1976 Fender Twin Reverb Silverface: random crackling noises

[Lambertini]

Hi there,

After rearranging the heater wires, replacing the hum digger pot, the caps and all the carbon resistors with RN65s, I've largely reduced the noises coming out of an old 1976 Fender Twin Reverb Silverface 100W.

I still have some random noises occuring.

EDIT: I initially thought on vibrato channel only, in the reverb circuit. The noise seemed to never occurs when V3 removed, but in the end, it do. Just looks like it happens more often and more quickly on Vibrato channel. Something to do with gain?

You can hear in the linked audio (I've applied a +26 dB gain to the file).

EDIT: https://soundcloud.com/ekovah/noise

Any clue or advice to identify where it comes from and how to fix?

Note: I don't have an oscillo.

Thanks!

[tdvt]

There are many far more experienced here that will hopefully weigh in, but to me that clip sounds like a noisy board &/or tube noise.

I have 2 '79 Pro Reverbs, one I rebuilt last spring, but just couldn't get all the noise out, so I am rebuilding it with a new board. The second is just sitting until the first one is done, but unfortunately I anticipate it needing a similar fix.

The wax sprayed on the board at the factory can be tough to remedy. How waxy is your board & components? I believe they used less wax early on, mine both have a substantial coating.

Look for stray voltage directly on the board around the eyelets feeding the plates & other higher voltage connections.

 Scrape away any wax that you can, then work any bad areas with a heat gun, paper towel, paint brush q-tips & isopropyl alcohol.

[Lambertini]

Not that much wax, I cleaned it up as much as I could, and all solder joints have been redone wax free.

The problem is located in the vibrato channel, and more precisely in the reverb circuit: if I remove V3, the random white noise never occurs.

[stratomaster]

There are clearly large amounts of wax still easily visible on the board in your image.  Whatever method you used to dewax the board was insufficient. 

Measure the board directly for DC.  If it's present then you'll have to do a more thorough decontamination of the board. 

I recommend removing all the pots, jacks and switches from the front panel, desoldering the grounds to the brass plate, and flipping the board up vertically.  This will allow you to see the condition of the underside of the board for any solder blobs or pools of wax.  It will also allow you to get a wire brush and clean the brass plate that the pots ground to. 

Once the board is vertical you can arrange paper towels at the bottom and use heat to melt the wax off the board onto to towels without having to remove individual components.

You can then use naphtha to clean both sides of the board. Only once as much of the wax is removed as possible then use 99% IPA to dry the board.  Do these same steps to the insulator board under the main component board.

All of this is a lot of work, so make sure it's necessary.  Measure for DC.  It's easy.  Then check for other sources of noise.  Loose connections, solder bridges, broken leads, dirty or loose tube sockets, noisy tubes, leaky caps, etc.  Since this is a silver face consider redoing the reverb driver cathode circuit to BF specs.  You didn't include an image of the reverb part of the circuit so I can't tell if this is one of the hot biased drivers with no bypass cap. 

[BrownIsound]

Yes, definitely check for dc leaking around the board.

Did you remove the board during the replacement of components?

[Lambertini]

Here is a pic of the whole circuit before changing the resistors, during which I did additional wax cleaning. I'll post an updated photo tomorrow and check the DC voltage of the board itself at the same time and let you know.

I'd be very upset if a brand new orange drop / mica cap would leak, there's not a single 50 years old cap remaining in the circuit.

Moreover everything is fine on normal channel, and vibrato channel without V3 works just fine as well. No noise either with V6 and power tubes only. => I'm 99.99% sure the problem is located in V3 circuit.

Tubes are brand new as well but I've already experienced brand new non-working tubes... Will change V3 to see if it helps.

Have you ever heard that white noise growling which is coming and vanishing? Can the reverb driver be involved in it?

I'd really like to avoid all the work you suggested @stratomaster. Nevertheless thanks for the operating procedure, I'll follow it if required.

@BrownIsound, I haven't removed the board to change caps and resistors.

[Lambertini]

I've swapped the V3 tube, the amp was quite quiet for a while but the weird noise came back... Now we know the tube is not the root cause. ://
I've done a new recording and updated the main post with it. The noise is very clear at the beginning of the audio file.

Here it is: https://soundcloud.com/ekovah/noise

I'll post an updated photo tomorrow and check the DC voltage of the board itself at the same time and let you know.

