Cascode gain. Did I do something wrong?

[Serge66]

Greetings All,

Last year, I was playing with the attached cascode circuit. The tubes are 6DJ8 / 6922. According to the theory, the gain is high. However, I could not get a gain higher than 30, which the mu of these tubes. Have I missed something? Theoretically, the gain can reach mu squared (900).
Thanks for your input and help.

Serge

[Serge66]

Hello!!!

[SY]

You'll never see anything like 900. The gain is the product of the gain of the bottom tube times the gain of the top tube. The bottom tube is working into a very low resistance load, so its gain is pretty low. The top tube's load is none too big, either.

Bypassing the cathode resistor (or using LED bias) will pump the gain up a bit, but an ECC88 cascode is more likely to have a gain in the 80-100 range.

[SY]

Let's look at the last example in your table. You've got 4mA of current. The cathode resistor is 810R. Plate resistor is 33k. At that current, the plate resistance is about 5k. But the effective plate resistance of the bottom tube is 5k + (mu + 1)0.810k, or nearly 30k.

Now, the cathode impedance of the upper tube is the total plate resistance divided by mu, or about 1k. The bottom tube is running less than unity gain!

[ErikdeBest]

Hello!!!

Hi Serge

I see that lots of questions asked here are of the 'broken record' type: every now and then they appear again, and well, most of the times the answers are the same. If not a straight answer, a refference where you can get it. In the case of the cascode with a E88CC your best bet is Valve amplifiers by Morgan Jones. Maybe you have it, so you can check pages 94-100. Now I did just a fast scan, and the example gives a gain of about 200 with 1.34mA and 100k plate resistor.

(It seems quite plausible that, although the cathode of the bottom valve is at 2.5V and therefore accepts a signal of 1VRMS without clipping, the output of the stage - which is fed from 285V - won't produce a 200VRMS swing.)

Hope this helps!

Erik

[SY]

Erik, do you get 200 from gm (lower) times RL (upper)? If so, did you consider the effect of the cathode resistor on gm? I think that 200 is correct if it's bypassed.

[Serge66]

Thanks Guys,

I'll buy the Morgan Jones book. Sounds like a good investment.
I'll try the LED bias. I do not really like the idea of bypass caps.

Serge

[Brett]

Originally posted by Serge66
I'll buy the Morgan Jones book. Sounds like a good investment.

Do. I can't recommend it enough. It's $A90 from Evatco but should be less from Amazon or the like.

[ErikdeBest]

Hi Stuart

You are right, the cathode is bypassed! That makes quite a difference, of course!

Erik

[Miles Prower]

Originally posted by Serge66
Last year, I was playing with the attached cascode circuit. The tubes are 6DJ8 / 6922. According to the theory, the gain is high. However, I could not get a gain higher than 30, which the mu of these tubes. Have I missed something? Theoretically, the gain can reach mu squared (900).
Thanks for your input and help.

A gain of u^2 is highly theoretical, and basically a rough estimate. You'll never see anything like that. It's like saying that the u-factor of a 12BY7A is 1100. No way will you ever get close to a gain that high with a single 12BY7A stage. Not gonna happen.

Still, a gain of 30 is pretty pathetic. There are a number of problems with that schemo. V1a operates as a grounded grid stage. You need to lose that R10. You also need to bypass its grid to AC ground (and make sure you're generous with the bypass capacitor) and make sure that you keep the wiring in that part of the circuit as short as possible. The cascode is quite capable of very high gains, and that always means a possibility of instability. Adding impedance to the grid will compromise the gain, and make the thing very liable to RF instability. This is why most UHF triodes designed for grounded grid service (the only way the thing will work at all at extreme frequencies) come in nine-pin miniature pinouts with four or five connections to the grid; these are supposed to be parallel connected to the ground plane to minimize as much as possible any grid-to-ground impedance.

Given that, are you sure it isn't oscillating? An overlooked RF oscillation will also rob AF gain.