diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   Simple opamp/mosfet shunt regulator (http://www.diyaudio.com/forums/solid-state/139825-simple-opamp-mosfet-shunt-regulator.html)

iko 5th March 2009 02:24 AM

Simple opamp/mosfet shunt regulator
 
2 Attachment(s)
I would like to bring to your attention a simple shunt regulator which, at least in simulation, shows some promising results. The topology is not novel, I've inspired myself from the Jung regulator, the salas regulator, and a few others. Just the other day I found that Erno Borbely has designed a similar regulator.

I also have another version with the opamp replaced by bjts, which is, of course, not as good on paper.

This opamp version shows some of the best psrr results of a number of regulators, including the Jung super regulator. It also shows a very good transient response to sudden changes in the load. Showed in the following posts you will see some of the simulated results. I have also built this regulator, and it looks good on the oscilloscope, but I don't have any specialized testing equipment to do more tests.

Resistor R13 should actually be 5K variable, that's what I used. That needs to be set to find the spot where oscillation does not occur. The output impedance is in the microohms until 100kHz.

All capacitors in the simulation were set to an ESR of .5 ohms and an ESL of 10nH.

Most people here will care about the sound and nothing but the sound. This is something I cannot comment on, everyone has their own idea about what sounds good.

Any comments, suggestions, are greatly appreciated.

iko 5th March 2009 02:25 AM

2 Attachment(s)
Here is the psrr curve.

iko 5th March 2009 02:29 AM

2 Attachment(s)
What if the load is a square wave with a few microseconds rise/fall times, between 75 and 100mA? -- here is the regulator output.

iko 5th March 2009 02:33 AM

2 Attachment(s)
And the output impedance plot, as well as the calculated zout (calculated by the spice directive .tf v(out) v1 )

--- Transfer Function ---

Transfer_function: 2.84731e-011 transfer
v1#Input_impedance: 0.00333333 impedance
output_impedance_at_V(out): 7.63865e-008 impedance

Salas 5th March 2009 03:19 AM

Congratulations! This is the most impressive set of simulated results for any reg that I have seen so far. The fact that you have built it and it does not oscillate is a major step. I don't want you to describe the sound, but did you perceive it as different considering your previous regs? Also have you got identical line to scope's gnd? I ask because such simulated results will manifest somehow even if realized at their half potential in practice.

iko 5th March 2009 04:00 AM

salas, you're too kind, but I think that it's too early for congrats.

Anyway, I thought I stop poluting your threads and create one for this regulator, since the results were surprising for me too.

Cannot say anything about the difference in sound because I haven't plugged it in the phono yet. I was looking around for a schematic to build a little "noise amp" because the oscilloscope shows as good a trace as the other one we discussed before, the opamp buffered regulator. I think better test equipment is needed to test this beast. Not only that, but it should be built with more care to layout; now it's a rat nest :) only a prototype to quench my curiousity, really. Certainly looks promising.

analog_sa 5th March 2009 06:12 AM

The shunt part is very similar to what i use at present. My 317 CCS is probably less good and eventually i intended using the same as Salas but the sound even with the 317 is better than Jung.

Can you publish CCS results alone?

What is the purpose of R15?

dodo 5th March 2009 07:19 AM

I would be intrested to see behaviour of that shunt with the large ( 1000uF) low ESR cap on the output


R15 with c5 build a RC circuit and plays ( I think ) the same role as in the Jung regulator

analog_sa 5th March 2009 07:44 AM

Quote:

Originally posted by dodo
R15 with c5 build a RC circuit

Difficult to argue with the above :)


The purpose of these parts in the Jung regulator is to make both dc and ac impedances at the opamp inputs equal. R15 doesn't seem to be helping much in this respect.

Bonsai 5th March 2009 08:40 AM

Connect R15 in series with the op-amp input and then put th e cap from th e op-amp input to ground - this is a filter.

There is no need in this type of circuit to focus too heavily on keeping the op-amp input resistances th e same since these source resistances are low enough no to cause any significant errors.

This is in any event a good sim result. I generally use 317/337 with the reference decoupled in my PSU's and then I decouple the op-amps with 22Ohms followed by 100uF right at the op-amp supply pins (I do this on each op-amp both + and - supply pins) which makes for a pretty low noise source.

I've sim'd the Jung and various shunt regulators and my view is that the shunt offers superior performance.


All times are GMT. The time now is 05:06 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2