[BrownIsound]

I’ve had new/not that old mica caps be leaky.

[Lambertini]

So I've checked the board for DC, found few spots at 0.005V and one at 0.02V. I unmounted the back board, cleaned it up completely, cleaned up the wax on the main board, checked the DC again, all spots around 0.002V like the rest of the board. Unfortunately the noise is still there. :(

@BrownIsound, it might be the 500pF cap before V3, the 560pF 3KV at the V3 cathode bias, or the the .002µF on the reverb output. I've some spare parts coming. I'll try to replace them when delivered.

[stratomaster]

Scrap the plate to cathode capacitor on V3.  Just cut it out and don't put anything in its place.  Put a 2.2k resistor bypassed by a 10uF or larger cap in place of V3 cathode resistor.  Put a fresh tube in V3.  This will revert the driver circuit to Blackface specs more or less.  By doing this you'll eliminate a plate to cathode capacitor and cool the bias of the 12at7 driver while maintaining relatively high gain to drive the reverb tank.

There is no high voltage at the 500pf at the input to the reverb driver, so that's the last thing I would suspect. 

[BrownIsound]

Scrap the plate to cathode capacitor on V3.  Just cut it out and don't put anything in its place.  Put a 2.2k resistor bypassed by a 10uF or larger cap in place of V3 cathode resistor.  Put a fresh tube in V3.  This will revert the driver circuit to Blackface specs more or less.  By doing this you'll eliminate a plate to cathode capacitor and cool the bias of the 12at7 driver while maintaining relatively high gain to drive the reverb tank.

Ah, this may explain some funnyness I’m having with my friends TR. some intermittent reverb channel issues, and the V3 tube being really hot, uncomfortably hot when grabbing its tube shield. I would imagine they may have gotten away with it at the time with RCAs, but some new production tubes might be running at the edge with this?

[Lambertini]

@stratomaster, thank you. I'll do the mod as drawn in the attachment. The parts should be delivered on Tuesday.

@BrownIsound, my V3 tube is also hot like hell, just can't touch it for a minute after amp is switched off.

[pdf64]

I suspect that the ‘removing the valve in V3 stops the noise’ thing may be a red herring.
Consider that noise generated by V3 circuit would be reverberised, shot noise would get ‘boinged’ by the tank.

[shooter]

the quick math says it's dissipating the max power, ~~~~~2.5W each half, (6vdc/470ohm = 1.27mA....400v*1.27mA = ~~5W/2)
add in some heater heat, leave fingerprints behind 

[Lambertini]

EDIT: Not good in math... corrected the calculation below.

The reverb driver is rather hot, around 458V instead of 440.
6 VAC / 470 ohms = 12.7 mA
458 VDC * 12.7 mA = 5.84W
<=> 2.92W per triode.


Would a 12AT7 last for more than 20 minutes?

That could explain why replacing the former TAD 12AT7 with a brand new Tungsol made the noise disappear for about 20 minutes?

[PRR]

> around 458V instead of 440

12AT7 is rated 300V design center (say 330V measured in production) so you are way over-volting it.

For what? You don't have to fry the tank, there is a gain-stage after it.

While it is un-original, I'd be thinking of a voltage dropper. Get the plate down below 350V.

[Lambertini]

The circuit is stock, the 458V feed the 2 plates, I'd assume 229V each, but I'm not comfortable with formulas...
You have the schematic 5 messages above.

EDIT: If I take the cathode voltage value from the schematic (6V) the plate-to-cathode voltage would be 452V.
Using https://robrobinette.com/Tube_Bias_Calculator.htm, I found 2.7W per triode, over the max 2.5W from tube spec.

I can measure the actual cathode voltage drop and increase the 470 ohm resistor value to 680 ohms or more, waiting for the mod proposed by stratomaster I'll do on Tuesday.

[stratomaster]

> around 458V instead of 440

12AT7 is rated 300V design center (say 330V measured in production) so you are way over-volting it.

While it is un-original, I'd be thinking of a voltage dropper. Get the plate down below 350V.

