The simplistic Salas low voltage shunt regulator

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
OOps, upside down. Sorry about the faint yellow for component outlines - it takes theis layer directly rom the original library specs - will change this later, too.
 

Attachments

  • scan0005.jpg
    scan0005.jpg
    74 KB · Views: 1,907

iko

Ex-Moderator
Joined 2008
First, you should not use the default diode models that come with ltspice. After you place the component on the screen, right-click on it and click "Pick new diode" then choose something that makes sense. If you don't have models for the proper zeners I can point you to some. You already found out how to include new models in ltspice right?

There is no need to include a bridge rectifier in this type of circuit. Just simulate the psu after the rectifiers, by inserting a voltage source with a DC component of about 30V and a sine wave function of desired voltage ripple and frequency (1V 60Hz or similar).
 
Formerly "jh6you". R.I.P.
Joined 2006
Salas said:

That is RDS On. Does not translate here. Has to do with pulse applications and heat efficiency.



I was thinking about the resistance, which is the "inverse of its transconductance" . . . So, if the regulated supply rail is assumed as an ac ground by the following amplifier circuit, the resistance should be counted on. No . . . ?


:)
 

iko

Ex-Moderator
Joined 2008
A bunch of .lib files can be found on this site:
http://www.elektronikschule.de/~krausg/Spice_Model_CD/
http://www.elektronikschule.de/~krausg/Spice_Model_CD/Mixed Part List/Spice-Models-collection/
http://www.elektronikschule.de/~krausg/Spice_Model_CD/Vendor List/Philips/Philips_Spice/
http://www.elektronikschule.de/~krausg/Spice_Model_CD/Vendor List/Motorola/Spice/
http://www.elektronikschule.de/~krausg/Spice_Model_CD/Vendor List/Zetex/
http://www.elektronikschule.de/~krausg/Spice_Model_CD/Vendor List/Siemens/
http://www.elektronikschule.de/~krausg/Spice_Model_CD/Vendor List/Spice-Models-collection/

IRF models: http://www.irf.com/product-info/models/spice/spice.zip

Most vendors provide spice models for their products, it's just a matter of searching. With the time your library will grow.

The way to include a model in your simulation is as follows. Let's assume I downloaded a file named pwrmos.lib which includes a model for the IRFP9240 which I'd like to use in my simulation.
1) I like to keep downloaded models in a separate directory, so I create a folder
c:\Program Files\LTC\LTwpiceIV\lib\sub\mylibs
ltspice is installed in c:\Program Files\LTC\LTwpiceIV on my computer, yours will be where ever you chose to install it

2) place the downloaded pwrmos.lib file in the created directory
c:\Program Files\LTC\LTwpiceIV\lib\sub\mylibs\pwrmos.lib

3) inside ltspice, include a spice directive by clicking on the .op button and then typing
.lib mylibs\pwrmos.lib
and place the directive on the screen next to your circuit

4) include a pmos component in your simulation if you have not done so yet, then while holding the Ctrl key pressed, righ-click on the component with the mouse. You will get a dialogue with a table of Attributes and Values. Chance the "Value" attribute by clicking on it, to IRFP9240. This is the name of the .model line in the file pwrmos.lib. Then press OK and you're done, you're using the right model. This works for models that come written as ".model ..... " There will be models which are written as ".subckt .... " and for those you have to change the "Prefix" attribute, and make it "X" . In general, when you download a .lib file, you need to look in it with a text editor such as notepad and find the model that you're looking for, see if it's a .subckt or a .model, and use it accordingly.

Hope this helps.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Babowana said:
The regulated rail is following the source of M1 (IRFP9240). So, the "resistance" will directly look at the ac voltage and its internal resistance, if any on the regulated rail . . . Hmm . . . I might be wrong . . .

I don't doubt about nice work of your circuit, anyhow . . .

Nice talk . . .


:)

No, that would be true if it was just a pass series open loop element. It feeds current to a shunt regulation loop that reacts by BC550gm*collectorLoad. Its the reaction that governs. The mOhm regions of output impedance for such shunts are real. Not as low as in sims due to layout, wires, parasitics, etc. but really low.
 

iko

Ex-Moderator
Joined 2008
Improved deluxe v2

And just when I thought I should give up on design improvements and start getting real, this comes up. The deluxe bleeding edge v2 edition, for those who can afford an extra few components. Scalable too, I think.

