Fender Twin - steps to reduce hiss?

[Roe]

I've removed all shared cathodes in mine, except for the phase inverter obviously

[daxliniere]

V4 on the The Twin schematic is the phase splitter for the output tubes. V8A//B is the reverb driver, V9B is reverb recovery gain, V9A is the post-8M2 gain stage.
Perhaps we should just stick with tube functions as the names rather than numbering.

[lonote]

Thanks Ionote. I read about halfway down the page and it was all getting a bit over my head.
But basically what you're suggesting is to replace C105 with 100uF and add 100uF in parallel with R56?

I'm definitely keen to try this, though reverb noise is lower down on my list of noise concerns for this amp. But, again, thank you.

Sorry, late to reply, but yes, those are the two points.

Not sure when they changed, but many of the later year silverface models used un-bypased cathode resistors on that tube, sometimes less than 1k, but the earlier versions used a bypassed 2.2k in that position. Changing to the higher value is often done as a mod to help the tube run a little less hot.

As far as the caps, you should be able to just temp in a large bypass cap across each of those cathode resistors, with out removing anything, to see if it helps with your issue.

[daxliniere]

As far as the caps, you should be able to just temp in a large bypass cap across each of those cathode resistors, with out removing anything, to see if it helps with your issue.

Yes, good point, that would be an easy way to test quieting down the reverb send/receive. I'm hoping I can make a lot more progress on the rest of the signal path, though.

[daxliniere]

Hey everyone! I'm doing the mods at the moment, taking a recording of the amp with a pre-recorded guitar loop between each component swap.

I'm up to separating the reverb driver tube which has 2 stages in parallel. Current value of cathode bias resistor is 1.2k - should this be 600R on each triode or 2k4?
Will also add the 100uF bypass caps.

Cheers!

UPDATE: I went with two 1K5 cathode resistors for V8A//B.

[B Ingram]

... 2 stages in parallel. Current value of cathode bias resistor is 1.2k - should this be 600R on each triode or 2k4?

2k4 each triode.

[daxliniere]

Thank you, sir!

Incidentally, increasing the reverb recovery cathode bypass cap from 22uF to 122uF (simple parallel test) seems to decrease LF hum, but increases white noise around 1kHz.
I'm going to swap the 1K5s for 2K4s now and record a new test.

[daxliniere]

I made all the suggested changes (except the Miller effect caps),
R126 (91K) - upgraded to Vishay metal film,
R101+103 (8M2, channel mixer) - upgraded to Vishay metal film,
R3+20 (1M, input bypass resistors) - upgraded to Vishay 3W metal film,
R4+5+21+22 (68K, input grid resistors) - upgraded to 10K Vishay metal film,
Split paralleled reverb send triodes of V8 and bypassed with 100uF and 2k4 each.

The reverb return is still incredibly noisy, though the head is currently disassembled, so is probably lacking some shielding.
The accusonics reverb tank isn't shielded on the underside: I wonder if gluing some simple aluminium kitchen foil underneath would help? With a ground connection, of course.

[Roe]

to improve the reverb, I'd recommend the SSS threads as well as the John Mayer ones. The strange wiring of the reverb pots on SSSs do work, and the 220k in parallell with a 500pf before the pot adds clarity to the reverb by giving a boost in the high mids

[daxliniere]

Hey Roe, thanks for that. Forgive my ignorance, but what is SSS? And John Mayer in this context?

Also, at some point I'll share the audio files taken after each mod.

[daxliniere]

So what do we think about changing the 6L6GC's G2 resistors (R60,61,64,65) for ultra low noise metal films?

[Roe]

noise/hiss is mostly generated by the first gainstages.

SSS is dumble steel string singer, e.g. https://ampgarage.com/forum/viewtopic.p ... 8&start=15

Mayer here refers to his Two Rock signature amp

[daxliniere]

Thanks Roe!

I did manage to almost completely silence the reverb. I had the idea to shield the open side of the tank (foil wrapped around cardboard, contacting tank's bare metal), but also happened to discover that the position AND orientation makes a HUGE difference. I would never guessed that upside-down vs right-side-up would make any difference whatsoever when the relative position of the two transducers stayed identical.

The only remaining quirk is that there seems to be an interraction between the reverb tank and the tubes, especially the two middle output tubes. Very high levels of gain or extremely close proximity generates quiet, HF feedback (Usually 5.5kHz, but frequency is variable +/-500Hz). Is that likely to be a faulty output tube? The output tubes aren't microphonic when doing the chopstick test.
The tank originally lived in the bottom of the combo, but now that I've 'beheaded' it, the tank needs to be closer to the rest of the amp.

I did end up installing WIMA 100pF gate-cathode caps on V1 and V7 (CH2 and CH1 triode #1s). Last thing I'm waiting on is stock of the 470R output tube resistors.

Anyways, the amp is much quieter now and the quality of that noise is less obtrusive. I have learned so much during this process and wanted to thank everyone for sharing their expertise. I truly appreciate it.


All the best,
Dax.

!!8c0c5241-5eb9-4862-981d-440eba0c8e61.jpg

[Roe]

Thanks Roe!

...
The only remaining quirk is that there seems to be an interraction between the reverb tank and the tubes, especially the two middle output tubes. Very high levels of gain or extremely close proximity generates quiet, HF feedback (Usually 5.5kHz, but frequency is variable +/-500Hz). Is that likely to be a faulty output tube? The output tubes aren't microphonic when doing the chopstick test.
The tank originally lived in the bottom of the combo, but now that I've 'beheaded' it, the tank needs to be closer to the rest of the amp.
.
Dax.

sounds like oscillation in the reverb circuit. normally, snubbers are added to the reverb driver (and/or recovery). But separating the two cathodes of the driver tube and installing seperate grid resistors on it is often enough

[daxliniere]

Here are some photos of the finished areas.

PXL_20250117_205214953.jpg

PXL_20250121_000628242.jpg

PXL_20250117_205221537.jpg

PXL_20250117_205241122.jpg

PXL_20250117_205157933.jpg

PXL_20250117_211907645.jpg

PXL_20250117_211941810.jpg

PXL_20250121_000616179.jpg