Build thread all octal (6SJ7+6SK7) 5F2A with 6L6

[Jerry garrcia]

Higher value cathode bypass capacitor increase gain because it decrease local feedback and cathode does not follow drive signal too much which would decrease actual drive which there is needed between cathode and control grid.

Without drawing loadline I am quessing that perhaps you have biased that pentode driver tube too hot when its anode voltage is less than 100V and it limit its gain? Sometimes tube datasheets have tables for several voltages and anode and cathode resistors and what the gain and output impedance then come.

The first gain stage (6SJ7) is the pentode and biased it according to RCA charts and the same as previous builds.

The triode wired pentode (6SK7) was the one with a rather cold Bias I think.
My main problem now is the 6SK7. Will do more reading. I think I should just increase the cathode resistor a bit more. There is a 4.7k now with a Rp 150K. Ether lower Rp to 100K or further increase the Rk will probably get the Vp>100.

I would like it to have the desired output without the bypass cap. On the finishing stretch.

Edit. I can’t decide what to do with the driver 6SK7. Ether go for a lower Rp like 33-47K and a Rk of 1.2-2.2K or Rp of 100-150K with a Rk of >4.6K. Probably the latter. Will have to try and listen.
Or accepting that I have to bite the dust and use 6J5 with a known characteristic.

 

[2L man]

I think yuu should at least test cathode bypass capacitor for 6SK7. 4,7k cathode resistance is quite high and local NFB comes higher.

I have not yet used, only test, but it is possible to use LED on cathode to bias tube. Obviously here few series? I don't know 6SK7 but its 7,1V bias is high against typical triodes. Using those voltages and resistances anode current comes 1,8mA and cathode current 1,5mA when they should be the same when screen is tied to anode. Perhaps because of resistor tolerances or control grid "leaks"?

 

[Jerry garrcia]

I think yuu should at least test cathode bypass capacitor for 6SK7. 4,7k cathode resistance is quite high and local NFB comes higher.

I have not yet used, only test, but it is possible to use LED on cathode to bias tube. Obviously here few series? I don't know 6SK7 but its 7,1V bias is high against typical triodes. Using those voltages and resistances anode current comes 1,8mA and cathode current 1,5mA when they should be the same when screen is tied to anode. Perhaps because of resistor tolerances or control grid "leaks"?

Will try that. I thought that the difference was quite strange as well but just thought that my calculations were wrong.

 

[2L man]

There flow 4,8mA to B+3 but about 1,8mA thru 6SK7 and it looks like 6SJ7 anode resistor is 200k something so its anode current is lower. Perhaps B+3 15uF electrolyt leaks? Or it comes from resistor tolerance? If electrolyt leaks perhaps it won't regulate B+3 well enough and it screws tube amplification?

Voltage loss B+1 to B+2 12k resistance make current 3mA which two tubes about use and choke resistance is out of that calculance which will drop actual current bit more.

 

[Jerry garrcia]

Or it comes from resistor tolerance?

I have developed this bad habit of checking all resistors before they get tackled in due to walked over a few mines before.

I will definitely look in to the B3 filter cap. After spending half the night reading on different low mju triodes I think the cathode resistor needs to get up or lowering the plate load.
What I should have done from the beginning was to (on breadboard with pots) decrease the anode resistor while increasing the cathode resistor and do plot data points along on a chart.

 

[Jerry garrcia]

I guess I’m done…
Really like the sound out of the amp. Tried it with an archtop, Frankencaster and a Wilshire with mini humbuckers. Good clean output up to 11 when in cathode bias. When switching to grid leak bias a different amp and much “hotter.
What I’ve done since last:
- decreased Rp from 150K to 100K on the 6SK7.
- followed @2L man (thanks for all help. You thought me a lot) tips and added a 22uF bypass cap on the 6SK7.
- Added a 470K grid leak resistor on the 6L6.
- changed the master volume pot from a 500K to a 1M log pot.
- added a 1.5K grid stopper on the 6L6
- removed the 220uF bypass cap on the 6L6 (did also try a 50 and a 100uF) but to much background noice and the amp became to “gainy”.

- will need to change the switch to the NFB. It’s not a total off switch since the off function showed a 68K resistance and not a total block between the lugs.

