pultec fixed/cathode bias

does anybody know the meaning of the combined cathode bias and positive fixed bias in the 12au7 tubes of the pultec?

i had an idea what could be going on...

when you have a higher Vg point for high input swings, you might end up with a loadline/Qpoint that has less stellar distortion signatures.

perhaps this technique gives you both?

first cathode bias a high enough grid voltage so you can maintain class a at high input signals.

then put some positive voltage through the grid so that your cathode voltage rises in reference to the plate voltage. and by doing so your platevoltage woking point finds a lower point with lower thd signatures.

I am pretty certain we have discussed this once before. A search might help.

Cheers

Ian

thanks Ian!

so far the only thing on the subject i can find is on CJs thread where he reverse engineered the S-217-D.
https://groupdiy.com/index.php?topic=51046.msg840467#msg840467

5v333 said:

thanks Ian!

so far the only thing on the subject i can find is on CJs thread where he reverse engineered the S-217-D.
https://groupdiy.com/index.php?topic=51046.msg840467#msg840467

Click to expand...

Didn't see anything obvious there. Can you post a schematic for us?

Cheers

Ian

EQP1A

i playd a bit with the 12au7 datasheet and my suggestion/theory didnt work out quite..

a B+ of 290V and a cathode R of 4700R would give a working point of about Vg -15V. i presumed first that it would be about 9V and then "elevated" to 40V.

interesting technique non the less. dont understand how to do calculations though.

take care!!!

Can you post the actual schematic you are looking at.

Cheers

Ian

http://www.proreplicas.com/gfx/pultec-details/SchematicPultecEQP-1A.jpg

It sets the cathode current *exactly*.

Self-bias with cathode resistor may vary 20%.

Here a fixed voltage about 4X higher than self-bias swamps-out variation of self-bias Vgk, and controls cathode current to about 5%.

I would suspect they wanted to reduce unbalanced DC current in the output transformer.

thanks prr!

i have never thought about that possibility!

looks like they wanted a well balanced circuit, see the *  that says select resistors alike?  but what does "alike" mean when you using a 50's  Simpson analog meter?  2% might be  possible with an eagle eye,

how much DC will that transformer take?  the data sheet says it wants zero. what gives?

well, rather unusual for an output, it has an 80 Ni Nickel core, which means that it will start saturating at about 5 K gauss.

we have a 3/8" x 3/4" core with 3150 turns of #36 wrapped around it in the form of two coils,  the spread sheet says that at 22.84 volts RMS into the primary , we will have exactly 5000 Gauss.  this thing is about a 4:1 so max output will be 5.71 volts. 

unbalanced DC will only matter in half of the winding since this is push-pull,

let's try 2 ma of unbalanced DC and see what happens,

half the primary  turns is 1575 turns x .002 = 3.15 amp turns,

magnetic force from DC = 1.257 x amp turns so F = 3.15 x 1 257 = 3.959  Gilberts

wasn't Gilbert a guy on Leave It to Beaver?  i did not know that he designed transformers.

so about 4 Gilberts, how do we get DC flux?  divide by the gap length, but we have no gap.does that mean infinite DC flux?  no, we have air between those L lams,  how much is anybody's guess,

books say that it is probably around .001 or .002, depending on the lam shape and size,

lets pray for 2 mil and covert to metric, 2 x 2.54 = .00508, so DC flux estimate is

4/.000508 = 787 Gauss,  we only have 5000 Gauss to play with, and a lot of that is used up by the AC signal, so you can see why we need to keep the DC down,  if we let it slip to 4 ma DC we have 1.5 K Gauss and this also lowers the perm of the Ni core so we get bass distortion,

this transformer was probably better suited to a Parafeed type circuit since it has a fragile core pertaining to DC. so the Pultec guys must have been aware of this, but they went for the gold and turned out a pretty nice box.

so, the transformer wasnt optimal for PP?
and therefore a circuit with stiff dissipation/balance was developed?!

