[OctavianH]

Matched is better! Use the 14.3 and 12.3 together and of course the matched ones at 10.5 and 10.6. It's going to be interesting to see how accurate those measurements are ones you receive your amp

Well, my dilemma was roughly ended because one of the weaker tubes died. So now I have 3. Anyway, I measured the Filament voltage of the tester and seems fine.
If I want to measure in socket the Plate voltage, it is clear to me which one is the Plate (pin 3) but where I can take a ground?

 

[SonicTrance]

If I want to measure in socket the Plate voltage, it is clear to me which one is the Plate (pin 3) but where I can take a ground?

On a 6V6 you should measure plate voltage from pin 3 to pin 8. You always measure tube voltages in relation to the cathode. Grid voltage is grid -> cathode and anode voltage is anode -> cathode. In Infinity and Eternity the cathode is at ground so we don't have to measure at the cathode, just ground.

 

[SonicTrance]

Testing in progress

 

[OctavianH]

On a 6V6 you should measure plate voltage from pin 3 to pin 8. You always measure tube voltages in relation to the cathode. Grid voltage is grid -> cathode and anode voltage is anode -> cathode. In Infinity and Eternity the cathode is at ground so we don't have to measure at the cathode, just ground.

I was thinking the same, I am glad that you confirmed it. I will soon post my measurements because these are raising some questions and might be useful for all of us to look at them and discuss 2-3 cases.

Testing in progress

Wow, this looks impressive. Silver front is a surprise for me. But it looks really nice, a pair of Mullard 6J5 + a pair of Philips EL3N. Cool!
A picture of from the top view would be useful to understand better the proportions and layout. The 6SN7 slot is in front in the center rotated 90 degrees due to lack of space?

 

[SonicTrance]

Wow, this looks impressive. Silver front is a surprise for me. But it looks really nice, a pair of Mullard 6J5 + a pair of Philips EL3N. Cool!
A picture of from the top view would be useful to understand better the proportions and layout. The 6SN7 slot is in front in the center rotated 90 degrees due to lack of space?

Thanks!
More, proper, pics will come later after testing is done. This was just a quicky.

I rotated the 6SN7 socket as I thought it looked better, less cramped. All sockets are rotated the same way though. Only the fastening is different on the 6SN7 socket.

 

[hpamdr]

Testing in progress

Nice lab and nice infinity... Do you have small bias meter on top and bigger vue meter on front ? (or is is just a perspective distortion)

 

[hpamdr]

Which raises the question: What is more important, electrically matchig or strong/weak measurements?
I have the same dilemma with my 4 x Tronix 6V6G, to are stronger and 2 are almost perfectly matched (Vgrid for 200V 20mA are -14.3, -10.5, -12.3, -10.6).
I can choose the "stronger" ones with bigger bias values but not so good matched, or the "weaker" which are very close.

I remember in old ages of ecc88 and el84 amplifier, the tube amp specialist (of the shop i was always visiting) was doing cross matching..
Let say you have a dual triode tube with not matched section, he was in purpose using a stronger output tube in the lesser driver section trying to compensate a bit..

@SonicTrance
I do not know if with Tomas design (excluding Odyssey) this is still adequate ? For Odissey as it is old school design, I would expect same behavior as old amp.

 

[OctavianH]

Ok guys, it seems Friday Beerday transformed in Weekend Measuring, Matching and Beer drinking.

We have a victim, a brave 6V6G who died on the Altar of Science. But we have also some measurements:

I measured output stage for plate 200V / current 20mA. There are 2 results because I try to match my pairs and decide which are properly matched or not.
On input stage I measured for plate 180V / 4mA single triodes as a pair (similar to outputs) and double triodes as T2 and T1 (pin 2 and pin 5 anode).
The values in red are the medium bias which puts both triodes as close as the ideal value as possible. The markings column is not relevant for anyone except me, are some distinctive signs for me to identify the tube in the pair, so ignore it.

And here, we have some intersting results. For example GEC L63 (brown base, straight clear glass) measures bias around -6.7V. But when I measured GEC CV1067 (grey glass, ST shape, older L63) I've seen that the bias goes to -6V. This is problematic because if we want to use a common bias for 6J5, we might go out of range with one of the types. I was not aware that such differences are happening in the same type. For example, the CV1067 at the bias of L63 goes around 2.4mA. Not quite in the tolerance of the input stage:

Another interesting part is related to KT66. The MOV KT66 grey glass (WW2) biases around -18V while 2 pairs of GEC KT66 brown base clear glass go up to around -21V. Deja vu? Similar with the GEC 6J5 grey glass vis the clear glass? Here the same point mentioned before, but since it is output stage and we have a much better tolerance I have not measured one at the bias of other.