Output impedance compared to already the excellent v1. We're talking about 25 microohms in the audio band, and less than 100 microohms at 200kHz.

reg-salasiko-new1d-zout.jpg


Check out the output voltage with 30V DC, 200mV AC input, and a sine active load of 65mA DC, 15mA AC, 18kHz (to test for load varying in the higher frequency audio band).

reg-salasiko-new1d-transresp.jpg
 

Attachments

  • reg-salasiko-new1d.jpg
    reg-salasiko-new1d.jpg
    45.6 KB · Views: 2,128
Just when you thought it was safe to get back into the water!

Hi Ikoflexer,

I am eagerly awaiting your book on shunt regulators......:D

I guess you would call a cascoded shunt, first one I have seen....cool.

I have one thought, have you considered replacing the shunt element with a complementary pair and would there be an advantage in doing so?

Keep up the good work.

Regards,
Jam
 

iko

Ex-Moderator
Joined 2008
Re: Just when you thought it was safe to get back into the water!

jam said:
Hi Ikoflexer,

I am eagerly awaiting your book on shunt regulators......:D

I guess you would call a cascoded shunt, first one I have seen....cool.

I have one thought, have you considered replacing the shunt element with a complementary pair and would there be an advantage in doing so?

Keep up the good work.

Regards,
Jam

Yes, good idea, I have thought of it but couldn't get it to work acceptably. At the moment I think I can live with this shunt regulator if it works as a real circuit. If so, I am curious if other people will see a noticeable sound improvement; my ears I don't trust, I'm afraid.
 
These designs appear to be offering excellent performance! I am not smart enough in this area to know much about how they work, but understand the primary application is in noise sensitive phono amps?

How about then a dual polarity (+/-15V) application, possibly with active ground also, for op-amp applications? I am thinking for active crossovers myself. Is it completely unrequired here with the inherently high PSRR of op-amps and relatively high signal voltage levels? Will an LM317+LM337 duo with standard grounding give comparable results here? I am seeking the best solution for powering my active crossover when I implement it (though op-amps will "only" be OPA2314) :)
 

iko

Ex-Moderator
Joined 2008
Dr.EM said:
These designs appear to be offering excellent performance! I am not smart enough in this area to know much about how they work, but understand the primary application is in noise sensitive phono amps?

How about then a dual polarity (+/-15V) application, possibly with active ground also, for op-amp applications? I am thinking for active crossovers myself. Is it completely unrequired here with the inherently high PSRR of op-amps and relatively high signal voltage levels? Will an LM317+LM337 duo with standard grounding give comparable results here? I am seeking the best solution for powering my active crossover when I implement it (though op-amps will "only" be OPA2314) :)

Initially salas developed the shunt regulator to be used with his phono stage design, but I think it grew past that since. My effort lately was to extend lower output impedance in the higher frequency band specifically so that it can be used with medium power amps, or even power amps (yes, I'm dreaming). salas knows people which already use his regulator with some power amps. The lower the output impedance, the more constant the output voltage will remain when the power amp needs to draw more current suddenly (such as is the case with strong bass, etc.). I have often hooked up the oscilloscope to Vout of some regulator and "watched" the music.

We've got people working on pcb layout for the original regulator, and I'm currently working on a layout for the experimental versions as well. I would not want to give you any guarantees that this or that is better than I don't know what. It's small parts count and easy enough to build to see for yourself if it fits the bill.

As far as the PSRR of various opamps, watch out for not so good performance in the higher frequency range, even within the audio band. Attached is a plot of line regulation for the salas v1 (blue) and the experimental version (green). In the fantasy land of simulators these are not very shabby results.
 

Attachments

  • reg-salasiko-new1d-linereg.jpg
    reg-salasiko-new1d-linereg.jpg
    88.4 KB · Views: 1,948
diyAudio Chief Moderator
Joined 2002
Paid Member
Dr.EM said:
These designs appear to be offering excellent performance! I am not smart enough in this area to know much about how they work, but understand the primary application is in noise sensitive phono amps?

How about then a dual polarity (+/-15V) application, possibly with active ground also, for op-amp applications? I am thinking for active crossovers myself. Is it completely unrequired here with the inherently high PSRR of op-amps and relatively high signal voltage levels? Will an LM317+LM337 duo with standard grounding give comparable results here? I am seeking the best solution for powering my active crossover when I implement it (though op-amps will "only" be OPA2314) :)

User Marinos has made such +/- shunts for OPA627s. He reported serious gains over batteries he used for years. Seems like high PSRR op amps don't zero the shunt goodness.

The circuits

His mini report

Maybe he can say or show more in this current thread at a point.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.