ampage Tube Amps / Music Electronics	For current discussions, please visit Music Electronics Forum.

ampage archive

Vintage threads from the first ten years

Zener diode regulator

7/17/1997 8:23 AM
Richard Hult	Zener diode regulator Does anyone have any experiences with zener diode regulation for the B+ supply? Just the most simple circuit with a zener diode, a resistor and an electrolytic cap... is it worth the trouble? I will use it for a tube preamp with one pair of 12AX7's and two FETs used as source followers (thanks to R.G. for that idea! The B+ is around 150 volts right now and I'd like to stabilize it a bit. Thanks in advance! Richard (d4hult@dtek.chalmers.se)

7/17/1997 8:23 AM

Richard Hult

Zener diode regulator

Does anyone have any experiences with zener
diode regulation for the B+ supply? Just the
most simple circuit with a zener diode, a
resistor and an electrolytic cap... is it
worth the trouble?

I will use it for a tube preamp with one pair
of 12AX7's and two FETs used as source
followers (thanks to R.G. for that idea!

The B+ is around 150 volts right now and I'd
like to stabilize it a bit. Thanks in advance!

Richard (d4hult@dtek.chalmers.se)

7/17/1997 3:18 PM
R.G>	For a preamp with a pair of tubes, it's hardly worth the trouble. These tubes always run at almost a constant current, so if the DC level is OK, just put more decoupling capacitance on the B+ line. Better, decouple each tube or each amplifier section separately with 10K/10uF. You could spend one of those power fets on a capacitance multiplier, but again, it's hardly worth the change in sound you'd get in most cases.

7/17/1997 3:18 PM

R.G>

For a preamp with a pair of tubes, it's hardly worth the trouble. These tubes always run at almost a constant current, so if the DC level is OK, just put more decoupling capacitance on the B+ line. Better, decouple each tube or each amplifier section separately with 10K/10uF.

You could spend one of those power fets on a capacitance multiplier, but again, it's hardly worth the change in sound you'd get in most cases.

7/18/1997 11:25 AM
Mike Donovan	I agree that the zener method is not necessary and could possibly even introduce "zener noise" into the circuit. I am intrigued by the use of FET's to add a gain stage or a "cathode follower" stage to a tube pre-amp. Do you have any recommendations as to which power mosfets are suitable for this task. I am looking through the Digi-key and Newark Electronics catalogs for N-channel MOSFETs with at least a 600v source-to-drain voltage rating. I've noticed that these devices can handle a hell of a lot more current and wattage than a 12AX7, and some of them seemed to be designed for "switching" applications. I'm a little confused. For a class A, single ended gain stage, do you bias a MOSFET like you would a 12AX7? (100k from "drain" to B+, 1Meg from "gate" to ground, 1.5k from "source" to ground.) To me, this sounds too good to be true! I'd prefer to stay away from any mathematics more involved than ohm's law! Is there hope?

7/18/1997 11:25 AM

Mike Donovan

I agree that the zener method is not necessary and could possibly even introduce "zener noise" into the circuit.

I am intrigued by the use of FET's to add a gain stage or a "cathode follower" stage to a tube pre-amp. Do you have any recommendations as to which power mosfets are suitable for this task.
I am looking through the Digi-key and Newark Electronics catalogs for N-channel MOSFETs with at least a 600v source-to-drain voltage rating. I've noticed that these devices can handle a hell of a lot more current and wattage than a 12AX7, and some of them seemed to be designed for "switching" applications. I'm a little confused.

For a class A, single ended gain stage, do you bias a MOSFET like you would a 12AX7? (100k from "drain" to B+, 1Meg from "gate" to ground, 1.5k from "source" to ground.) To me, this sounds too good to be true! I'd prefer to stay away from any mathematics more involved than ohm's law! Is there hope?

7/18/1997 1:53 PM
R.G.	The substitution is approximate, but is good enough to be very valuable if you need another tube section. Mosfets with BVds of 500 and up are easily available, and do in fact work as a drop in for any EXTERNALLY BIASED 12AX7 cathode follower stage, if you pay some attention to the quirks. The mosfets are enhancement, not depletion devices like the tubes. Therefore, the source will sit at 3-8V lower than the gate, where a triode would have the cathode at a volt or so higher than the grid. This is why you can't use the self bias circuit. That being said, if a triode is biased by an external voltage supply, like a resistor string or the plate of the preceeding tube, you can drop in a mosfet and it will operate as a follwer just as the tube did. You might want to put a 100 ohm damper resistor in series with the gate terminal to suppress parasitics, but I have gotten away without it so far. It's a decent follower. As a directly coupled plate follower, the mosfet is probably an order of magnitude faster than the 12AX7 as a guess, so nothing the driving tube does can "outrun" it. It saturates to lower voltages than the driving tube can, so the thing is still linear when the driving tube is saturated. The gate capacitance is large, maybe 1000nF, which seems like a bad thing, but in the follower mode, this input capacitance is effectively reduced by the 100% feedback of the source follwer mode, so you get full bandwidth. By actual test, the bandwidth of the ones I've used go well beyond audio. This leads to one selection criteria for a device - get the smallest current rating mosfet you can. Power mosfets are actually hundreds or thousands of small fet-cells in parallel on the fet die, and the current rating is not made with bigger cells, just more in parallel. Gate capacitance is proportional to drain current rating, so get as low as you can. I think that if someone had the equipment, cutting a powerfet die down by about 9/10 to a 100ma device would actually make an ideal tube-style follower. Alas, my 20kW CO2 laser die deleter has not come in yet, so I can't test this theory. You should not use these for class A single ended gain stages at all, at least not in the sense you mean. It is too good to be true, and in fact is not true. It won't bias up that way. However -- if you use a jfet "bottom end" as the input to a hypothetical class A gain stage, self biased as a class A gain stage, it IS a depletion mode device, and works fine this way. It just can't take the voltage. If you then connnect a power mosfet in common gate mode, gate tied to about a 12V zener about ground, source tied to the drain of the jfet, and tie the 100K "plate" resistor to the drain of the power fet, you have a composite "device" that is a class A, self biased, linear amplifier that will operate as an amplifier much as a triode does. It will not necessarily sound the same. The transfer characteristics of the "device" will be that of a super high-voltage jfet. Jfets are very much like triodes in the low ends of their curves, shifting over to pentode-like in the high end. The sound has some advantages over bipolar since jfets are square-law devices over much of their range and as such have only second harmonic distortion up to clipping, so it might sound good - just not the same. Drive it into clipping and it will be VERY different from a triode. Good engineering and experimentation would still be necessary, and you would find that no matter what it sounds like, good or bad, there will be a rabid contingent that will insist that no matter how it sounds, it's not good because it's not vacuum, even if they can not pick out which is which in a well run test.

