Individual Fixed Bias Supplies?

[jbefumo]

Tried searching the forum (and google) and came up empty, so hope I didn't miss the obvious:


I'm building a Tweed OD on the carcass of an old Sundown that was originally set up to make 100W from a pair of 6550s (since replaced with KT88s). Was hoping to set it up with cathode biasing, but think the combination of my existing PT (B+ = 497V) and OT (which I'm assuming to be around 4.5k primary) would not be optimal (as the data sheet specifies 9k primary for cathode biasing).


Anyway, if I'm going to go that route, I'd like to have individual bias control for each tube.  There are plenty of circuits out there, but nothing looks quite right to me (which is to say, doesn't work in MultiSim.)


(Oh yeah, fwiw, my PT has no center tap and no separate bias tap, so I'm pretty much stuck with a FWB and that B+ voltage.)


Datasheet specifies Vg1 as ~ -34V, while the original schematic for the Sundown notes the bias as -56V.


I've come up with the attached schematic mainly by starting with the original Sundown circuit and futzing around with values until I attained a range of -29V to -64V.


Ultimately I'll probably replace the KT88s with KT66s for a more manageable power level, but for now, just want to get it working with what I have on hand.


Anyway, any observations, critiques, suggestions, etc, will be greatly appreciated.


Joe

[kagliostro]

If your OT has 4 and 8 ohm secondaries

connect an 8 ohm speaker to the 4 ohm secondary

you will see the primary impedance rise from 4.5k to 9k

Franco

[shooter]

fwiw, never used it, but it's in my *all things amp folder*

[Fresh_Start]

I would want resistors isolating the two bias supplies regardless of pot adjustment. Separate, second filter caps for each supply. Don't want to use wiper for bias supply if possible so you have a bias Voltage (big negative preferably) even if the pot's wiper fails somehow.

Drawing in Post 3 here looks like a reasonable starting point:
http://music-electronics-forum.com/t5415/


I'd make that first cap 100uf instead of 47uf, but a Pi filter with two 47uf caps bridged by 15K resistor does a fairly good job of filtering.


Chip

[Willabe]

I'd make that first cap 100uf instead of 47uf, but a Pi filter with two 47uf caps bridged by 15K resistor does a fairly good job of filtering.

There's a time constant formed by the cap(s) and the R. If too long (bias filter caps too large) the power tubes will be stressed or even start to red plate before the caps are charged up and the tubes get their full - bias dcv.

[PRR]

> Datasheet specifies Vg1 as ~ -34V, while the original schematic for the Sundown notes the bias as -56V.

At what screen voltage??

I suspect datasheet likes 300V on screens. I suspect Sundown thought that was too much trouble, took screens to near 497V, and naturally needed at least 497/300 or 1.66 times the control grid bias. 34*1.66 is dang near 56V.

Unless you re-think the screen supply, I'm sure you will be needing more like 56V than 34V.

A sweet trick to get a 60V AC winding (gives ~~84V DC, which can be trimmed as needed) is to work a 120V:12V transformer backward across the 6V winding.

[ernie_jr]

PRR, glad you still like my trick, just did 2 more amps using this as the bias supply. Bought several of those trannies from radio shack when they were having the going out of business sales.

[Jack_Hester]

Here are a couple that I did.

Jack

[jbefumo]

Actually, it has 8 and 16-ohm secondaries, but given that, wouldn't the procedure be to connect the 8-ohm output to a 16-ohm load?  I have on hand a NOS 15" 8-ohm Altec Lansing alnico speaker that is hands-down the best transducer I've ever heard, and am kind of fixated on using that (but could certainly find some other home for it.)  If what you suggest is correct, then connecting the 16-ohm output to the 8-ohm speaker/load would do the same thing, right? But somehow that just doesn't seem intuitive to me.

If your OT has 4 and 8 ohm secondaries

connect an 8 ohm speaker to the 4 ohm secondary

you will see the primary impedance rise from 4.5k to 9k

Franco

[jbefumo]

Thank's much -- downloaded and will digest shortly.


While on the topic of driving thumbtacks with a sledgehammer -- is there any value in using one of those high (negative) voltage LDO regulators to present a rock solid bias voltage, or is that just asking for problems and adding unnecessary complexity?

Here are a couple that I did.

Jack

[jbefumo]

Ahh, never thought about the separate transformer approach, and since little current is needed, wouldn't have to be big -- THANKS!
 

> Datasheet specifies Vg1 as ~ -34V, while the original schematic for the Sundown notes the bias as -56V.

At what screen voltage??

I suspect datasheet likes 300V on screens. I suspect Sundown thought that was too much trouble, took screens to near 497V, and naturally needed at least 497/300 or 1.66 times the control grid bias. 34*1.66 is dang near 56V.

Unless you re-think the screen supply, I'm sure you will be needing more like 56V than 34V.

A sweet trick to get a 60V AC winding (gives ~~84V DC, which can be trimmed as needed) is to work a 120V:12V transformer backward across the 6V winding.

[jbefumo]

Yet another data point in my knowledge base -- thanks!

I'd make that first cap 100uf instead of 47uf, but a Pi filter with two 47uf caps bridged by 15K resistor does a fairly good job of filtering.

There's a time constant formed by the cap(s) and the R. If too long (bias filter caps too large) the power tubes will be stressed or even start to red plate before the caps are charged up and the tubes get their full - bias dcv.

[jbefumo]

Being one of those "if some is good, more is better" mentalities, this is a most useful observation.  Glad I didn't grab those 20,000u capacitors from my junk bin in order to make it SUPER STEADY!

I'd make that first cap 100uf instead of 47uf, but a Pi filter with two 47uf caps bridged by 15K resistor does a fairly good job of filtering.

There's a time constant formed by the cap(s) and the R. If too long (bias filter caps too large) the power tubes will be stressed or even start to red plate before the caps are charged up and the tubes get their full - bias dcv.