Hi,
it has been told to me that a 12AX7 tube cannot drive a KT88 triode-stripped tube: so, since I'm a newbie, I'd like to know wath is Miller capacitance, and which requisites should a driver have to drive such a tube?
I mean, I was thinking that a tube can be driven by a voltage, so why should I provide a "stronger" driver than 12AX7? I can suppose there's some current to drive, but which one?
Is there some form of a capacitance between the grid and the rest of the tube? And so, I have to drive it?
How can I calculate the needs for the driver?
Thanks,
Giovanni
it has been told to me that a 12AX7 tube cannot drive a KT88 triode-stripped tube: so, since I'm a newbie, I'd like to know wath is Miller capacitance, and which requisites should a driver have to drive such a tube?
I mean, I was thinking that a tube can be driven by a voltage, so why should I provide a "stronger" driver than 12AX7? I can suppose there's some current to drive, but which one?
Is there some form of a capacitance between the grid and the rest of the tube? And so, I have to drive it?
How can I calculate the needs for the driver?
Thanks,
Giovanni
Just a quick and partial answer: Miller capacitance is the capacitance seen by the driver looking into the control grid. It is composed of the natural capacitance between the anode and grid, but becomes multiplied by the mu of the valve. Triodes exhibit this.
The driver needs to supply current to overcome this.
The driver needs to supply current to overcome this.
"it has been told to me that a 12AX7 tube cannot drive a KT88 triode-stripped tube: so, since I'm a newbie, I'd like to know wath is Miller capacitance, and which requisites should a driver have to drive such a tube?"
That's true, the 12AX7A isn't a very good current source. As for Miller capacitance, this problem occurs with the common cathode (emitter, source) topology. The Cgp acts as a positive feedback path. Whenever you have positive feedback, the admittance of the feedback element appears to be multiplied by a factor of: Cgp`= Cgp(Av + 1). This adds to the Cgk already present, which is why common cathode amps have a limited high frequency response. Furthermore, changing the Vgk doesn't have an immediate effect, but only once the input + stray capacitance is all charged up. Fast charging requires a high current, otherwise, you run into slew rate limiting and thus distortion. Feeding adequate current to a triode final requires something like a high gain pentode (6CB6A for smaller power triodes, 12BY7A for larger ones) or a cathode follower (6SN7's good for this) or a MOSFET source follower (very good: MOSFET can source current better than any VT), or an SRPP. All these have been used at one time or another.
"How can I calculate the needs for the driver?"
SR= 2pi * f * Vp (slew rate in volts/sec)
i(t)= C(dv/dt) (where dv/dt= slew rate, and C equals the sum of Miller + Cgk + Cstray. Figure Cstray ~= 30pF)
That'll tell you how much current. Although, it's often not necessary to calculate this with any great degree of accuracy. Usually, any medium u triode driver operated at reasonable plate currents will be able to handle the job. You might run into problems where very large power triodes are being used, like 845's or 833A's, or something like that.
That's true, the 12AX7A isn't a very good current source. As for Miller capacitance, this problem occurs with the common cathode (emitter, source) topology. The Cgp acts as a positive feedback path. Whenever you have positive feedback, the admittance of the feedback element appears to be multiplied by a factor of: Cgp`= Cgp(Av + 1). This adds to the Cgk already present, which is why common cathode amps have a limited high frequency response. Furthermore, changing the Vgk doesn't have an immediate effect, but only once the input + stray capacitance is all charged up. Fast charging requires a high current, otherwise, you run into slew rate limiting and thus distortion. Feeding adequate current to a triode final requires something like a high gain pentode (6CB6A for smaller power triodes, 12BY7A for larger ones) or a cathode follower (6SN7's good for this) or a MOSFET source follower (very good: MOSFET can source current better than any VT), or an SRPP. All these have been used at one time or another.
"How can I calculate the needs for the driver?"
SR= 2pi * f * Vp (slew rate in volts/sec)
i(t)= C(dv/dt) (where dv/dt= slew rate, and C equals the sum of Miller + Cgk + Cstray. Figure Cstray ~= 30pF)
That'll tell you how much current. Although, it's often not necessary to calculate this with any great degree of accuracy. Usually, any medium u triode driver operated at reasonable plate currents will be able to handle the job. You might run into problems where very large power triodes are being used, like 845's or 833A's, or something like that.
Thanks to all of you!!!
Miles:
You gave me a good idea, I'm pretty skilled with Mosfet, I built in the past several Mosfet amplifier (Nelson Pass Aleph) and, thus, building a input/driver stage with suh a device should be simple, from my point of view.
The idea that is coming into my mind is to build a Mosfet differential input (LTP) with a CCS that delivers something like 20mA; this is usually more than enough to drive any kind of power mosfet (I drove 3 paralleled Mosfets with such a configuration), and, then, enough also for a tube.
I will try to make some calculation on it.
Another possibility is to use two ECL84 I have laying around (I taken apart an old oscilloscope) as drivers. Could it be one? As far as I know such a tube has a "low" plate voltage, I do not know if it is enough.
Thanks,
Giovanni
Miles:
You gave me a good idea, I'm pretty skilled with Mosfet, I built in the past several Mosfet amplifier (Nelson Pass Aleph) and, thus, building a input/driver stage with suh a device should be simple, from my point of view.
The idea that is coming into my mind is to build a Mosfet differential input (LTP) with a CCS that delivers something like 20mA; this is usually more than enough to drive any kind of power mosfet (I drove 3 paralleled Mosfets with such a configuration), and, then, enough also for a tube.
I will try to make some calculation on it.
Another possibility is to use two ECL84 I have laying around (I taken apart an old oscilloscope) as drivers. Could it be one? As far as I know such a tube has a "low" plate voltage, I do not know if it is enough.
Thanks,
Giovanni
Hi Giovanni,
The triode strapped pentode section of ECL84 surely would make a "little monster" driver when being triode strapped. Even suitable for LTP with its fairly high gain. Quite linear, too.
Have a look at the triode strapped plate curves at my site:
http://www.tubes.mynetcologne.de/roehren/daten/ecl84pentode_as_triode.pdf
Tom
The triode strapped pentode section of ECL84 surely would make a "little monster" driver when being triode strapped. Even suitable for LTP with its fairly high gain. Quite linear, too.
Have a look at the triode strapped plate curves at my site:
http://www.tubes.mynetcologne.de/roehren/daten/ecl84pentode_as_triode.pdf
Tom
- Status
- Not open for further replies.