Final measurements:
6SJ7
Vp 94V
Vg2 46V (might decrease the Rg2)
Vk 1.6V

6SK7
Vp 114V
Vk 10.4V

6L6
Vp 455V
Vg2 390V
Vk 30V

Ia 60.1mA
Plate dissipation 90%

I’ll include two videos. First without bypass cap on the power tube and second with a 100uF. No NFB in neither videos.

 

[Jerry garrcia]

Update. Been stuck with two other builds so haven’t got time to do the final tweaks. I like the tone but it lacks the push.
Thinking of redoing the preamp and drive circuit. I can’t get the 6SK7 do deliver enough to the 6L6. Don’t know if it’s my lack of knowledge or that it’s the wrong tube for the job.

So my two ideas I’m considering is either to change the triode strapped 6SK7 to a 6J5 or a 6SQ7, or
putting a 6SQ7 up front and using the 6SJ7 as the driver?

 

[2L man]

For us Europeans russian 6SN7 and 6SL7 substitutes (I don't remember their type#) are often good buys so I have not tried to use those unususal pre amp tubes like 6SQ7 and 6SK7.

 

[Jerry garrcia]

For us Europeans russian 6SN7 and 6SL7 substitutes (I don't remember their type#) are often good buys so I have not tried to use those unususal pre amp tubes like 6SQ7 and 6SK7.

I do have both old 6SN7 and 6SL7’s at home but hate to leave one side unused. The 6SQ7 is like a half of a 6SL7. Both with gain 70. The one used in the EH-150/185. So that might be better than the 6J5 since the gain of that is around 20.
So would the pentode be a better driver than the triode (Merlin seems to think that) but if so, will I loose the glorious 6SJ7 tone?

 

[NTC]

Can you measure the gain at each stage? Also, what is the cathode voltage on the 6L6?

I would use the 6SQ7 or 1/2 a 6SL7 in the driver position in place of the 6SK7. A 6SQ7 has a gain of 100, not 70 like 6SL7.

 

[Jerry garrcia]

Can you measure the gain at each stage? Also, what is the cathode voltage on the 6L6?

I would use the 6SQ7 or 1/2 a 6SL7 in the driver position in place of the 6SK7. A 6SQ7 has a gain of 100, not 70 like 6SL7.

I don’t know what I got 70 from. Possibly from some forum or just a mistake. I can’t measure the gain, I think. Will read up on that.
I agree. Deep in my sole I feel
that the pentode should be up front.

 

[Jerry garrcia]

While doing some cross country skiing today I had time to think a bit.
Got a tip from a guy who has built a few SE 6L6 amps with a high gain pentode up front.
He suggested to abandon the loved triode strapped 6SK7 and use the following setup:
6SJ7 - gain pot - 1/2 6SN7 - tone stack - Master volume - 2/2 6SN7 - 6L6.
Seems worth a try since he’s been using it with success and a lot of hifi guys uses the 6SN7 to drive tone stacks like the 12au7. They also appears to be good with math and aims to optimise a circuit .

Been reading up on the 6SN7 and different circuits. It seems like in hifi the common setup is Rp 47K and Rk 860R with B+300V. Is that for maximise the delivered current to the tone stack?
The Center point on the loadline would be about a Rp 100K and Rk 1.8K.
Would that seem like a good setup for the two halves of the 6SN7? The first for the tone stack and the higher gain setup to drive the 6L6? Need to learn more about driver tubes. Think I can squeeze in another Rp, Rk and coupling cap on the turret board if I mount some stuff on the socket and board.

 

[Jerry garrcia]

So now all redone. Skipped the 6SK7 and the 6SQ7 driver. Added a 6SN7 as tone stack driver and then a second gain stage. The driver stage Rp 33K and Rk 860R. Gain stage Rp 100K and Rk 1.5K.

Since the eyelet/turret board was made for only one triode as a driver and no second gain stage it’s a bit cramped up in there. The two 6SN7 plate load resistors are on top of each other…
6SJ7
Vp 86
Vs 41
Vk 1.45

6SN7 1/2
Vp 144
Vk 4.4

6SN7 2/2
Vp 91
Vk 3.3

6L6
Vp 415
Vs 386
Vk 26

Plate dissipation 105%. Might have to increase the cathode resistor of the 6L6
it is LOUD and clean. Possible to get it to break up when adjusting the gain pot. So happy!
@2L man

Edit: wrong on the schematic. The bypass cap on the 6SN7 is a 1uF