Maker Pro
Maker Pro

High voltage attenuation

J

Jon Slaughter

Jan 1, 1970
0
I have a circuit that has around 500V on it. I want to reduce the voltage to
some sub-circuits. Effectively there is a load that has around 500V and I
wish to attenuate this with regulation. A simple emitter follower with
degeneration should do the trick. The bjt needs to be on the high side
because of the common ground between the different subcircuits.

The only problem is actually programming the base current. I want to avoid
having high voltages on a pot that will be used to set the voltage
attenuation for safety reasons(else I would just use the pot directly
instead of the bjt).

I eventually want to program the base using a uC instead. The load current,
which should be irrelevant to the problem, is around 50mA. I don't need full
scale range but from about max Vcc(95%+) to around 1/4 Vcc. The main thing
is proper attenuation and regulation(which can be done with caps
afterwards). It does not have to work "fast" either.

Any ideas?
 
J

Jon Slaughter

Jan 1, 1970
0
whit3rd said:
Well, a grounded-base NPN transistor, base => +5V and emitter =>
resistor
to a DAC, can translate a safely low DAC output to a higher voltage.
You can add an op amp, change the NPN for NMOS, etc. if
the emitter/source bias point is an issue. Be prepared for
power-sequence issues, of course. The collector might want a fuse,
the emitter a clamp.

But this changes ground. The bjt essentially becoming a low side resistor
with the load being "lifted" off ground by the the bjt. This causes a
problem with transfering signals between circuits since they will be
shifted. If could "restore ground" somehow then I guess it would work.

What do you mean by "power-sequence issues"?

To 'regulate', you need in addition some kind of high-voltage sensing,
my favorite is a R//C sense resistor diode-clamped to GND, just
sink current from the resistor until it matches ground. The
capacitance
of the diodes and error amplifier is swamped by the bypass capacitor
on the high value R.

I'm not sure I follow. Simply adding capacitance across the bjt and emitter
to ground is not any better(although maybe slower)?
 
Top