Thinking of setting up Hybrid cascoce using this for a negative supply.

[SoulFetish]

In order to increase the gain of my input stage, I'm experimenting with using a jfet input to drive the cathode of V1 in a cascode. I was initially going to set the grid of the upper triode using a voltage divider from the B+, but I had an idea!
I can ground the grid at 0V, and use the zener I've already installed to drop the B+ as my negative supply using "back biasing". This is the set up. Lemme know what you think.





edit: do I need to capacitively couple the input?

[Helmholtz]

With load connected the voltage at the second PS node will be lower by the zener voltage + resistor drop.

I wouldn't like to have different grounds.

Also the voltage across the zener in your arrangement is likely to have some ripple which would be amplified by the input stage.

Why don't you just supply the grid with +8.2V from a zener regulated supply (current limiting resistor + zener with parallel filter cap)?

[SoulFetish]

With load connected the voltage at the second PS node will be lower by the zener voltage + resistor drop.

No it won't. I should have drawn some more components to show how the connections are made, but let's assume that we are using a star ground for simplicity; All grounds are connected to the cathode of the Zener Diode. Aiken shows a bleeder connected to this point to maintain a load current at all times.

I wouldn't like to have different grounds.

I'm a little sketchy about it as well, but that's why were talking it over.

Also the voltage across the zener in your arrangement is likely to have some ripple which would be amplified by the input stage.

Aiken's simulation shows remarkable regulation. I would have suspected a large ripple component.

Back Biasing (see improved zener model)

Why don't you just supply the grid with +8.2V from a zener regulated supply (current limiting resistor + zener with parallel filter cap)?

Well, I already set up the circuit using a resistor divider of 1M(R1) & 91k(R2). So, I'm already using the traditional power supply set up.

[Helmholtz]

No it won't.

You're right. Was confused by different grounds, my bad.
Actually all supply voltages for the rest of the amp will be lowered by the zener voltage.

Aiken's simulation shows remarkable regulation. I would have suspected a large ripple component.

Even if the the residual ripple (and zener noise) across the zener is in the mV range, it might be too much to apply to an input of the first amp stage.
It seems that Aiken's proposals are meant for biasing a power stage having a typical gain below one.
But in your case the available gain after the first stage might quite large.

[SoulFetish]

I just tested it using a resistor divider from the B+ at the first stage node. Powered it on monitoring the voltage at the upper tube plate so I can adjust the jfet source resistance to bias the stage. Total failure....
Couldn't get any voltage drop across the plate load at any range of the 300R source pot. Thought I toasted my LSK170s, but it turned out to be the upper tube wasn't working. So I swapped in another tube, worked like a charm. Running a signal into the front end and biasing the upper tube anode at 63V, it looked like the input might be a little too sensitive, so I backed it off to 80V and play tested it.
My initial thoughts are that it's incredibly quiet. Really quiet. I need to adjust the tone stack or dump some of the HF at the input (not unexpected considering), and I might bias it back to 65V. The supply voltage will end up being around 105V-100V and I'm using a 9k1 load resistor on the plate. You have any suggestion on how I can improve the stage?

[Helmholtz]

Improve in what respect?

I understand the stage is low noise and seems to have lots of gain (what actual gain do you measure?).
What is missing?

[SoulFetish]

So, I decided to scrap the hybrid cascode on the input. Gate sensitivity for the LSK170 is two much, and it's to easy to clip the input. I loved how quiet the amp was using this circuit, but I think I like this for other purposes. For example, it makes a great reverb recovery amplifier.
I just redesigned the input for a traditional tube cascode setup using (2) 7586s. The great thing about these tubes is the high transconductance while operating at low plate voltages. The data sheets quote a gm of 7000 µmho, and gain of 31 with a plate voltage of 26.5V. This frees me up from having to modify the existing power supply, and I can work with the +105V available at that node. So, these should work out quite nicely.
Here is the scematic showing the upper tube with a grid leak bias (with an alternative fixed bias option for the screen using a voltage divider, show in blue.)



