6V6 Bias on SF Princeton Reverb

[terminalgs]

A buddy brought me an all original ca.73 SF Princeton Reverb with complaint that volume "drops out".  This amp sounds great, everything checks ok.  original cap-can,  and coupling caps not leaking any DC.  no noise or hum or anything.  just this drop out issue.  I couldn't get it to repeat the behavior but I didn't really play it dimed for long (at insistence of ears).


Tubes are all original or at least from 70s.


5U4GB: drop out could be this tube.  When I give the amp back to him, I'll tell him if it drops out again to try another 5U4GB.  B+ stays high with this 5U4GB (at least on the bench), so it doesn't act like the worn rectifier tubes that my limited experience has encountered. 
FWIW, 5U4GB tests 100% strong on tripplet 3403 tester.


on the other hand, the old 6V6GT's test as "poor" in tripplet and with the original Fender bias circuit, the bias on the 6V6s was low, like 18mA and 19mA @ 400VDC.  so I popped in a set of matched JJ 6V6S's to compare bias and voltage readings of old 6V6GTs to 6V6Ss.


Which circuit?: the BF version was AA1164, and the 1970 & later SF version was AB1270, which this particular amp should fall into. However, this one has bias circuit with a 100K, 22K, and 50/70V cap (plus diode).  so it's a hybrid of the 1164 and the 1270:

• AA1164 =  100K, 22K, 25uf/50V
• AB1270  = 100K, 27K, 50uf/70V
• This one = 100K, 22K, 50uf/70V (and looks all original FWIW).

https://el34world.com/charts/Schematics/files/Fender/Fender_princeton_reverb_bf_ab1270.pdf


https://el34world.com/charts/Schematics/files/Fender/Fender_princeton_reverb_aa1164_schematic.pdf


So in go the 6V6S JJ's and bias is hot, both at 416VDC plate to cathode, 27mA for 11.2W


I changed the resistance of the 22K in the bias circuit and got these two:


431VDC (plate to cathode)  @ 22mA  for 9.49W (67% of JJ's 14W)
418VDC @ 26mA for 10.8W  (77% of JJ's 14W)


pretty hot all around but sounding good.  more articulate and punchy compared to the worn out tubes, which were more tubey, mid-rangey and rounded (BTW this amp has a big 40W or 50W ceramic Pyle). I'd say with the JJ's, very Jerry Garcia rig sounding,  with the old tubes... more lo-fi blues amp sounding. Both sound good.  I was prepared to send it back to said buddy with the lower 22mA setting..  and say "how's this?  to your liking?"


But said buddy delivered Princeton with a box of NOS 6V6GTs so I decided to dig into those.  all singletons, no matched pairs of course.  four coin-based JAN tubes (so guessing 70s era), a 7408, a few others, and I managed to find two 6V6GTs that biased out within 10% of each other.  With those I've tried some different biases:

• 416vdc  (plate to cathode) @ 27mA for 11.2W (93% of 12W) -- original bias circuit with 22K R.
• 435vdc (plate to cathode) @  19mA for  8.265W   (68% of 12W) -- 22K + 3.3K.
• 436vdc 17mA for 7.76W (64% of 12W) -- 22K+4.7K (pretty much AB1270's 27K)

as you can see as I lower the bias, plate vdc rises.  at the 5U4GB B+ rises to 440vdc with the tubes at 19mA.  One thing that makes me nervous is the 5U4GB beats the 6V6's to warm up and for 5-10 seconds B+ vdc is near 500V on the 450V rated cap can (now sorta wish these had a stand-by to let the tubes warm up!).


the schematics don't show what the original BF or SF bias current was, but it seems like it must have been quite high. I can't find any threads on forums or passages in books that say for a class AB amp 70%-95% of max dissipation is the way to go for 6V6s in a Princeton.


If any of you have biased a Princeton and remember the details, please share whether they confirm what I see or contradict.


how concerned should I be with the warm up time exceeding 450V cap can rating?


Also:  thoughts on the 14W JJ 6V6 (and possibily the 14W EH 6V6GTs)?  I had speculated that maybe JJ made a beefier tube to handle the high bias Fender put on Princetons and Deluxes,  however it biases even hotter. Are these really 6V6GT replacements? (the same way one might have asked "are these really 5Y3s?" regarding those early Sovtek 5Y3's didn't really seem to be 5Y3s other than a 2A 5V filament and pinout.. with less internal resistance than a gz34).


PS: I'll add that I put unknown tubes in this tripplet 3403 tester to check for shorts, and take the bad/poor/good ratings with a gain of salt / data point.  in the case of these 6V6s that amp came with, which worked...  they were low on the "good" range compared to all the other NOS & JJ instances.

[mresistor]

I had a 68 come in that was doing the exact same thing.  What I found was a bad .02 coupling cap prior to the phase invertor.  Check all the coupling caps in the signal path.  In my case someone had installed tropical fish and physically damaged one on installation, I could wiggle one lead and the amp would cut out.

[pdf64]

The 12 - 14W anode rating is a red herring, the anodes themselves all have the same dissipation capability. The difference is the rating system used to derive the limit.
My understanding is that all 6V6 probably have a 12W anode dissipation limit under the design centre rating system, and a 14W limit under the design max rating system. The latter was introduced in 1958. Valve info that was published prior to that will usually be noted to be under the design centre system, info published after that will usually be noted to be under the design max system (design max ratings are intended to be derated before use, whereas design centre ratings can be used 'as is').
By definition, all 6V6 anodes, when rated under the same system, will have the same limit.