Farfisa Tube Amp High Pitch Oscillation RVRB Problem

[sarakisof]

Hello everyone. I was given a 60s Farfisa full tube guitar-organ amp from a good friend and after a long restoration and repair project got this beautiful amp working great again, except for reverb. ECL86 (triode+pentode) was missing from chassis and driver transformer was dead (primary windings open reading). Did my best to save it, unwind as much as i could, but unfortunately the break seemed to be deeper in the core. After some hundreds of turns i quit.

It uses an ECL86 (triode pentode) ---> driver transformer ---> rvrb tank ---> AC151 for recovery stage.

Schematic:
http://b.baldach.free.fr/farfisaFR40/FarfisaFR40_Schematic.jpg

I finally found an ECL86 from an old kind man here, told me it was working back then. Slightly used. He hadn't test it though.
Now about the transformer, found some guys with same issue that used Hammond 1750AX (7500Ω-8Ω) and worked perfectly.
I was suggested any trafo about P:5-10Κ and S:4-8Ω would work there.
So i found an OT from an old defective tube radio i had laying around for parts. Radio uses one EL84 (ECL86 is half EL84, exactly my case).  Radio's OT is P:7000Ω S:4Ω and about 3-5W, so it would work perfect as replacement.

OT 7K/4Ω and ECL86 were installed.
Everything was ready, fired it up, push the reverb footswitch, tapped the springs with my finger and could hear the big splash at least.
Then, while i was grabbing my guitar i heard some loud weird soft scratchy/heartbeat noises, like not good contact or smthng, signal was coming and leaving.
I tried a new rvrb tank with similar specs (4AB3C1B) , as the original one had one transducer's magnet broken (one spring only defective, not both) just to see if that caused the issue.
To my surprise, when i powered on and tubes get warmed up, i got this oscillation high pitch kind of noise you can hear in my vid below.  This same thing happens since then. Seems like it starts from around 1650hz and gets decreased to lower frequency range.

- Noise appears only with the ECL86 installed on chassis. Pulled it off and the issue disappears indeed.
- With the tank disconnected, same issue again.
So it's either the tube itself or the transformer and not after that, like in recovery stage, AC151 etc.
- Osc noise is there with all pot volumes down.
Don't have another ECL86 to test and they are pricey, so wouldn't like to get one just to see the issue remains.

I uploaded a quick video so you can hear the noise.
*WARNING: Be careful with volume. Headphones aren't recommended. 

https://we.tl/t-317qJZTEsV

What do you think guys? Would it be wise to disconnect trafo again (as it was before, when original was dead) in order to isolate the problem to ECL86 only, or it could be smthng else?

[sarakisof]

Oscillation problem solved. I had forgot to ground the trafo's mount to chassis, so the secondary gnd winding lead (the one that goes to rvrb input gnd sleeve RCA) was ungrounded.
It was also feedback network issue, swapped primary windings and osc completely gone.

Now the only thing remained is to get signal mixed with reverb. At this moment i get only the effect splash sound when tapping the springs and the dry signal.
Input driver troubleshooting just started.

Any opinions are welcomed.

[anotherjim]

I have an idea (needs checking) the ECL86 is an improved ECL82? The 82 is still easy to find. Pin out may be different.

If the input of the spring is fully isolated, you could try driving with a small solid-state amp with driver tube taken out ( because no load). Spring sounds but no signal implies nothing from the driver tube?

[sarakisof]

I have an idea (needs checking) the ECL86 is an improved ECL82? The 82 is still easy to find. Pin out may be different.

If the input of the spring is fully isolated, you could try driving with a small solid-state amp with driver tube taken out ( because no load). Spring sounds but no signal implies nothing from the driver tube?

ECL86 is more or less half ECC83 and EL84 in the same glass.

Any ideas of solid state amp circuit i could use there?