Now maybe I measured wrong, or the tubes are not burned in since I've never used them, but the question is, in input stage, if we have such difference between 6J5 and equivalents, what we can do. In my case, I have a special request for Eternity, a switch to use fixed bias vs variable bias, fixed bias being 6J5 / KT66. This switch which I call during week time "Tube rolling off" and during weekend "Beer rolling on" might be problematic if I roll these 2 types of tubes. I am curious if anyone measured at 180V / 4mA any 6J5 and had similar results.

And now, regarding EL32 and EL38, I measured 2 pairs, Mullard EL32 grey glass ST shape and Mullard (rebranded Philips) EL38 black base clear glass:

My results are these ones, but if someone might measure these and share results I would be grateful.

In conclusion, what I marked green are the "electrically matched pairs" at the operation points of Eternity, the prio 1 tubes I will try in the amplifier. What I marked yellow is a decent tube and I will leave to the end the orange. The red will not be used at all. The EL38 will be used since I want to try it, even if the matching is not as good as expected. I made several close measurements to be sure I am on the right path. Now I'll just wait to see what others measure or find and compare with my results. I have somewhere 2 additional Sylvania 6V6G and I might obtain better matching results if I will remember where are those.

 

[SonicTrance]

I remember in old ages of ecc88 and el84 amplifier, the tube amp specialist (of the shop i was always visiting) was doing cross matching..
Let say you have a dual triode tube with not matched section, he was in purpose using a stronger output tube in the lesser driver section trying to compensate a bit..

@SonicTrance
I do not know if with Tomas design (excluding Odyssey) this is still adequate ? For Odissey as it is old school design, I would expect same behavior as old amp.

I'd not recommend compensating a week input tube with a strong output tube. The circuit is most happy with good strong tubes and that goes for all amps. But, especially Infinity and Eternity.

@OctavianH
It's good to see you start measuring! It's fun huh?

I have a special request for Eternity, a switch to use fixed bias vs variable bias, fixed bias being 6J5 / KT66. This switch which I call during week time "Tube rolling off" and during weekend "Beer rolling on" might be problematic if I roll these 2 types of tubes. I am curious if anyone measured at 180V / 4mA any 6J5 and had similar results.

Don't worry about tube differences. There's a whole lot of tolerance to these kinds of circuits.

Here're my bias measurements for Infinity:

Output tubes Infinity 200Va, 40mA Ia
KT66/7581A: -17.3
EL39: -10.0
EL38: -9.0
EL12sp Tesla: -7.8
EL12 spez Telefunken: -6.7
EL3N: -3.9 @ 35mA
EL32: -11.7 @ 35mA
EL12: -6.2

Input tubes Infinity @ 180Va, 4mA Ia
6J5/6C5/6SN7: -6.30
EL11: -7.80
6N7 in 6J5 sockets: -3.65
6N7 in 6SN7 socket: -2.35
E88CC: -4.90
E80CC: -3.80
12AV7: -4.50
6CG7: -6.20
MHLD6: -7.55
6SL7: -765mV @ 3mA

I chose to run EL3N and EL32 @ 35mA to put a little less stress on them. Same thing with the 6SL7 @ 3mA.

 

[OctavianH]

Here're my bias measurements for Infinity:

Output tubes Infinity 200Va, 40mA Ia
KT66/7581A: -17.3
EL39: -10.0
EL38: -9.0
EL12sp Tesla: -7.8
EL12 spez Telefunken: -6.7
EL3N: -3.9 @ 35mA
EL32: -11.7 @ 35mA
EL12: -6.2

Input tubes Infinity @ 180Va, 4mA Ia
6J5/6C5/6SN7: -6.30
EL11: -7.80
6N7 in 6J5 sockets: -3.65
6N7 in 6SN7 socket: -2.35
E88CC: -4.90
E80CC: -3.80
12AV7: -4.50
6CG7: -6.20
MHLD6: -7.55
6SL7: -765mV @ 3mA

I chose to run EL3N and EL32 @ 35mA to put a little less stress on them. Same thing with the 6SL7 @ 3mA.

Many thanks on that! I'll measure for fun EL32 on 35mA and EL38 on 40mA just to compare with your results. For me, in Eternity, output stage will run around 20mA so I guess my measurements are close to what will be. At least I expect so. I am really curious what bias you will find for 6J5 and KT66 on Eternity. My results are problematic, as mentioned above.

