Biasing SE class A, fixed bias

[Dave H]

...I'd be looking for something around 375vdc and 2K5 OT which could be done with his 5K OT and half the speaker load on the next higher tap.

That could do it. I’ve never used a KT88 so I don’t know for sure but it looks like it could work at either 450V @ 5k or 350V @ 2k5 and the data sheet is no help as it doesn’t have any SE examples presumably because they didn’t see the point of a low power SE amp when the tube can make 100W push pull. Anyway I don’t think it will be 20W either way as that’s the absolute max for class A from a 40W tube and I only get 4W from an EL84 and 8W from a EL34 so a KT88 would be 12-13W.

[jbefumo]

My mistake -- must be getting dyslectic in my old age -- 474V, and that WAS under load.

Actually, I calculated the impedance for the OT from the formula : = Va^2 / Pa (from The Valve Wizard -Single Ended), which gave me" 474^2 / 44 = 5106.28. Is that an erroneous formula?

[Steve Conner]

To a first approximation, the right bias point for a single ended stage is with an idle current of half the maximum current that flows when Vgk=0.

What I would do is choose the idle current for 75% plate dissipation (old designs often used 100% but new production tubes may not survive long) and then use the tube curves to estimate the required screen voltage that would limit peak current to twice the idle current.

Then I would choose the load impedance equal to plate voltage divided by idle current, maybe a little less to allow for the fact that the plate "bottoms out" around 50-100V. Note that this would give the same answer as Merlin's formula if you started out with 100% plate dissipation. So his formula has the "little less" built in.

Finally after building it, I would adjust the bias voltage to get the idle current I calculated, and everything else should fall into place.

This assumes that you started out knowing what supply voltage you wanted to use. Lower supply voltages allow higher idle currents, so it affects everything.

[jbefumo]

Got it, thanks. So I'm thinking I should start out shooting for around 65mA? My choice of PT was dictated by what I had on hand, so moving forward, I can change that to something more appropriate. The (old) Genelex datasheet lists maximum Ik as 230mA. So, for best tone should I be shooting to get my idle current up higher, say, in the 150mA range by reducing voltage? Thanks again for all the advice!

[jazbo8]

Got it, thanks. So I'm thinking I should start out shooting for around 65mA? My choice of PT was dictated by what I had on hand, so moving forward, I can change that to something more appropriate. The (old) Genelex datasheet lists maximum Ik as 230mA. So, for best tone should I be shooting to get my idle current up higher, say, in the 150mA range by reducing voltage? Thanks again for all the advice!

Found these load line plots and THD vs Load plot on my drive that might be useful. In the plots, Vg2=300V, so you may want to lower it since your plate voltage is higher.

[jbefumo]

Thanks -- I'll examine the plots in a bit. Just got the thing working again - original circuit (i.e., haven't introduced fixed bias yet.) With cathode bias via a 1.5K/5W resistor and 100u cap I'm reading 52.9V at the cathode idling, so that would mean 35mA current flow, right? (Wanted to start at a way safe place and work from there -- don't want to be replacing that tube or that OT!)

[jbefumo]

I'm pretty comfortable with preamp/triode calculations, but am still trying to get a handle on the power tubes, so sorry if I'm being obtuse. . . . Referring to the info at The Valve Wizard -Single Ended, in his OT impedence calculation he says:

Z = Va^2 / Pa
Where:
Va = Anode voltage.
Pa = Maximum anode dissipation.

Since that page is focused on SE topology, and his example is using an EL34, I assumed by his 25W figure that the "Maximum anode dissipation" in the formula referred to the value from the data sheet, and not the ~50% value one would be targeting for an SE application. This is how I came up with the 5k for the OT. Obviously, if I were using 21W instead of 42 watts for the KT88, I would have opted for a ~10K transformer. Am I making sense here?

FWIW, I read 461V at pin 3,, 453V at pin 4, and measure a 3.65V drop across the 1K resistor that connects my G2 to the 'downstream' side of my choke.

[Dave H]

I'm pretty comfortable with preamp/triode calculations, but am still trying to get a handle on the power tubes, so sorry if I'm being obtuse. . . . Referring to the info at The Valve Wizard -Single Ended, in his OT impedence calculation he says:

Z = Va^2 / Pa
Where:
Va = Anode voltage.
Pa = Maximum anode dissipation.

Since that page is focused on SE topology, and his example is using an EL34, I assumed by his 25W figure that the "Maximum anode dissipation" in the formula referred to the value from the data sheet, and not the ~50% value one would be targeting for an SE application. This is how I came up with the 5k for the OT. Obviously, if I were using 21W instead of 42 watts for the KT88, I would have opted for a ~10K transformer. Am I making sense here?

