Ramping a reference voltage up from, or down to zero volts, over ~200us

[ocset]

Hi,
Please assist us in  ramping a reference voltage  down to zero volts, or back up to the reference voltage. This being done  at random times, on demand. The ramp should be over some 200us. It should be linear, but its not too bad if its not that straight….it could even  be say  in 10 steps (up or down).

The reference voltage could be any level from 0V to 0.5V.

….So eg it could be at, say, 0.23V…and every now and again we would  need to ramp it down to zero volts….then some time later, ramp it back up to 0.23V.

Would you agree that the attached method is the best way?
It just uses a microcontroller to PWM the buffered reference voltage…variable PWM to make it ramp over 200us (please ignore RC values in schem) …as discussed, it could be 10 steps, so 10 different duty cycles inside the 200us ramp time.
Please confirm there are no nice analog modules which do this, and no nice sub circuits which can do it better than this?

(I wasnt going to post, but thought this may be of interest to  others...and i believe i have provided a solution, though unfortunately it uses a microcontroller)

[rstofer]

How does V_DS (ON) play into this?  It isn't 0.000V - it might get down to 0.09V, give or take, but always greater than 0V.  Maybe reaching absolute 0.000V isn't really necessary.

If it doesn't matter how long the output is 0V and it doesn't matter how long it stays at the reference value and the primary consideration is the 200 us ramp, I would use a relay.  That way I can get to 0V (given suitable gold plated bifurcated contacts).

[ocset]

Thanks, actually the slight vds(sat)  of say a 2N7002 NFET will be fine.

Its basically a power supply which is getting switched in and out at certain times...and the reference needs to gradually ramp down (or up)  in ~200us....but there's no real precision needed here.

[ocset]

Hi
Actually, now i think  200us is too small a time for a micro to do the ramp….so I have done the attached circuit…however, it needs a +/-5v supply and the ramp is a little less linear than I’d like.
(LTspice sim and pdf schem attached)
Any improvements much appreciated.

[David Hess]

Pulse width modulation is not a bad way to do it.  I would have used an analog switch but a 2N7000 should work also.

A buffered RC filter is much simpler and the exponential waveform is usually sufficient.  A very linear slew rate limiter can be made with current sources and a diode bridge.