CJ said:

looks like they wanted a well balanced circuit, see the *  that says select resistors alike?  but what does "alike" mean when you using a 50's  Simpson analog meter?  2% might be  possible with an eagle eye,

how much DC will that transformer take?  the data sheet says it wants zero. what gives?

well, rather unusual for an output, it has an c.

we have a 3/8" x 3/4" core with 3150 turns of #36 wrapped around it in the form of two coils,  the spread sheet says that at 22.84 volts RMS into the primary , we will have exactly 5000 Gauss.  this thing is about a 4:1 so max output will be 5.71 volts. 

unbalanced DC will only matter in half of the winding since this is push-pull,

let's try 2 ma of unbalanced DC and see what happens,

half the primary  turns is 1575 turns x .002 = 3.15 amp turns,

magnetic force from DC = 1.257 x amp turns so F = 3.15 x 1 257 = 3.959  Gilberts

wasn't Gilbert a guy on Leave It to Beaver?  i did not know that he designed transformers.

so about 4 Gilberts, how do we get DC flux?  divide by the gap length, but we have no gap.does that mean infinite DC flux?  no, we have air between those L lams,  how much is anybody's guess,

books say that it is probably around .001 or .002, depending on the lam shape and size,

lets pray for 2 mil and covert to metric, 2 x 2.54 = .00508, so DC flux estimate is

4/.000508 = 787 Gauss,  we only have 5000 Gauss to play with, and a lot of that is used up by the AC signal, so you can see why we need to keep the DC down,  if we let it slip to 4 ma DC we have 1.5 K Gauss and this also lowers the perm of the Ni core so we get bass distortion,

this transformer was probably better suited to a Parafeed type circuit since it has a fragile core pertaining to DC. so the Pultec guys must have been aware of this, but they went for the gold and turned out a pretty nice box.

Click to expand...

CJ how did you determined  "well, rather unusual for an output, it has an 80 Ni Nickel core, which means that it will start saturating at about 5 K gauss.".  Permalloy or similar Ni78's have typical  7000-8000 KGauss. Right?  I really doubt it's made of Ni78 (80) because Peerless stated +21dbm max level. It looks like Ni50. Right?

i believe it is a custom blend made by Arnold Engineering, Mike over at Magnequest has a few pounds of the 12-L original stuff but he does not want to sell any.

you are probably right, Opacheco has a real one, maybe he can help figure it out by doing some tests.

the pics he put up make it look like something with a lot of perm,  21 dBm would mean 36 volts on the primary, which would saturate 80 Ni, but that 21dBm figure might be inflated, would have to talk to a Pultec owner to see what that transformer will do,

5v333 said:

does anybody know the meaning of the combined cathode bias and positive fixed bias in the 12au7 tubes of the pultec?

Click to expand...

A combination of fixed and cathode bias gives excellent control of the operating point against tube ageing/dispersion.
That's exactly the technique used with transistors, in order to accomodate large manufacturing variations and significant temperature drift.

CJ said:

i believe it is a custom blend made by Arnold Engineering, Mike over at Magnequest has a few pounds of the 12-L original stuff but he does not want to sell any.

you are probably right, Opacheco has a real one, maybe he can help figure it out by doing some tests.

the pics he put up make it look like something with a lot of perm,  21 dBm would mean 36 volts on the primary, which would saturate 80 Ni, but that 21dBm figure might be inflated, would have to talk to a Pultec owner to see what that transformer will do,

Click to expand...

Yep, 36volts. If it's true, than it must be some "lamination mixture" or maybe Ni60 what is obsolette alloy long time ago. If I can have just one lam i can do spectral analasys... bu i have no 
Edit: also, I still dont understand does transformer core " see " complete tube current (i think something around 8ma per tube) or just the tubes  current difference?  Generally it should be the seccond one but I'm  not sure. Can someone confirm this?
Btw, Opacheco, are you around to do some test?

My ASSumption, which may not answer your question as you pose it, is

1) the transformer sees whatever heat is dissipated in the series windings, minimal, but you can see how someone putting a 6L6 on one might cook it quickly, so gauge must be sized for any potential current flow. 

2) then assuming that is within acceptable bounds, with balance the current doesn't matter, with imbalance distortion goes up and low frequency response goes down, as those parameters are designed with zero current in mind.  With perfect balance DC current counts as zero, so long as 1) above is observed.