[anotherjim]

ECL82 is the same kind of tube. E=6v heat, C=triode, L=pentode. It was pretty much standard for ceramic cartridge record players. I think the ECL86 has slightly higher output power from the L (but less than an EL84) and higher voltage gain from the C. But if you're prepared to consider the reverb as a sideshow, I'd have thought the 82 was worth considering as a substitute.

As long as the reverb tank input is low impedance, any amplifier should be able to push enough audio into the reverb. I only suggest it for testing purposes and it looks like it's got a phono plug connection on the tank?

[Rob Strand]

Input driver troubleshooting just started.

Any opinions are welcomed.

If the thing oscillated because of the transformer connection issues then you would assume the reverb driver amp is working.   In fact I'd say both tube stages seem to be doing something.  Can you trace the signal feeding the reverb driver amp?  Maybe the signal isn't making it to the driver amp.

[sarakisof]

 I just tried an ECL82, started to get blue sparks inside the tube and i immediately turned it off. So it's not the tube itself. Hopefully, ECL82 didn't get damaged. Neither blown fuse or smthng bad.
So whatever tracing i will do, it will be with no ECL86/82 installed. Don't want to shot them. 
I think i had blue sparks with the ECL86 too, when first installed it. But since then no, maybe i managed already to damage a part of it.

- Before the ECL82, i had tried the ECL86 once again, to check tube's voltages. It was OK, I had 340V on pin6 (schem calls for 320V), BUT when that loud pop appeared on speakers and was like "smthng bad is happening for a moment" voltage suddenly decreased to 190V for a second and immediately came back again to 340V. 😯
Could be smthng wrong about pins3,5 that are connected with B+? (arrow symbol in schem corresponds to B+)

- I get splash sound either with Reverb footswitch enabled or disabled. Don't know if this is normal and explained from schem, as i cannot find the switch "Pedal R" (like Pedal T for tremolo) in schemo. If it's this one at S1 - P1 points in schem, then it's not normal that i get splash even if R switch is pushed off.
Could you please make this area on schem clear to me? As i understand it, by closing the switch the rvrb out signal after Rvrb vol pot gets shorted to ground and stops there before getting to S1-P1/PI stage, right? Don't know if this has to do with signal not going to driver amp though.

[sarakisof]

EDIT Add: I get continuity between S1-P1 point to GND (20Ω) , whatever i do with rvrb footswitch. Either on or off position. 😯

[PRR]

> I get continuity between S1-P1 point to GND

That's wrong. Remove stuff until it un-shorts. May be a stray bare wire. May be the switch has sprung its guts.

[Rob Strand]

I get splash sound either with Reverb footswitch enabled or disabled. Don't know if this is normal and explained from schem, as i cannot find the switch "Pedal R" (like Pedal T for tremolo) in schemo. If it's this one at S1 - P1 points in schem, then it's not normal that i get splash even if R switch is pushed off.
Could you please make this area on schem clear to me? As i understand it, by closing the switch the rvrb out signal after Rvrb vol pot gets shorted to ground and stops there before getting to S1-P1/PI stage, right? Don't know if this has to do with signal not going to driver amp though.

A brief explanation of the signal flow is this:

- The two preamps mixed via the resistors before the 3xcap network at the top/centre (2x4n7 + 47n + 1M etc).
   (The caps remove thumps from the tremolo.)
- The output of the 3xcap network feeds down to S8.   That's essentially the dry signal which couples across from S8 to P8.
- The dry signal then feeds the dry input of the power amp via the 100k and feeds the reverb drive circuit directly.
- The output of the reverb comes off the transistor stage and feeds the 10k pot,
   the switch and then goes to S1, then P1.
   The switch kills the reverb by shorting the output of the 10k pot to ground. [That's why there's continuity to ground.]
- The reverb mixes back with dry signal via the 100k+250k pot just before the input of the power amp (22nF + V5).
   The dry signal at this junction is the 100k dry input I mentioned two points back.

