CCS in Placid-BP Bipolar Shunt Regulated Power Supply

[planet10]

Originally Posted by maximus

the posts that were deleted

no posts were deleted. They were just moved into this new thread in a different forum.

[maximus]

Hi Russ,

I usually drive the output of regulators from a defined source impedance at a variety of frequencies up to 1 MHz, both sine and square wave and observe the results on a 250MHz scope. I adjust the drive impedance to suit the load current swing required for the circuit under test. Sine wave drive gives the nominal output impedance at any given frequency and square wave drive gives the transient response information.

Hi ikoflexer,

Have you cascoded M2? Or are you just referring to the other parts that are cascoded.

Regards
Paul

[Russ White]

Hi Paul,

That is basically the same as I have been doing. But I add an instrumentation amp to get a more clear picture of the error voltage. But Since Placid is not a local reg I didn't care so much about F > 500khz. Still I have checked, and it does very will even way up there.

Placid is actually very easy to tune.

I added a cascode today to test, and did not observe any meaningful change in performance. Definitely nothing worth the added complexity in this situation.

Cheers!
Russ

[ikoflexer]

Originally Posted by maximus

Hi ikoflexer,
Have you cascoded M2? Or are you just referring to the other parts that are cascoded.
Regards
Paul

Sorry Paul, I don't know which schematic you're referring to. I've cascoded the CCS mosfet in my circuit, without much improvement. I found that a simple capacitor multiplier before the regulator improves psrr a lot more. It's probably off topic here, so you're more than welcome to see the circuit I'm talking about following the link in my signature.

[maximus]

Hi Russ,

How close are your actual measurements to your simulation software results? I ask because the post you addressed to Allen and I, quoted the spec below,

Allen, if your question was directed to me, then yes, I have indeed measured my Placid BP the best I know how to. The output Z is below 5 millohms up to 30khz. This was measured by running an amplifier generated 20ma peak sine current at the output and measuring the output voltage ripple after a 30db gain stage. This was on top of a static 100ma load.

The output Z is somewhat dependent on the shunt current. But it remains excellent even with shunt currents as low as 30ma.

For reference this is the output Z as predicted by LTSpice. Real life performance is of course not quite as good, but close.

This spec is very different to the simulator curve you show in post 11 on this thread. I am asking because I am considering how useful simulations may be to my work. If they are reasonably accurate they may save me some time.

Hi ikoflexer,

I was referring to the circuit in post 1 from the thread link you provided. My apologies if you are now using a different circuit.

Capacitance multipliers are pretty good at removing ripple and noise from noisy voltage sources when operated under continuous current draw, as would be the case when supplying a constant current shunt reg. It is quite conceivable that you could get a better overall performance improvement compared to cascoding a simple current source mosfet. The topology of the complete current source circuit will dictate the overall performance of the current source. Some topologies are better than others. Unfortunately there is not much I can say here as the ones I use are subject to non-disclosure agreements.

Regards
Paul

[digger945]

Originally Posted by Russ White

Now for local regulation requiring even better transient response I have a CCT coming specifically for that. Something probably not unlike your three pin regulator replacements.

Happy to hear this news. I will wait to see what you have before going any further with builds or mods.

It would be great if these could be used or modified for specific use if needed, either digital or analog.