7/18/1997 1:53 PM

R.G.

The substitution is approximate, but is good enough to be very valuable if you need another tube section.

Mosfets with BVds of 500 and up are easily available, and do in fact work as a drop in for any ** EXTERNALLY BIASED ** 12AX7 cathode follower stage, if you pay some attention to the quirks.

The mosfets are enhancement, not depletion devices like the tubes. Therefore, the source will sit at 3-8V lower than the gate, where a triode would have the cathode at a volt or so higher than the grid. This is why you can't use the self bias circuit.

That being said, if a triode is biased by an external voltage supply, like a resistor string or the plate of the preceeding tube, you can drop in a mosfet and it will operate as a follwer just as the tube did. You might want to put a 100 ohm damper resistor in series with the gate terminal to suppress parasitics, but I have gotten away without it so far.

It's a decent follower. As a directly coupled plate follower, the mosfet is probably an order of magnitude faster than the 12AX7 as a guess, so nothing the driving tube does can "outrun" it. It saturates to lower voltages than the driving tube can, so the thing is still linear when the driving tube is saturated.

The gate capacitance is large, maybe 1000nF, which seems like a bad thing, but in the follower mode, this input capacitance is effectively reduced by the 100% feedback of the source follwer mode, so you get full bandwidth. By actual test, the bandwidth of the ones I've used go well beyond audio.

This leads to one selection criteria for a device - get the smallest current rating mosfet you can. Power mosfets are actually hundreds or thousands of small fet-cells in parallel on the fet die, and the current rating is not made with bigger cells, just more in parallel. Gate capacitance is proportional to drain current rating, so get as low as you can.

I think that if someone had the equipment, cutting a powerfet die down by about 9/10 to a 100ma device would actually make an ideal tube-style follower. Alas, my 20kW CO2 laser die deleter has not come in yet, so I can't test this theory.

You should not use these for class A single ended gain stages at all, at least not in the sense you mean. It is too good to be true, and in fact is not true. It won't bias up that way.

However -- if you use a jfet "bottom end" as the input to a hypothetical class A gain stage, self biased as a class A gain stage, it IS a depletion mode device, and works fine this way. It just can't take the voltage. If you then connnect a power mosfet in common gate mode, gate tied to about a 12V zener about ground, source tied to the drain of the jfet, and tie the 100K "plate" resistor to the drain of the power fet, you have a composite "device" that is a class A, self biased, linear amplifier that will operate as an amplifier much as a triode does.

It will not necessarily sound the same. The transfer characteristics of the "device" will be that of a super high-voltage jfet. Jfets are very much like triodes in the low ends of their curves, shifting over to pentode-like in the high end. The sound has some advantages over bipolar since jfets are square-law devices over much of their range and as such have only second harmonic distortion up to clipping, so it might sound good - just not the same. Drive it into clipping and it will be VERY different from a triode.

Good engineering and experimentation would still be necessary, and you would find that no matter what it sounds like, good or bad, there will be a rabid contingent that will insist that no matter how it sounds, it's not good because it's not vacuum, even if they can not pick out which is which in a well run test.

7/21/1997 6:32 AM
Dave Harris	I use ZETEX ZVN0545A, 450V, 45mA, 700mW. The package is E-Line (A flatter TO-92). They also have a complement ZVP0545A. Try and get a 12AX7 to do that ! When looking up the above I saw SIEMENS BSS135, 600V, 80mA, 1W, TO-92. It had a footnote which said. "Depletion mode, normally on". Now that IS interesting isn’t it ? I will see if I can get a fuller data sheet.

7/21/1997 6:41 AM
R.G.	I'd very much like to see some data on the Siemens part. Looks like I'm going to have to butcher some nine pin tube bases to make a plug in module.

7/21/1997 7:32 AM
Dave Harris	Data on its way. I hope.

Page 1 of 3

Next>

Last Page>>