As for the splash.   The thing to notice is the signals on P8 have a path to ground and that's the ground path
for the grid of the reverb driver tube input V6.  When the reverb switch is grounded it changes the impedance to ground.
The change in impedance is *very* small it's less than 100k + 100k + 250k pot  vs 100k + 100k + 250k pot  + 10k.
In order to have a significant splash I'd be suspicious there is DC on the P6 node.  How it gets there I don't know.
The small leakage from V8 would seem unlikely to cause it.   I guess it would be wise to measure the DC voltage
there and see what happens to with the reverb switch on and off.
   
--------------------------------------------------------------------------------------
EDIT:
Something that occurred to me later is if there is a poor connection on S6/P6 you could be getting
DC across the grounds between the two halves of the circuit.  You could measure the voltage between
the grounds.  Having power ground and signal ground passing over one pin is really asking for trouble in
the long term.

[anotherjim]

I just tried an ECL82, started to get blue sparks inside the tube

I did say check if the pin out was different. It is and seriously so! Never try substitution without checking first!
Really, I mentioned the ECL82 because if they are easier to get and cheaper, then the chore of rewiring the socket could be worth it.

[sarakisof]

Sorted it out, i think case solved.

- First, by chopstick tapping i realised i had a bad solder at 200uf EL34 cathode cap.

That was causing the 50hz hum i had before. Amp is buzz free now, almost dead silent.
Here, amongst others, i had spotted back then.
https://www.diystompboxes.com/smfforum/index.php?topic=126741.0


- Second, about reverb issue and S1-P1 shorting to ground whatever position rvrb switch was, after troubleshooting and removing stuff, connections were fine.
Turned out the problem probably (i say so cause I'm not 100% sure it was that, as it wasn't a "found that wire shorting and remove it" case, you will see what i mean) was at the rvrb pedal footswitch itself. By disconnecting the switch from circuit and continuity test it for functionality, i realised that in one position it was open as it should, but in the other position stayed "open" 2000ohms or smthng and was taking enough time until resistance was gradually dropping to 20, 10 and finally 5ohms beeped, as it should.
Read somewhere this can happen specially in old footswitches, i think they call it sticky symptom or smthng.
Used  my trusty Teslanol Oszillin T6 contact cleaner w/lube and problem solved. 
I finally have working reverb unit guys! 











EDIT: The only thing that worries me a bit now is that sometimes while playing with rvrb on, i get some random splashes once in a while. That could be due to vibrations, bad rca plugs, interference etc. Voltage spikes/Driver issue?. Maybe the ECL86 that isn't NOS, the trafo replacement that i use now or smthng with this

Having power ground and signal ground passing over one pin is really asking for trouble in
the long term.

?

[sarakisof]

Remember this thread?
Today i got sparks in ECL86 tube again. 

> I get continuity between S1-P1 point to GND
That's wrong. Remove stuff until it un-shorts. May be a stray bare wire. May be the switch has sprung its guts.

@PRR : Maybe it's not wrong. If you look at the scheme the upper leg of 10K lin Reverb pot (S1-P1 area) gets connected with gnd. So with rvrb pot all way closed to zero its center leg gets normally connected with the upper leg.
So, when pot is closed S1-P1 gets connected to gnd.
I had the pot at zero that's why i got continuity.


All those days things weren't exactly 100% normal as i always had cracks and splashes from tank only when i had rvrb foottswitch on.
As i have said above, when ECL86 removed, amp runs perfectly.
Smthng about rvrb circuit is wrong.
What other things should i check guys?

[Rob Strand]

All those days things weren't exactly 100% normal as i always had cracks and splashes from tank only when i had rvrb foottswitch on.
As i have said above, when ECL86 removed, amp runs perfectly.
Smthng about rvrb circuit is wrong.
What other things should i check guys?

To me it looks like there's a problem with the ground connection S6/P6.   That causes DC to be injected into V8 pin 1.

