Math for Dummies (cathode biased power section...)

[dehughes]

Okay...so, if I have 342v on the plates of my 6V6 pair (Tung Sol reissues), and they have a 250 ohm cathode resistor (RS-5 Dale) bypassed by a 25uf cap, and the cathode voltage is about 20v, then I'm running the tubes way too hard, huh. I've tried doing the math myself (based upon formulas found online, but for some reason I always get fuzzy when I try to figure out how hard I'm running a cathode biased power section. Help is appreciated... I have a few 300 ohm resistors I'm considering putting in there, but I want to get the math straight first.

Gimme 1 ohm resistors to ground on a fixed bias setup and I'm good...but then there's no math in that. If one of you has a handy website or list of formulas that would help, please point me in that direction, as this is basic stuff I want to get down.

Thanks.

[collinsamps]

12 watts per tube @ 38ma per tube based on those numbers and not taking screen current into consideration.

You are running them at the dissipation they are spec'd for, and depending on tube type(GTY, GTA, JJ's etc......not familiar with the tungsol reissues) they can run at 14watts dissipation vice the 12 that you are running them at all day long without a hiccup. Many amps are harder on them than yours with those numbers(BFDR for example).

No worries.

[Ears]

How did he get that? Here is one method:

20V dropped across the 250 Ohm cathode resistor = 80ma.
That is 40ma per tube.
Each tube drops 322V, that is, 342V-20V.
Each tube dissipated (322V x 40ma) watts = 12.8W

This ignores the approx 5mA screen current at idle.
Assuming your load is sensible it seems fine.
Mind you 70% of 14W is 9.8W and that is the conservative benchmark (edit: at least it is in case of fixed bias amps).

[Andy Le Blanc]

The old way R = E/I

In our case its, desired grid bias divided by the total current rating of the valve
equals resistance required

Use this and the tubes characteristic curves, after you've plotted load and the
maximum plate dissipation limit.

[tubeswell]

The New Sensor 'Tung-Sol' RI 6V6GT are spec'd for 12W max dissipation and the 6V6GTAs are spec'd for 14W.

The screen current is only about 2mA tops (not 5mA) in those things.

One of the other posters worked out your dissipation based on 342V at the plate and 20V at the cathode with 250R, with a tube current of 40mA per tube. This would be about 12.2W when you take out 2mA for screen current for each tube.

So they are about bang-on for Class A at idle, and being cathode-biased you don't have any worries about exceeding plate current specs under load. I bet they sound pretty darn good.

[Ears]

<snip>
The screen current is only about 2mA tops (not 5mA) in those things.
<snip>

I suppose it must drop for high plate voltages , GEC 6V6 GTA spec it for class AB1 at Isc idle = 5mA and 4ma for Vp= 250V and 285V respectively (Vsc=Vp)
But that's for two tubes I guess.

[dehughes]

Cool, guys. Thanks so much. I think I'm getting it now...

Even though I could probably leave them running as they are, I think I'm going to tack in that 300 ohm resistor and re-measure, for as things are now I'm getting a bit of "frying pan skillet" noise coming out of the power section....which sounds IDENTICAL to what I hear in my AC30 build (which is running the tubes way over spec....). So, I think though the numbers make sense, I still may be pushing these 6V6s a bit too hard. They do sound good, though.

I'll report back after I swap out some resistors and re-measure things...

[talbany]

Check this out... Plug in the numbers


http://www.webervst.com/tubes1/calcbias.htm


Tony VVT

[solderstain]

Good thread!

Cathode biasing (of power tubes) is a subject that I'm becoming more interested in from a design stand-point. I've done a handful of builds where I've just built an example of someone else's design, but I'm growing past that...

So, all you guys with the large brains (or more years of schooling than me... ):

What's the math to calculate the value of a cathode resistor necessary to arrive at a desired dissipation? Here's a place I can find myself:

A 2xEL84 amp, with 12AX7s in the front end. Building it into a gutted chassis of another manufacturer and reusing the original iron. With a given power supply voltage string and filter structure, I'd know what the plate voltage is likely to be. And it will be lower than what is considered 'normal' for the average EL84 application.

Since I know the tube characteristics, and I have a good idea what plate voltages I'll have, how do I arrive at the value of the cathode resistor?

I'm sure it's common math to most of you, but it's new to me. I don't want to just throw a bunch of different resistors in there until I arrive at my goal. I'd like to have a damn-close ball-park idea of what I need before I heat up the soldering iron.

And.... a similar question about the attending capacitor: how does one arrive that the 'best' value for the cap? Is that pretty-much up to experimentation and what my ears like?

Right now, I'm more interested in the math for calculating the cathode resistor value, though.

THANKS IN ADVANCE!

[dehughes]

Well, I tacked in a 300 ohm resistor on the cathodes of the 6V6s, and measured 22v across it, and 352 on the plates. Things are running a bit cooler, though that's still 35ma/12w per tube, which is kinda hot. I'm considering ordering some 350 ohm resistors to cool things off even more, but I dunno....

As well, I installed a parallel 10k ohm resistor before the plate resistors on both sides of the 6SN7, which raised the plate voltage from 100 to about 110 per triode. That stage sounds MUCH better now...amazing how that changes things. It softened up and warmed up nicely. I'm tempted to raise the voltage there even more...but I'll have to check the tube charts to see what my options are.