The Boogie Mark 1 reissue uses an 820 ohm cathode resistor with a 15uF bypass cap and a 22k series dropper between the PS node and the reverb transformer primary. They're also using a 12AT7 as the driver.  This arrangement has been stable in my Mark 1 for over a decade. The only changes I've made are using a 5w ceramic 22k in place of the carbon comp, replacing the 500v ceramic .003uF cap with a 2kV film, and swapping the order of the 470k and 100k voltage divider to get more signal into the reverb.  Neither affect bias/dissipation in the tube. 

Consider this an option for dropping the voltage, a 5w 10-22k series dropper. 

[Lambertini]

I suspect that the ‘removing the valve in V3 stops the noise’ thing may be a red herring.
Consider that noise generated by V3 circuit would be reverberised, shot noise would get ‘boinged’ by the tank.

I'm confused to tell you, you're right. I just left the amp on for a while w/o V3 and... the noise finally happens.
With V3 plugged, it seems to come just more often and faster.....

[PRR]

Consider this an option for dropping the voltage, a 5w 10-22k series dropper. 

Works, but leaks audio. Probably want to also try a 10++uFd cap at the junction to steady the idle end of the winding.

I know the 12AT7 will not blow-up at 301V, and I know thousands of these amps are still alive and reverberating. And if HOT reverb DRIVE had benefit, like hot speaker drive, then I'd try it hot. But this looks like Leo got lazy, 1960s RCA/GEs would stand it, and the general rise from 115V to 125V came later.

[stratomaster]

I suspect that the ‘removing the valve in V3 stops the noise’ thing may be a red herring.
Consider that noise generated by V3 circuit would be reverberised, shot noise would get ‘boinged’ by the tank.

I'm confused to tell you, you're right. I just left the amp on for a while w/o V3 and... the noise finally happens.
With V3 plugged, it seems to come just more often and faster.....

Any chance your phase inverter tube was once your reverb tube?

Also the plate to cathode cap is still in circuit reverb with the tube removed.  Removing that tube doesn't guarantee the reverb driver circuit is healthy.

Failing that, it might be time to start grounding grids or lifting coupling caps to see if you can get more narrow targeting of suspect parts of the circuit than just pulling tubes can provide. 

[Lambertini]

After few days of testing, the problem seems to be fixed without modifying the reverb circuit.

After noticing the "noise" (noises?) actually occurs even w/o V3, and having heard it w/o any preamp tubes at all after leaving the amp powered on for a fairly long time, I did the following:

• Replacement of all the old and loose power tube sockets (I didn't expect that would be so long and tedious!)
• Upgrade of the cheap screen resistors (previously replaced with low cost comps) by high grade 470R 2W resistors
• Replacement of V4 tube that was producing small static noises when moved

I haven't heard the crackling noise since.... yet.

[stratomaster]

Small static noises when moving preamp tubes can usually be resolved by cleaning the socket with a solvent only contact cleaner.  You can sometimes also bend the contacts to get a better grip. Finally, using tubes with thicker leads can be a last resort. I find the Chinese tubes to have the thinnest diameter contacts. 

If you're very adventurous, loop some 22ga buss wire a few times around the lead and leave it long at the beginning and end winds.  You've now made a heat sink.  You can tin the leads if you're quick about it.  This will gain a bit of purchase in the socket.  Practice on your cooked reverb driver.   

I'd still cool the bias in the reverb driver with the 2.2k bypassed resistor.  It'll reduce unnecessary wear on that tube.

You should also make sure your PI tube was never used as the reverb driver in the past.  It may be worn or damaged if it was. 

[Lambertini]

The PI tube is fresh and the components for retrofiting the reverb circuit to AB763 arrived. I'll do it anyway tomorrow and let you know.
I already tried to clean and thighten the socket slots, with no good results, the metal is very thin and used. I have few spare good quality 9 pins sockets for the 4 preamps tubes that still have the stock ones, will replace them as well.
Thank you!

[Lambertini]

V3 socket, Reverb transformer and V3 cathode bias replaced. V4B grid leak resistor changed to a 220K. New 12AT7 on V3. Hope this will save the tube.
Amp is still quiet but the reverb is now a bit overwhelming, need to fix that.
Anyway the initial problem is solved, thank you all!

[sluckey]

Amp is still quiet but the reverb is now a bit overwhelming, need to fix that.

Replace the stock 100K linear Reverb pot with a 100K audio taper pot. You will still have the same max reverb but will have much better control of the level with the audio taper pot.