I suspect the thing that makes it worse is the DC coupling into V8 pin 1.   You could probably reduce the chances of bad symptoms by putting an AC coupling cap (say 22n) between the node with V8 pin + 10M and the line to P8 (ie put a cap in place of that short vertical wire on the schematic).   If the resistor on pin 1 is  10M you might have to drop it to 1M as the grid leakage through 10M might affect V8 biasing.

[sarakisof]

Thank you Rob. So you think it has to do with DC coupling there.

Will try the mod and report back. You mean like that right?

[Rob Strand]

Thank you Rob. So you think it has to do with DC coupling there.

Will try the mod and report back. You mean like that right?

Yes, exactly like that.

Just to add a bit of explanation:  The first stage of V8 has quite a bit of gain.    The DC path to ground for V8 is on the "other side" of S6/P6.   If there is a DC drop across S6/P6 the DC coupling of V8 will amplify the DC drop by a large amount and possibly saturate V8.   It is a set-up asking for trouble, and even more so on old equipment with old connectors.    AC coupling V8 will prevent the excessive shifts.   I'm not 100% sure why/how the sparks are occurring.

[sarakisof]

Got it. This could explain random loud crackly solaahes, but what about sparks?

I'm not 100% sure why/how the sparks are occurring. 

What could be the reason for this according to schem? Should i check ECL86 nearby components, what test measurements should i do with ECL86 removed?

[Rob Strand]

What could be the reason for this according to schem? Should i check ECL86 nearby components, what test measurements should i do with ECL86 removed?

The only theory I have at the moment is suppose the DC at the input of V8 is jittering off and on erratically due to some bad connection.   The first stage of V8 will the turn off and on quite hard.   The off and on signal drives the second stage of V8 hard off and on.   The load on the second stage is a transformer.   If it's getting turned hard off and on then perhaps there a strong back emf and that's causing arcing.   For example if you get a battery and old it on an iron core transformer or inductor then let disconnect you get a very large arc.   I had great fun doing that as a kid.  It's like the ignition spark on a car coil (only smaller).

In this case the arcs would be short term and intermittent more than a continuous arc.

The ECL86 datasheet shows maximum voltages of 550V with a 320V rail it wouldn't be hard for the plate and grid 2 to get to twice 320V = 640V.

Can you see exactly where where is occurring?    I don't quite understand exactly what's happening when the fault occurs.

-------------
EDIT: Something that popped in my head was if the ground at the ECL86 is disconnecting maybe there is flashover to the heaters.  Bit of a long shot.

[sarakisof]

Do you remember this old thread? Since then i was living with ECL86 arcing and crackling drip reverb noises. The other day reverb stopped working.

Rob had mentioned this:

I suspect the thing that makes it worse is the DC coupling into V8 pin 1.   You could probably reduce the chances of bad symptoms by putting an AC coupling cap (say 22n) between the node with V8 pin + 10M and the line to P8 (ie put a cap in place of that short vertical wire on the schematic).   If the resistor on pin 1 is  10M you might have to drop it to 1M as the grid leakage through 10M might affect V8 biasing.

Actually there is such a mod already from factory there, although missing from schematic. They have put a Philips coloured ceramic 150pF at ECL86's pin1 (P8), as AC coupling cap.

Today, i realized that while recapping, i had replaced everything except for one tiny electrolytic. It was the 1uf/6V connected to the cathode resistor of ECL86's pin2. Actually they had put an old Ero 5uF/6V (damn Farfisa, common schem-chassis non-correspondings  😡). This cap measured 32pF instead of 5uf so i replaced it.

I still have no reverb, but could that 1uf/5uf pin2 cathode cap caused all those bad things and probably got ECL86 shot?



With orange arrow the 5uf cathode cap that measured 32pf.
With green arrow the AC coupling 150pF cap that actually exists, although missing from scheme.


Edit: I will try to find a NOS ECL86 (previous was used) and change ECL86's tube socket with a new one.

[PRR]

....could that 1uf/5uf pin2 cathode cap caused all those bad things and probably got ECL86 shot?...

No; but it sure could cause high pitch squeal. (100% NFB around power amp.)