Cascoding tubes

[Joshua_G]

HINT: If you want the tubes in a SRPP to last more than a few months, you must provide two seperate heater supplies. One for the lower triode, and one for the upper, with the upper heater line raised above gnd some 10-30V HIGHER than the cathode of this second tube will be at. This is to advoid heater cathode breakdown.

Regards, Allen

Thanks.
What happens when the filament supply is DC, not connected to the ground?

 

[Bandersnatch]

Joshua,

You can view a cascode as a "synthetic" pentode. Like a pentode it offers greatly reduced Miller capacitance and large single stage gain. It too has a very high O/P impedance. Unlike a pentode, cascoded triodes don't introduce partition noise. The PSRR of cascodes is WRETCHED, which makes regulated B+ a must.

Cascodes that function well employ triode's with high gm, at the bottom of the stack. Gain in cascodes is roughly = (gm_{Lower Device}) (net Z_L).

BTW, the example you provided is a SRPP, not a cascode. Both arrangements are examples of "Totem Poles".

hey-Hey!!!,
As Eli said, the circuit characteristics are for the most part based on the bottom triode. There are a few important characteristics that depend strongly on the upper element. Ability to swing close to its control grid voltage being the most important; this is a pentode that can't come any where near driving its plate below its 'g2'. For the upper element a MOSFET is about the best IME. It can swing the composite's 'plate' very near its g2 voltage, and presents the nearest-to-vertical load line to the triode underneath.
cheers,
Douglas

 

[Joshua_G]

hey-Hey!!!,
As Eli said, the circuit characteristics are for the most part based on the bottom triode. There are a few important characteristics that depend strongly on the upper element. Ability to swing close to its control grid voltage being the most important; this is a pentode that can't come any where near driving its plate below its 'g2'. For the upper element a MOSFET is about the best IME. It can swing the composite's 'plate' very near its g2 voltage, and presents the nearest-to-vertical load line to the triode underneath.
cheers,
Douglas

That's very interesting. Can you link a schematic of MOSFET upper half?

 

[Bandersnatch]

That's very interesting. Can you link a schematic of MOSFET upper half?

It is really simple: treat it just like a triode with a few exceptions. You'll want to clamp gate-source voltage with a good Zener( banded end to gate ), it will want a gate-stopper, and you don't ever have to worry about even miniscule grid currents.
cheers,
Douglas

 

[Gary P]

A variation I've tried with pentodes and will be trying with cascoded triodes in the near future is to put a very good CCS on top of the upper cascode element and connect the plate resistor to ground. The CCS provides all the current for both the cascoded triodes (or pentode) and the plate resistor. This shortens the signal path to include only the cathode resistor, the amplifying element and the plate load. The CCS provides very good power supply isolation.

For an example look at the "driver experiments" section on my web page.

Gary

PimmLabs

 

[Bandersnatch]

A variation I've tried with pentodes and will be trying with cascoded triodes in the near future is to put a very good CCS on top of the upper cascode element and connect the plate resistor to ground. The CCS provides all the current for both the cascoded triodes (or pentode) and the plate resistor. This shortens the signal path to include only the cathode resistor, the amplifying element and the plate load. The CCS provides very good power supply isolation.

For an example look at the "driver experiments" section on my web page.

Gary

PimmLabs

Now that is absolutely Brilliant! I believe you'd also be able to take advantage of the mu output for a low output Z drive too. Going to have to give this a try in the next amp...very sweet; thanks Gary.
cheers,
Douglas

 

[Jaap]

2 ideas / questions

A variation I've tried with pentodes and will be trying with cascoded triodes in the near future is to put a very good CCS on top of the upper cascode element and connect the plate resistor to ground. The CCS provides all the current for both the cascoded triodes (or pentode) and the plate resistor. This shortens the signal path to include only the cathode resistor, the amplifying element and the plate load. The CCS provides very good power supply isolation.

For an example look at the "driver experiments" section on my web page.

Gary

PimmLabs

* I wonder if this idea could be stretched to a hedge circuit (2 cascodes as differential amplifier, with ccs under the cathodes)

* I also would be interested to know if anybody has tried to make a cascode with two different tube types, say a E88C on top of a 76 or ecc40 (=e80cc) or E88C on top of triode of ECL82/86

I could try this out on my test mule with ecl82

 

[TonyTecson]

hi,

please look here: http://www.diyaudio.com/forums/tubes-valves/65656-question-about-single-nos-triodes-cascode-fun.html

 

[Bandersnatch]

* I also would be interested to know if anybody has tried to make a cascode with two different tube types, say a E88C on top of a 76 or ecc40 (=e80cc) or E88C on top of triode of ECL82/86

hey-Hey!!!,
I have used different triodes in cascode and liked the result. 6J6 diff amp with a 6H30 on top for instance. I kept going in this general direction, and use MOSFET's on top now...🙂 The MOSFET allows me to avoid some rather questionable operating points the composite would require. I have upped the game to 6H30 on the bottom and cannot afford the 2x Ec2 voltage as a lower limit to plate excursion; now I can get within a few volts of Ec2 and not leave the 'pentode' portion...🙂
cheers,
Douglas

 

[TonyTecson]

hey-Hey!!!,
I have used different triodes in cascode and liked the result. 6J6 diff amp with a 6H30 on top for instance. I kept going in this general direction, and use MOSFET's on top now...🙂 The MOSFET allows me to avoid some rather questionable operating points the composite would require. I have upped the game to 6H30 on the bottom and cannot afford the 2x Ec2 voltage as a lower limit to plate excursion; now I can get within a few volts of Ec2 and not leave the 'pentode' portion...🙂
cheers,
Douglas

may i know your operating point for the 6H30?

 

[Bandersnatch]

may i know your operating point for the 6H30?

hey Tony,
It has the MOSFET gate at ~110V and is passing 15 mA at idle per triode. It is arranged as a LTP over a CCS, and I don't worry about the value of the bias; that takes care of itself. B+ is over 550V, and the MOSFET drain is ~330V( all from ground ). It can take a 6H6pi with only a slight reduction in gain and these are generally better matched section-section( not to mention being much cheaper ).
cheers,
Douglas

 

[timgrindell]

Yes Yes Yes

A variation I've tried with pentodes and will be trying with cascoded triodes in the near future is to put a very good CCS on top of the upper cascode element and connect the plate resistor to ground. The CCS provides all the current for both the cascoded triodes (or pentode) and the plate resistor. This shortens the signal path to include only the cathode resistor, the amplifying element and the plate load. The CCS provides very good power supply isolation.

For an example look at the "driver experiments" section on my web page.

Gary,
after reading your driver experiments page I had the same idea except to use a hybrid cascode as per AW. Even to using 2 or 3 MAT02 on the bottom for huge TransC. Then with a mu output impedance of circa 500R it is almost ideal to drive a 600R LCR riaa.
Cool
tim
PS AW would call this vapourware (I think)