Bias Resistor Calculation

[lkrasner]

I'm working on calculating cathode bias resistor values for various power tubes I'd like my amp to be able to take. I would like someone check my calculations here. Now, I'm somewhat ignoring the fact that the B+ will change a bit with the different tubes due to the increased current draw. I figure I should be fine with a bias point anywhere between about 80% and 95%, so I'm aiming for a bit over 90, assuming it will actually be a bit lower when the voltage drops for the bigger tubes.

The preamp is contains 2 12AX7s (3 gain stages and a cathode follower) and 1 12AU7 (FX loop, cathode follower and gain stage). I believe this should add up to 15-25mA (not that important for this step).

The power amp is an EL84 in parallel with a 6V6 OR a single bigger octal.

I'm using a PT rated at 225-0-225 and 150 mA. So, pencil in 20mA for the preamp and 80mA for the poweramp, the total current is 100mA. the rectified B+ should be somewhere between 1.4*225=315 and 1.3*225 = 293. Maybe a bit more due to a high mains voltage on the primary. Dropped across a 100 ohm resistor, this should become about 280-305VDC. Let's call that 300 for the rest of the calculations.

Now for the tube bias calculations. Here's what I got from looking at the max dissipation and transfer characteristics and trying to make an educated guess.

Code: [Select]

EL84:
Ia = 35mA
Rk = 180R

6V6:
Ia = 40mA
Rk = 390R

6L6GC:
Ia = 55mA
Rk = 380R

EL34 / KT66:
Ia = 72mA
Rk = 360R

KT88 (I couldn't find transfer characteristics anywhere):
Ia = 95mA
Rk = 160R
How do these look? The 6V6, 6L6, EL34 values seem very close to me, if I'm right about them, is it even worth putting a switch posistion for these, or should I just settle for something that works for everything?

[shooter]

When I do cathode biasing I start with someone else's design that has the tubes I wanna use, that way I know I'm close and error to the cold side.  I don't solder in my R's, just gator-clip them.  2 meters hooked up, 1 across Rk, the other on plate, power up n wait til the filaments are hot and volts stabilize.  Power down, do the math, adjust Rk til I'm there.  Then I tack it down, power up and let everything cook in for an hr.  Make sure you account for G2's current and the volt drop across Rk when calculating Max plate pwr.  after about a week of daily use I go back and check everything again, usually with a dummy load and check 0 signal and Max distorted signal.  Then if all is good, Permanently solder Rk and just play it til is smokes

[lkrasner]

When I do cathode biasing I start with someone else's design that has the tubes I wanna use, that way I know I'm close and error to the cold side.  I don't solder in my R's, just gator-clip them.  2 meters hooked up, 1 across Rk, the other on plate, power up n wait til the filaments are hot and volts stabilize.  Power down, do the math, adjust Rk til I'm there.  Then I tack it down, power up and let everything cook in for an hr.  Make sure you account for G2's current and the volt drop across Rk when calculating Max plate pwr.  after about a week of daily use I go back and check everything again, usually with a dummy load and check 0 signal and Max distorted signal.  Then if all is good, Permanently solder Rk and just play it til is smokes

That works well if you have a big supply of power resistors. I keep a bunch of 1W metal films around that work great for most things, but cathode resistors generally need to be 3W-5W  I believe, which are harder to find in decent sets.

Maybe I'll just order a couple values around each so I can test.

[shooter]

I do mostly cathode biased so I have 2ea of the commons, 10w starting at 150 up to 470, then when I order for a build I just buy whatever I settled on n the test ones go back in the box. 

[PRR]

Why are you aiming 6L6GC current lower than EL34?

6L6GC has a higher rated Pdiss than EL34.

Do not exact-copy the 6L6GC datasheet values. Most of these were un-changed from the original 6L6 sheet, and that tube was first rated 21W then hastily re-rated for 19W. The 300V 4K5 load condition cooks the tube at 15 Watts, which a 6V6 can do (for a while).

(Today I suspect both 6L6GC and EL34 may be the same plate-stamping; historically EL34 was worked well past the "25W" number, into 6L6GC turf, and why make different parts?)

Load impedance, divided by Mu(g2), then times 0.6, should be a fair trial cathode resistor.

This may melt or freeze. Adjust B+ or Rl until in ballpark.

Since Mu(g2) is "10" for 6V6, 6L6(GC), EL34, and most large audio-market bottles generally, once you nail your OT, then you know your rough Rk (but have to compute safe B+). EL84 Mu(g2) is 18, so it will be different.

[shooter]

Mu(g2),

I'm ignorant to Mu(g2) Prr, I get the .6 and load impedance

[HotBluePlates]

Mu(g2),

I'm ignorant to Mu(g2) Prr, I get the .6 and load impedance

Check out the 6L6GC data sheet. As PRR said, conditions listed were repeated from very early sheets, but look at the top of page 2 for the Triode Operating Condition for the tube. It lists an amplification factor of 8.

Amplification factor in this case is the Mu from control grid (G1) to screen grid (G2), which in triode operation of the 6L6 is connected to the plate. The amplification factor you're used to seeing for actual triodes is the amplification from grid to plate; here it is the amplification from G1 to G2. In other words, a change of control grid voltage is 8 times more effective at impacting plate current than a change of screen grid voltage.

Say the screen of the 6L6 is at 320v above the cathode voltage. 320v/8 = 40v, so if you biased the control grid to 40v below cathode voltage (by applying fixed bias of -40v to G1 or elevating the cathode to 40v in cathode bias), plate current should be nearly shut-off.

Theoretically, Class A would idle halfway to cutoff, or a bias of 1/2 * Screen Volts / Mu_g1-g2. But things get squirrely at low plate current, and typically the tube gets biased off the mathematical half-way point, hence the 0.6 in the equation.

PRR took it a step further and noted the cathode bias resistor has a similar relation to the plate load.

[shooter]

Thanks HPB and PRR, I did some surfing and reading, pretty soon I may know half of what everyone else has forgot about tubes!

[jjasilli]

Note that Weber sells:

RHEO500-25

$8.00

500 ohms, 25 watt   rheostat used for experimenting with adjustable cathode bias.

https://taweber.powweb.com/store/resord.htm