Simple Cascoded Regulator

[annex666]

I had an idea for a simple cascoded voltage regulator that uses a zener diode as a reference that is follower by an emitter follower.

The current source for the zener is shielded from variations on the input by virtue of the cascoding used. Therefore better rejecting input variations on the output.

I'm sure this (or something much like this) has been done before and don't claim that it's a new idea.

Any thoughts anyone?

[annex666]

Note that the 15V and the 4.4V are reversed - PSpice went a little crazy on me for a while!

[hjelm]

Neat!

Question how do you see the balancing between I7 and I8?

If I8 is smaller than I7 that could cause some trouble if there is a very light load but i assume that it is just necessary to have i8 bigger than I7.

Is there a need for using a cap across the zeners to reduce the noise?

Is it possible that it could ocillate since you have a feedback?

Have you tested the BW?

[annex666]

The idea was to make I8 bigger than I7 to keep T1 and T2 in the forward linear region therefore maintaining the cascoding. If this were not the case when there is no load there would be no cascoding effect.

The diodes would be bypassed with caps. to bypass ac signals.

I've not measured the bandwidth as I'm quite new to PSpice and I'm having a few troubles simulating - anyone else feel free if you want to sim. the circuit and post results.

Has anyone seen something like this used before?

[Nelson Pass]

Another option you can explore is simply to regulate the
supply twice, and yet another option is to place lots of
capacitance to ground both transistors' emitters, as well
across the Zener diodes.

[skp112233]

A cascoded 6V shunt regulator was featured in the June 1995 issue of Audio Amateur under the title 'Quickie DC for Radio Shack's CD-3400' with very good specs.

[annex666]

Here is an updated schematic showing in detail the construction of the current sources - this is just a first draught so if anyone has suggestions for better part values.

Note that I did this quickly so there may be some mistakes, also I was limited to the transistors that are in the standard PSpice library (not ideal!).

Cheers for looking all.

[annex666]

Here is a graph of PSRR of the circuit. As you can see it's not great, but shows the principle works (at least in simulations).

There is a LOT of room for improvement so if anyone has suggestions for different components/circuit changes post your ideas.

[janneman]

I think the basic decision to make is either to stay with the open-loop supply as it is (and which means that the name regulator is a bit misleading) or to add a regulation loop and make it a real regulator.
If you stay with the open-loop as it is, the improvements will be limited. Adding a control loop would give you a jump in performance. But since that would involve feedback you may not want to do that for personal reasons.

In the circuit as is, I wonder what the effect on the output quality is of the circuitry around Q7-Q8. In fact, Q7 only adds extra load current. Since the reference comes from the final output, the anode voltage of D9 is very stable and clean. I would try a simulation with everything around Q7-Q8 removed. You *may* need a resistor from the anode of D9 to ground for start-up purposes only, but other than that I expect no change.

Jan Didden

[annex666]

I have no objections to feedback what so ever - although do not believe it the definitive solution to any problem.

I believe in general that if a circuit can be improved before the feedback loop is closed this should be done, then the loop closed.

Please feel free to post your ideas for changes to the above schematics. It would be interesting to have schematics from others on the thread.

Thanks for the posts one and all.