What's more, is now I have a hissy problem in the preamp that wasn't there before, so I'm hoping it's just my poor solder joints (what with the tacked in resistor on the 6SN7 stage. We'll see.

Anyway, is there a way to predict how the 6V6s would run should I install a 350 ohm resistor? I'm curious as to if that'd make a significant difference in tone/feel/current draw. I like how it is now (better than with the 250 ohm resistor), but don't want to kill my 6V6s prematurely.

[iknowjohnny]

Speaking of how that changes things, try separate cathode resistors. One for each tube of twice the value you are using how, IE: instead of a single 250R, use two 500 R's, one for each tube. the tone will change a lot doing that too, and IMO for the better depending on the amp and the tube type. Then you can use unbalanced bypass caps like say a 50uf on one and a 100uf on the other and it really changes thing in the same way you're talking about....smoother, more sponge factor, etc. YMMV, but worth a try.

[tubeswell]

<snip>
The screen current is only about 2mA tops (not 5mA) in those things.
<snip>

I suppose it must drop for high plate voltages , GEC 6V6 GTA spec it for class AB1 at Isc idle = 5mA and 4ma for Vp= 250V and 285V respectively (Vsc=Vp)
But that's for two tubes I guess.

I reckon that is just a conservative number new sensor put in there. My experience with them is about 2mA tops. I have tried them in a couple of amps last year (a 5G9 and a 5F2A), but have since gone back to JJ6V6s. The screen support rods are quite piddly in the 'Tung-Sol' RI, and I am now a bit dubious about their longevity - after 9 months or so they started to crap out with the tell-tale static and buzzy noises. Still, they sound okay when they are new. Using a beefier cathode resistor in cathode bias is probably not a silly idea for these tubes.

[thepacifist]

AFAIK, the cathode voltage is to be decided from the datacharts on the basis of desired bias right?

If yes, let me know. I have a excel model for this ready which I can put up here and can be tweaked as per inputs from experts.

pacifist

Good thread!

Cathode biasing (of power tubes) is a subject that I'm becoming more interested in from a design stand-point. I've done a handful of builds where I've just built an example of someone else's design, but I'm growing past that...

So, all you guys with the large brains (or more years of schooling than me... ):

What's the math to calculate the value of a cathode resistor necessary to arrive at a desired dissipation? Here's a place I can find myself:

A 2xEL84 amp, with 12AX7s in the front end. Building it into a gutted chassis of another manufacturer and reusing the original iron. With a given power supply voltage string and filter structure, I'd know what the plate voltage is likely to be. And it will be lower than what is considered 'normal' for the average EL84 application.

Since I know the tube characteristics, and I have a good idea what plate voltages I'll have, how do I arrive at the value of the cathode resistor?

I'm sure it's common math to most of you, but it's new to me. I don't want to just throw a bunch of different resistors in there until I arrive at my goal. I'd like to have a damn-close ball-park idea of what I need before I heat up the soldering iron.

And.... a similar question about the attending capacitor: how does one arrive that the 'best' value for the cap? Is that pretty-much up to experimentation and what my ears like?

Right now, I'm more interested in the math for calculating the cathode resistor value, though.

THANKS IN ADVANCE!

[dehughes]

AFAIK, the cathode voltage is to be decided from the datacharts on the basis of desired bias right?

If yes, let me know. I have a excel model for this ready which I can put up here and can be tweaked as per inputs from experts.

pacifist

Good thread!

Cathode biasing (of power tubes) is a subject that I'm becoming more interested in from a design stand-point. I've done a handful of builds where I've just built an example of someone else's design, but I'm growing past that...

So, all you guys with the large brains (or more years of schooling than me... ):

What's the math to calculate the value of a cathode resistor necessary to arrive at a desired dissipation? Here's a place I can find myself:

A 2xEL84 amp, with 12AX7s in the front end. Building it into a gutted chassis of another manufacturer and reusing the original iron. With a given power supply voltage string and filter structure, I'd know what the plate voltage is likely to be. And it will be lower than what is considered 'normal' for the average EL84 application.

Since I know the tube characteristics, and I have a good idea what plate voltages I'll have, how do I arrive at the value of the cathode resistor?

I'm sure it's common math to most of you, but it's new to me. I don't want to just throw a bunch of different resistors in there until I arrive at my goal. I'd like to have a damn-close ball-park idea of what I need before I heat up the soldering iron.

And.... a similar question about the attending capacitor: how does one arrive that the 'best' value for the cap? Is that pretty-much up to experimentation and what my ears like?

Right now, I'm more interested in the math for calculating the cathode resistor value, though.

THANKS IN ADVANCE!

DUDE, an Excel model would be killer. I used to have something like that, but it won't work with my latest Mac software (I think the older one I had was from '97 or before...).

BTW, I installed a 300 ohm Rk on the 6V6s, and it sounds killer... Turn up the Brown Deluxe side 2/3rds of the way up and it KILLLS, even through my non-broken-in G12H30 (perpetually on the fence with that speaker). It's that "oh my goodness" tone that I'm always shooting for with amps, but so rarely find.

I also installed a 1k dropping resistor before the 100k plate resistors on the 6SN7 side, and got the voltages up to nearly 120 per plate. I'll check that out today when I get home and see how it sounds...couldn't crank it up at 1am.

[solderstain]

DUDE, an Excel model would be killer.

AMEN! Post that sucker!