I observed that several pairs are "slow starters". This means you need around 4-5 minutes to obtain coherent measure results. For example GE 6550A (KT88 replacement, expected to bias close to KT88 around -24V) provided confusing results, I made a separate note to take again a look. Same was for Genalex KT77 which honestly measured out of range. And even if you might think these are coherent, I have a feeling something is not right there. I will investigate further.

Regarding tolerance, I try to remain in the 3-5mA and 15-24mA as you mentioned.

And yep, measuring is fun. I think beer drinking, sex, tube measuring are on top but if the pandemic goes off, a Black Sabbath concert might change the order. LOL

 

[SonicTrance]

I am really curious what bias you will find for 6J5 and KT66 on Eternity.

The 6J5's will be the same as in Levi's Infinity if I use the same tubes. KT66 is another story of course.

Regarding tolerance, I try to remain in the 3-5mA and 15-24mA as you mentioned.

Yep, that's good.

And yep, measuring is fun. I think beer drinking, sex, tube measuring are on top but if the pandemic goes off, a Black Sabbath concert might change the order. LOL

Haha, yes I agree!

 

[OctavianH]

The 6J5's will be the same as in Levi's Infinity if I use the same tubes. KT66 is another story of course.

According to my measurements -6.3V will make my 6J5 jump over 5mA. But most probably you measured directly on the amp and other factors were involved.

 

[SonicTrance]

I posted pics here: https://www.head-fi.org/threads/oblivion-ultrasonic-studios.902926/post-16236147

Here's a pic from the top

 

[OctavianH]

I posted pics here: https://www.head-fi.org/threads/oblivion-ultrasonic-studios.902926/post-16236147

Here's a pic from the top

It is an incredible beauty. I love it!

 

[hpamdr]

Ok guys, it seems Friday Beerday transformed in Weekend Measuring, Matching and Beer drinking.



We have a victim, a brave 6V6G who died on the Altar of Science. But we have also some measurements:

I measured output stage for plate 200V / current 20mA. There are 2 results because I try to match my pairs and decide which are properly matched or not.
On input stage I measured for plate 180V / 4mA single triodes as a pair (similar to outputs) and double triodes as T2 and T1 (pin 2 and pin 5 anode).
The values in red are the medium bias which puts both triodes as close as the ideal value as possible. The markings column is not relevant for anyone except me, are some distinctive signs for me to identify the tube in the pair, so ignore it.





And here, we have some intersting results. For example GEC L63 (brown base, straight clear glass) measures bias around -6.7V. But when I measured GEC CV1067 (grey glass, ST shape, older L63) I've seen that the bias goes to -6V. This is problematic because if we want to use a common bias for 6J5, we might go out of range with one of the types. I was not aware that such differences are happening in the same type. For example, the CV1067 at the bias of L63 goes around 2.4mA. Not quite in the tolerance of the input stage:



Another interesting part is related to KT66. The MOV KT66 biases around -18V while 2 pairs of GEC KT66 brown base clear glass go up to around -21V. Deja vu? Similar with the GEC 6J5 grey glass vis the clear glass? Here the same point mentioned before, but since it is output stage and we have a much better tolerance I have not measured one at the bias of other.

Now maybe I measured wrong, or the tubes are not burned in since I've never used them, but the question is, in input stage, if we have such difference between 6J5 and equivalents, what we can do. In my case, I have a special request for Eternity, a switch to use fixed bias vs variable bias, fixed bias being 6J5 / KT66. This switch which I call during week time "Tube rolling off" and during weekend "Beer rolling on" might be problematic if I roll these 2 types of tubes. I am curious if anyone measured at 180V / 4mA any 6J5 and had similar results.

And now, regarding EL32 and EL38, I measured 2 pairs, Mullard EL32 grey glass ST shape and Mullard (rebranded Philips) EL38 black base clear glass:



My results are these ones, but if someone might measure these and share results I would be grateful.

In conclusion, what I marked green are the "electrically matched pairs" at the operation points of Eternity, the prio 1 tubes I will try in the amplifier. What I marked yellow is a decent tube and I will leave to the end the orange. The red will not be used at all. The EL38 will be used since I want to try it, even if the matching is not as good as expected. I made several close measurements to be sure I am on the right path. Now I'll just wait to see what others measure or find and compare with my results. I have somewhere 2 additional Sylvania 6V6G and I might obtain better matching results if I will remember where are those.

You did a wonderful job here
What to do with your KT77, heat the tube and wait long, 10 minutes with the one that measure weak and retest....... until measure is stable. (Also verify limits on heater circuit 1.4A is already a bunch ) I use an external heater on my uTracer.

If it is not better, try to get another Kt77 but keep it in case you get another KT77 with same measure you will have to bias like a 6v6 !

Measuring is very cool and gives you an idea of what have been purchased.