Pico's Simple Shunt Regulator

[2 picoDumbs]

I started out with the Salas Simple Shunt Regulator http://www.diyaudio.com/forums/powe...listic-salas-low-voltage-shunt-regulator.html but with a TL431 as the reference with a trim pot for an easily adjustable power supply however I really wanted to use lm329 but you don't have the convenience of adjustable Voltage Out like you do with TL431.
Then I had an idea...........

Here it is.
Simplified constant current source at input (I already have these laying around ie IXTH20N50D (not IXTT package)) it does the job.
D1 is LM329 but I don't have a model for it so I used a 6.2V zener just in simulation don't let that give you a brain melt down.
R2 is a 5k Trimpot set at 1k for around 24V out.
For a supply voltage below 24V either increase the value of R9 eg 18k or consider using a precision 2.5 V reference etc. There are quite a few different ways to achieve the supply voltage you require.

Output impedance.

Feel free to tweak this anyway you wish, I'm sure you guys can think of some improvements or modifications (I have thought of a few other ideas also)

 

[2 picoDumbs]

Note: There should be a resistor at the gate of M1. Around 100 Ohms should suffice
Doesn't hurt to put gate resistors on 2n3819 either. Say 1k should cover it.

 

[AndrewT]

The input voltage is shown as 32Vpk, This is near the limit for the bc847 smd. OOPS I see the 847 is also a 45Vceo transistor and is available in a, or b, or c grades.
Allow upto 40Vpk as the maximum input voltage and you can use BC850b, or BC850c, which are rated as 45Vceo

You could change R1 to a CCS (2mA to 4mA)

and raise the voltage reference to 12V, or 15V

 

[2 picoDumbs]

I'm planning on using 2sa970 and 2sc2240, I just used bc847a during simulation cause the model more closely matched the hfe of the devices I have.
Other people will probably want to use bc550 bc560.

I should have mentioned that.

 

[as_audio]

This diagram reminded me immediately of an early apprentice piece that I did for
"Audio Amateur" and german "Elrad" magazines in 1990, "Parallel Regulators for
Audio".

See attachment Fig. "Bild 7". This one is in german.

Sorry, the complete file is said to be too large by a small margin:
"Your file of 1.06 MB bytes exceeds the forum's limit of 976.6 KB for this filetype"
Perhaps moderator can change this for one time?

PS I love your forum nickname.

 

[needtubes]

Looks like there is an LM329 SPICE model in this thread. Perhaps it will work for you?

 

[2 picoDumbs]

This diagram reminded me immediately of an early apprentice piece that I did for
"Audio Amateur" and german "Elrad" magazines in 1990, "Parallel Regulators for
Audio".

See attachment Fig. "Bild 7". This one is in german.

Sorry, the complete file is said to be too large by a small margin:
"Your file of 1.06 MB bytes exceeds the forum's limit of 976.6 KB for this filetype"
Perhaps moderator can change this for one time?

PS I love your forum nickname.

That looks like a great article.
I have an amazing ability of being able to reinvent the wheel.

 

[2 picoDumbs]

Looks like there is an LM329 SPICE model in this thread. Perhaps it will work for you?

Thanks.

 

[Mark Johnson]

I'm not a big fan of the current source load; it'll be finicky and annoying to set the JFET+resistor current equal to exactly half of the tail current. I suggest adding two additional parts (gasp!): use a degenerated PNP current mirror instead. 2xPNP + 2xEmitterResistor. BTW PNPs are cheaper and more plentiful than JFETs.

Voila, guaranteed 50/50 current split between legs of the differential pair. And since the collector-emitter voltage is guaranteed to be extremely low (0.7V on the left, 2V-4V on the right), you can fit low voltage high beta PNPs without worry. Like the supremely wonderful ZXTP07012 or 2SA1954.

And if you want that last factor of 3x reduction in output impedance, you could replace the PMOS with a power PNP bipolar. Since its gm is about 3x higher, it's (1/gm) contribution to open loop output impedance is about 3x lower. Being a follower, it doesn't change the open loop gain, so the closed loop output impedance is the same feedback-factor (100X??) lower than the open loop output impedance. Voila, closed loop Zout is 3x lower.

If you stick with PMOS you may want a 16V zener, to prevent excess Vgs if and when something goes very wrong.

edit- stability components?

 

[2 picoDumbs]

Thanks. I'll try the current mirror.

 

[kevinkr]

<snip>
Sorry, the complete file is said to be too large by a small margin:
"Your file of 1.06 MB bytes exceeds the forum's limit of 976.6 KB for this filetype"
Perhaps moderator can change this for one time?

PS I love your forum nickname.

Perhaps this will work?: https://smallpdf.com/compress-pdf

 

[as_audio]

Perhaps this will work?: https://smallpdf.com/compress-pdf

Seems not to work on Linux.

 

[as_audio]

Now, it was possible to compress on their homepage
also, but the quality of the scan is impaired. Fig. 7 on
third page in this context.

"Parallel Regulators for Audio" in german, the version
in "Audio Amataeur" is essentially the same, but had
some editorial mischief.

 

[2 picoDumbs]

Yeah that worked.
Any chance you can do an english translation of the article?
Looks like a good article.

 

[as_audio]

Any chance you can do an english translation of the article?
Looks like a good article.

Somebody may want to do a scan of the article in
The Audio Amateur 4/90. A missing figure, "Bild 11"
appeared in the next(?) issue of TAA.

 

[as_audio]

I found a historic reference to shunt regulators, it is from 1969 and written by Walt Jung.
http://waltjung.org/PDFs/Dont_Shun_the_Shunt_Regulator.pdf
Compare fig. 2 (D1 is a transistor used as Zener reference) to the diagram in post 1.
Don't worry about the grounding, this one is going negative. Another excellent article by W.J.

 

[2 picoDumbs]

Yeah got that one.
I have most of Walt's articles.
He is a good guy.

 

[as_audio]

The pdf in post 16 disappeared because Walt Jung homepage is down.
Let me try to attach it here :

 

[as_audio]

In fig. 2 of this pdf an arrow is missing for Q5 (npn).
This must be the only error in a Walt Jung article ever ..

 

[Bigun]

Nah, must be a bug in the pdf document encoding