One question I had was whether a grid leak bias was the better option for an input stage? Or if it was better suited in a stage where drive levels where larger? I suppose it's no big thing to try both and see which one I think sounds better.
Anyways, too be continued....


edit: corrected the label for tube type in the schematic.

[mozz]

Think this would work with 6dj8/ecc88, with changes to plate resistor 300v supply? I have a bunch and been waiting on a schematic, other forum a guy use to brag about using 6dj8 in cascode but never posted a schematic. Seems rather simple and basic.

[SoulFetish]

Think this would work with 6dj8/ecc88, with changes to plate resistor 300v supply? I have a bunch and been waiting on a schematic, other forum a guy use to brag about using 6dj8 in cascode but never posted a schematic. Seems rather simple and basic.

6DJ8s were developed specifically to operate as a cascode! It would definitely work, in that you would want to make the appropriate changes to plate load/grid voltage/bias/etc,
If you wanna post a data sheet, we can talk about setting em up

[Helmholtz]

I just redesigned the input for a traditional tube cascode setup using (2) 7586s.

Circuit says 7589s?

[SoulFetish]

Circuit says 7589s?

]

typo. I fixed it in the illustrator file and thought I had replaced the JPEG.
Ill fix it a little later on.

[Helmholtz]

No wonder I couldn't find a 7589 datasheet.

How did you come up with the 1M value for the grid leak bias resistor?

[SoulFetish]

How did you come up with the 1M value for the grid leak bias resistor?

1M was chosen following Blencowes instructions in chapter on Cascodes in Designing Valve Preamps for Guitar and Bass, where he goes into grid leak biasing the upper triode. The interesting thing about the 1M grid leak resistor value is that at idle, the upper triode is zero biased. He explains it as follows:

"At quiescence, the upper triode is zero-biased of course, but when we input a signal a charging current flows into the capacitor via the grid-leak, which lowers the voltage on the screen grid. The larger the signal, the greater the bias developed, which creates a subtle compression effect on sustained passages, rather like power-supply sag. This method is simple and works well for guitar. "

Also, because the cathode of V2 is tied to the anode of V1, as the input signal swings more positive, the anode of V1 drops along with the cathode of V2 causing the grid of V2 to go positive with respect to the V2 cathode and draw grid current. The grid current will bleed of some of the charge of the bypass capacitor with every half cycle of audio signal, dropping screen voltage closer to the design values. The larger the signal the lower the screen voltage. This causes the compression effect assosiated with this arrangement.

So, I tested it out tonight. Everything measured nearly identical to what I had mathematically expected (how often does that happen on the first shot?). I'll get into the sound and response of the amplifier more tomorrow.
gonna get some sleep..

[tedmich]

oops misread title, expected an AI melding of these two

[SoulFetish]

okay, I'm back. Here's my initial thoughts on the grid leak biased cascode in the front end. First, it blasts the hell out of the front end of the amp with gain. Too much gain actually, and that's with the stage loaded by a tone stack and volume. I think you definitely get a "squishy' type of feel in the overdrive characteristics. Is reminiscient of having a super hot fuzz face in front of the amp which doesn't really clean up with the guitar volume until you reach the roll it back to the first 10-15% of the rotation. This might be just the thing that some players might like. It really has that feel and sound characteristic.
I am not looking for that in the normal operation of my amp, so I'm going to adjust the upper grid voltage to a fixed 40V using a divider from the B+. This should increase the input sensitivity to 3V (p-p) Also, I'll experiment with the load resistance from 9k1-12k, which should provide a gain in the range of 63 (just above a nominal center biased 12AX7/100k load) - around 84. Without testing it, I figure I'm probably looking for a gain of around 70 or so.


ps. tedmich, I don't know if the world is ready yet for a hybrid Cass-Croce.


- here is a schematic reference: