with the tube.  What I was suggesting is by measuring the cathode voltage you can imply current is flowing simple because the cathode voltage is non-zero.   You can also calculate the current as I = Vcathode/Rcathode; and for accuracy you could measure value of Rcathode.

Tube voltages are fine, didn't measured everything but when i have done back then were all fine. Remeasured B+ and anodes at most tubes again, they follow schemo's volts.
Trafo's res: 580ohms input, 3ohms output(tank in). It's ok, it was the first thing i checked when i got pop arcing.
Q

It's looking like the tube part of the circuit is OK.  (except for the arcing)

By signal tracing (1khz to input then diy tube tracing cable to xtrnal cheap amp) i got signal through hole path until rvrb trafo out (tank in). Then i got pop/arcing again and immediately switched off. It's scary 🤣
So i guess ecl86 is fine.
I don't get splash sound by tapping springs. Tank is ok, so no recovery stage.

So my next step is to check recovery stage. Could bad AC151 germanium cause all this?

OK got it.  Yes, could be the AC151.  You can measure the DC voltages around the AC151.

It is possible the transformer is breaking down between primary and secondary and that is causing zaps.   That would take out the AC151.   Alternatively the arcing fault is on the tube side.  For example a bad contact on the output tube socket.   That would release the DC current through the transformer and cause a large voltage spike on the transformer - similar to a car ignition spark but less aggressive.

I doubt the AC151 can be taken out like that due to the reverb isolation. 

[sarakisof]

You can see my previous post's "EDITs".
I see, i will check ac151 voltages and will be back.

This amp has driven me crazy. The other Italian tube amp (Geloso 227) was easy and had been  flawlessly repaired bsck then.
This one is a bad guy 

[Rob Strand]

Please disregard my the last part of my post - see my last comment.

You can see my previous post's "EDITs".

Also I meant the output tube of the reverb circuit (not the main amp).   There's two power amplifiers.

[sarakisof]

Ok but again, it could be a bad germanium from the beginning right? I don't have smthng to loose if i measure ac151 volts i guess? I'm saying this, cause it's a whole story to disassemble the other chassis (preamp stage S points) 

I think it's either the ecl86, tube socket, trafo or ac151. Wish i could know for sure so i could order just one thing.

[Rob Strand]

Ok but again, it could be a bad germanium from the beginning right? I don't have smthng to loose if i measure ac151 volts i guess? I'm saying this, cause it's a whole story to disassemble the other chassis (preamp stage S points) 

Well at this point it would be best to pursue the AC151.

I think it's either the ecl86, tube socket, trafo or ac151. Wish i could know for sure so i could order just one thing.

I honestly don't know.  However before you had one problem, arcing, but now you have two problems, arcing and no recovery.  So now you might need to get two things.

The chances of a bad transformer is very low.  Still possible but low.    Corroded tube sockets on old organ sounds quite possible.   You might be able to clean the socket.

When you probe the secondary of the transformer and get an arc, does that happen all the time or even often?   That would be quite odd and might be a hint for the cause of the arcing.

[Rob Strand]

RG did a nice write-up on arcing,
http://www.geofex.com/ampdbug/popping.htm

At the end of the list. Cracked resistors is certainly another possibility.
Some crazy faults are caused by the smaller components.

[sarakisof]

ECL86 cathode voltages.
pin2: 0.12 V
pin7: 10.2 V
Is Triode Cathode normal?
Edit: seems fine.

[Rob Strand]

ECL86 cathode voltages.
pin2: 0.12 V
pin7: 10.2 V
Is this normal?
Edit: seems fine.

Yes seems OK.

I don't know if the voltage are exactly correct for that circuit but the voltages don't look unreasonable.

Stage 1: Ia1 = 0.12/150 = 800uA which would produce a 176V drop across the 220k plate resistor.
Stage 2: Ia2 = 10.2/560 = 18mA

[sarakisof]

When you probe the secondary of the transformer and get an arc, does that happen all the time or even often?   That would be quite odd and might be a hint for the cause of the arcing

Sorry my fault. In that early post i wanted to say i got arcing when probed at the tank output.
I got news. Tried again to probe there and i get no signal. So the chain stops there. If i get signal at tank's input, should i suppose to also get signal at output transducer?
That means that my tank is defective/dead output transducer? But again this is a brand new appropriate MOD tank. And it worked... (Trafo thoughts....)

[sarakisof]

Receive transducer reads 220ohms so it works. Probed again, still no signal at out receiver transducer (tried with both rca connected and disconnected). Should the signal supposed to travel across springs and can be heard at the receiver?

[Rob Strand]

Sorry my fault. In that early post i wanted to say i got arcing when probed at the tank output.

From an electrical point of view it would be virtually impossible to invoke an arc by probing there.    I suspect the arcing was a random event or perhaps the act of probing cause some vibration through the chassis and the real cause (bad connection?) created the arc.

I got news. Tried again to probe there and i get no signal. So the chain stops there. If i get signal at tank's input, should i suppose to also get signal at output transducer?
That means that my tank is defective/dead output transducer? But again this is a brand new appropriate MOD tank. And it worked... (Trafo thoughts....)

The bad AC151 might be shorting the signal.   The signal at the output of a reverb tank is *very* small so maybe hard to hear.  Unlikely to be the reverb tank.  You can measure the DC resistance of the output coil of the tank, you might expect 200 ohms but it could be anything from 40 ohms to 1k ohm.   Try measuring in circuit first and if the measurement  doesn't make sense you might have to desolder one of the wires - if the tank has connectors then it's easy to pull out the cable.

[Rob Strand]

Receive transducer reads 220ohms so it works. Probed again, still no signal at out receiver transducer (tried with both rca connected and disconnected). Should the signal supposed to travel across springs and can be heard at the receiver?

You beat me to it.  Yes, right on the mark.
So perhaps the AC151.

[Rob Strand]

You can check the AC151 by injecting a signal into the reverb recovery amp.
- disconnect the reverb tank output.
- inject a *low level* signal in the recovery amp input
- slowly increase the reverb level
  you should hear a clean signal coming through.

[sarakisof]

Just did it. Unfortunately AC151 works. I say unfortunately because troubleshooting to be continued ..

[Rob Strand]

Just did it. Unfortunately AC151 works. I say unfortunately because troubleshooting to be continued ..

Well in some ways it's good news.

So upto now
- reverb driver works, signal makes it to output of transformer
- reverb recovery works, signal passes through recovery amp

So the unknowns remaining are the connections to the reverb tank, both tank input and tank output.

Can you tap the reverb and hear a spash?   I thought a few posts back you said that wasn't happening (?).
That would point to a connection issue between the reverb tank output and the input of the recovery amp (AC151 etc).

[sarakisof]

Before continuing to the steps you mentioned, i just wanted to make sure i indeed had signal at transformer (so driver works), because earlier that i got signal it was like indeterminate and with this amp you never know 
And guess what. This time i had no signal by probing at trafo's output. And smthng else. Tried to probe at trafo's input (so to find if it is the curlpit) for about some secs (no signal again by the way) and the 1.5K 10W power resistor (between the two 32u filter caps) got extremely hot. Was smelling and took about 10 minutes to be able to touch it!
I mean really really hot.
Repeated (for just 3secs) just to verify and it started to get extremely quickly hot again and i switched off.

Is this normal? Maybe it could be a hint to the problem.
Input transformer = ECL86 pin3 and pin6...

[Rob Strand]

Before continuing to the steps you mentioned, i just wanted to make sure i indeed had signal at transformer (so driver works), because earlier that i got signal it was like indeterminate and with this amp you never know 
And guess what. This time i had no signal by probing at trafo's output. And smthng else. Tried to probe at trafo's input (so to find if it is the curlpit) for about some secs (no signal again by the way) and the 1.5K 10W power resistor (between the two 32u filter caps) got extremely hot. Was smelling and took about 10 minutes to be able to touch it!
I mean really really hot.
Repeated (for just 3secs) just to verify and it started to get extremely quickly hot again and i switched off.

Is this normal? Maybe it could be a hint to the problem.

Good idea.

The lack of signal is top of the list of things to debug.   A short across the reverb input or transformer output could cause that.

As for the hot resistor.  Yes, it is suspicious but it could be normal.  I can't read the expect DC voltage on the high voltage side of the 1.5k 5W resistor.   The output side is 330V.    If you measure the voltage across the 1.5k resistor you can work out the power dissipated as V^2/1.5k if it is close to 10W then it will get hot, like 200C.

However, more useful is to measure the 330V rail.  If there is a short or fault the high current will pull the voltage on that rail down.   It should be clear there is a fault.    If the 330V rail got pulled down to 250V then that would definitely increase the power dissipated in the 1.5k resistor beyond normal.