FWIW, I read 461V at pin 3,, 453V at pin 4, and measure a 3.65V drop across the 1K resistor that connects my G2 to the 'downstream' side of my choke.

It’s 42W you want to use in that formula. 21W is the maximum possible power output in class A from a tube dissipating 42W at idle (because the maximum efficiency of class A is 50%) but the tube doesn’t have to be running at 100% of its max rated Pa you can be kind to it and run it at 75% as Steve said which gives An OT impedance of 7k at 472 volts but you have a 5k OT so if you want to bias it in the centre it would have to be at about 80mA or 0.08 x 461 = 37W or 88% of max Pa. The tube should manage 400V peak from a 440V supply and that is 80mA (peak) into 5k. 80mA peak x 400V peak = 32W peak power or 16W average (RMS) into the OT.

[jbefumo]

Got it, thanks! 80-86 mA was what I was thinking from the start. Will stick with 80, since volume really doesn't seem to be an issue. Just in the process of connecting the fixed bias circuitry. Sounds pretty good so far. Will post some clips and the full schematic on my site and a link here in the next few days. Believe it or not, I started this 'project' in July, 2009, with the intention of turning a $100 Crate V18 into a Dumble, and not a clue what I was doing other than cook booking from stuff on the web. Quickly realized that I'd have to bite the bullet and really figure out what I was doing . . . obviously, an ongoing process. Great way to keep dementia at bay!

[jbefumo]

Just stumbled onto this discussion: AX84.com - The Cooperative Tube Guitar Amp Project

Probably a fortuitous link. Going to take a bit of time to digest it, but my initial impression is that my voltage is too high, my transformer impedance too low, and most critically, my screen resistor too small. I'm wondering if, given what I have to work with, if the most efficacious approach might be to just use an EL34 (I have a Mullard reissue on hand), and rebias accordingly? I'm gathering that my assumption that the KT-88 is as brawny as it looks was erroneous.

[jazbo8]

Yes, it may be easier to use an EL34 with your PT and 5K OPT, btw, does the OPT have multiple taps on the secondary?

[jbefumo]

No, this one only has the 8-Ohm tap.

I also have a couple of fat-bottle 6CA7s on hand, as well as an old USA-made 6L6 . . .

So, I'm gathering that if I were to start with a clean sheet, I'd want to shoot for B+ of around 300-330V with a 2500 primary OT? I realized early on that I had bought too high a voltage PT, and figured the KT88 (which I had on hand in my Sundown SD1012C (100W P-P from 2 6550s) HAD to be super robust . . . really hadn't thought through the SE differences or the relatively fragile grid. I'm thinking I may be better off using this PT for a push-pull amp (I have a Hammond 1650H on the shelf), and spring for a more appropriate PT. Of course, now I'm working around the 5000k SE OT . . . costly miscalculations!

Live and learn . . . ;^)

[jazbo8]

Well, too bad about the single-tap OPT, you can still use them even if you are not getting the optimal output power, you may even like the sound because of it, you never know Here is one possible load line with the 5K OPT, Vp at 450V and Vg2 at 200V.

[jbefumo]

Okay, by way of sanity check:

Rather than connecting my grid resistor right after the choke, I moved it to the lowest point in the chain, where I draw for my first preamp stage (not sure of the correct terminology).

The grid resistor is still 1k/5W

My readings are:

Plate voltage: 458V
Vg2 = 206V
Drop across that 1K grid resistor is 3.1V, so I calculate 3.1mA and hence grid power dissipation as: 206V * .0031 = 639mW.
Bias is -17.7V

Cathode current measures 70mA, so output is 16W

So, if I wanted to go back to cathode bias, would a 250R resistor put me at the same point?

Is this all correct?

Thanks again,

Joe

[jbefumo]

Sorry if this is a duplicate -- thought I posted it, but it doesn't appear to have taken...


Okay, by way of sanity check:

Rather than connecting my grid resistor right after the choke, I moved it to the lowest point in the chain, where I draw for my first preamp stage (not sure of the correct terminology).

The grid resistor is still 1k/5W

My readings are:

Plate voltage: 458V
Vg2 = 206V
Drop across that 1K grid resistor is 3.1V, so I calculate 3.1mA and hence grid power dissipation as: 206V * .0031 = 639mW.

Cathode current measures 70mA, so output is 16W

Bias is at -17.7V. Based on this, if I were to go back to cathode bias (hum from the fixed supply, and don't feel like messing with it just now), a 250R cathode resistor would be equivalent, and if I were to go with a 270R, I'd be safe with either the KT88 or an EL34?



Is this all correct?

With the EL34, would there be a benefit to returning to the higher G2 voltage?



Thanks again,

Joe