High Voltage Reference

[runeight]

hey tube folks, if I wanted to replace a voltage reference like an 0A2 with an SS equivalent, what would be the best way to do it?

for example, in a simple tube series pass regulator with a tube error amp working against a VR tube, what is the best way to replace the VR tube with its sand equivalent(s)?

HV zeners are not stable nor robust enough for sure.

is there a good SS configuration that would be temperature stable as well as being a stable reference?

did a little searching and didn't find anything, but please point to links if i've missed something.

thanks

 

[jackinnj]

you can use a TL783 from Texas Instruments -- just lift the ADJ pin above ground with a Zener diode -- the TL783 on its own can handle 125V. The adjust pin current is pretty minimal.

Look at this circuit for a similar idea (my repair of a Fairchild 255A) http://www.tech-diy.com/fairchild_255a.htm

Nat Semi has a couple easy to implement HV circuits on their website.

 

[fdegrove]

Hi,

is there a good SS configuration that would be temperature stable as well as being a stable reference?

Assuming you want to keep the series pass device tube and the error amp as it is the OA2 or other VR can be replaced with a stack of zeners to arrive at or close to the original voltage.

The best behaved zener reference are supposedly the 5.6V (or thereabout) but they'd still require some damping to reduce their tendency to emit the occasional noise outburst.

Either way you look at it, it's still an awful lot of diodes just to replace that gorgeous looking (snicker, snicker) and far more reliable OA2...

Cheers, 😉

 

[SY]

I've had very good results from using a jFET current source stacked on top of a large resistor. You have to protect it at power-up and power-down, but that's just a couple of diodes.

 

[jackinnj]

I've had very good results from using a jFET current source stacked on top of a large resistor. You have to protect it at power-up and power-down, but that's just a couple of diodes.

which is worse, the noise from a stack of zeners, or the
SQRT 4KRT *10^9 resistor noise ?

 

[SY]

Hard to say, but once they've both been through an RC filter, there's probably not much difference between any of the reference options- I haven't experienced the Zener string problems that runeight describes. The FET-CS/resistor combo was merrily lifted from a Joe Curcio design.

 

[jackinnj]

Hard to say, but once they've both been through an RC filter, there's probably not much difference between any of the reference options- I haven't experienced the Zener string problems that runeight describes. The FET-CS/resistor combo was merrily lifted from a Joe Curcio design.

In my early days of regulators -- trying to remake an Audio Research SP3 from scratch -- I learned the hard way that a string of zeners is better than one high voltage zener ! and the diodes were much more expensive then (early 19980's).

 

[Gary P]

I've had very good results from using a jFET current source stacked on top of a large resistor. You have to protect it at power-up and power-down, but that's just a couple of diodes.

Another option that is available now is to do the same circuit but replace the J-Fet with a Supertex LND150N3 depletion mode mosfet. The part is rated for 500 volts and won't need protection at startup/shutdown. IDSS is in the 1.5ma range. Add a current setting resistor to drop the current to the .2ma to .4ma range. Cascade 2 of them for a much more thermally stable circuit. I use this circuit to generate the voltage reference in a couple of my CCS designs and it has worked great.

Gary

 

[runeight]

Gary, which of you schematics should I look at?

 

[Gary P]

Look at the voltage reference in either the Rev. 5 or selfbias CCS's. Both circuits are the same, basically a CCS feeding a resistor with a cap across it. In the CCS circuits they are tapped to provide different bias voltages, one of them adjustable for the 2 mosfets.