135w UL Twin Reverb bias math

[winterturtle]

Good afternoon- I'm checking the bias on the twin. 1980 Silverface Twin Reverb -- dead stock except the electrolytics are all new, grid resistors are new, PI tube is new. Pots and tube sockets have all been D5->F5. She had a mix of power tubes so I slapped in some JJ 6L6GCs. Tweaked the matching and hum balance pots then used the "transformer impedance" method to check my dissipation. Thinking my approach was appropriate but hoping someone doesn't mind checking over my shoulder & offering experiences =)

https://imgur.com/AGCzwXR

I've heard these amps usually run cold but this seems excessively low?

[Latole]

I've heard these amps usually run cold but this seems excessively low?

When I do bias, I alway check sound with guitar. Most of the time, I did not ear sound difference at 50 %

[winterturtle]

When I do bias, I alway check sound with guitar. Most of the time, I did not ear sound difference at 50 %

It definitely sounds great to my ears. I have it hooked to a Jensen 2x10 and it sang. I'm very happy with the level of hum -- that's down quite a bit too.

[pdf64]

483V seems a little low. What mains and heater AC voltage do you measure?
What’s the idle DC voltage at the OT CT-2k7-standby switch node?

[winterturtle]

483V seems a little low. What mains and heater AC voltage do you measure?
What’s the idle DC voltage at the OT CT-2k7-standby switch node?

121.1vac with her on and stable at the power transformer leads.
6.34vac across the heater wires
493vdc from ground to that node where the big 2k7 ceramic and standby voltage connect. I measured with speakers connected and no input. Fingers crossed she's not acting up in other ways!

[Latole]

483V seems a little low.

480 is fine.
Many Silverface Twin R  schematics show 470 V and less

[pdf64]

483V seems a little low.

480 is fine.
Many Silverface Twin R  schematics show 470 V and less

Indeed, on some it’s a lot less, but the 135W has 500V noted.
So the additional test results now reported help to verify all is well 
https://el34world.com/charts/Schematics/files/Fender/Fender_twin_reverb_sf_135_schem.pdf

[Latole]

483V seems a little low.

480 is fine.
Many Silverface Twin R  schematics show 470 V and less

Indeed, on some it’s a lot less, but the 135W has 500V noted.
So the additional test results now reported help to verify all is well 
https://el34world.com/charts/Schematics/files/Fender/Fender_twin_reverb_sf_135_schem.pdf

With 480 or 500 amp is fine.

[pdf64]

Thanks for the North American perspective, it’s just that in the UK, they typically run higher than than the schematic nominal levels.
To the point that I typically set them at 260V on our 240V mains, to get them in line and cool them down a bit.

[winterturtle]

I did a little resistor shopping and decided to try the jumper method - then soldered as in the final table (2x47k in parallel = 23.5k) at 67/66% which sounded best to my ears. Placebo? Certainly a risk haha -- but this seemed liveliest.

Is there a high level conclusion or concept to learn about that would help me understand the implications of idle voltages? As I stand I'm now at 475v and still hoping I don't have anything immediate to worry about.

*the green resistance #s were taken when nice and warm

[shooter]

understand the implications of idle voltages?

a base-line value that can be used for future reference
a stable value that makes ohms law easy
IF cloning, comparison on how close you got

[pdf64]

When the design maximum system was introduced, it was expected that the limit would be derated (by an application appropriate degree) before being used in such calculations.
Unfortunately people seem to overlook / ignore / be unaware of the distinction, between design centre and design max systems, and they’re used interchangeably.

Unless undertaking bias checks in a limit condition, eg with the mains variac’d up to 130, I suggest to derate design max type limits by about 13% to approximate its design centre equivalent.
So about 26W for a 6L6GC, hence its hottest fixed bias idle guideline becomes about 18W, rather than 21W.
To achieve adjustable bias here, the 33k resistor can be replaced by a 22k in series with a 22k trimmer.



https://www.one-electron.com/Archives/RCA/RCA-AppNotes/RCA%201958%20AN-174%20Design-Maximum%20System%20for%20Rating%20Electron%20Tubes.pdf

[HotBluePlates]

... I've heard these amps usually run cold but this seems excessively low?

It doesn't look excessively low, especially for a 135w Twin (where the old blackface operated the same tubes for a bit over 80w).

... 67/66% which sounded best to my ears. ...

Is there a high level conclusion or concept to learn about that would help me understand the implications of idle voltages? ...

You don't play an idling amp.  The Voltage & Current & resulting Dissipation that matters happens at or a little before maximum clean output power.

Amps like this rigged to run further into Class AB (towards the Class B end of the spectrum) draw large peak plate currents, and could overheat their tubes.  So the tube is idled cooler, in the expectation this will cause it to be shut off longer at high power outputs, allowing the late som time to cool off.  Restating, that's:

     - Idle cooler to
     - Shut the tube off longer to
     - Offset high peak plate current (and higher average plate current), which then
     - reduces average plate dissipation

What idle plate dissipation results in the best idle bias is only ever a guess.  It's possible to rig a test for average plate current with signal, and average plate voltage with signal, to know the plate dissipation in-use.  But then you need a signal generator & a dummy load, as well as more than 1 meter at a time.

[shooter]

It's possible to rig a test for average plate current with signal,

I use 1 ohm resistors, cathode to ground on each PA tube  (assuming two tubes or two pair)
I connect 2 meters that have MIN/MAX/AVG settings
then I jam hard for 5 minutes, evaluate the results using AVG for the math, re-tweak bias as needed
repeat til the tech in me is happy