Thoughts on this low power regulator ?

[rbroer]

I would appraciate any thoughts on this low power 5V regulator.
I can cram it in 5 cm2, so it can be placed very close to the point where it's needed.
A current feedback opamp like AD811 can supply up to about 100mA, datasheet mentions voltage noise is quite low.
I measured current draw of 59mA, when feeding 10V and outputting 5V into 100 ohms.
BTW, there can't be an output capacitor !

Regards, Rudolf.

[janneman]

rbroer,

I don't see any reason why it wouldn't work, except maybe the start-up. Did you in fact check that it does start reliably? And why can there be no cap?

Cheers, Jan Didden

[gromanswe]

That is alright.
You TL431, one of my favourites, low noise referens.
Op amp is not that important
as long it can handle currents.

Referens is feed by the regulated votage.
Good !

As long as you can get it to start up.
In such design, sometimes Circuit will not
start up.
That is because if referens is zero,
the output of OPamp wil be
zero, and the referns can not
get any current to rise up to 2,5V.

Otherwise very good design!

gromanswe

[rbroer]

janneman

It works, I tested it last night. I had some concerns about starting up, but I guess since the output won't go to negative rail (=gnda), but only as low as about 2V, startup shouldn't be a problem.
No output cap, current feedback opamp don't like driving capacitive loads.

I'm just curious if it is worth making tiny pcb's;
so, will it better the average LT1085 or TL431 in respects of noise, frequency response, dynamic behaviour ?

Regards, Rudolf.

[gromanswe]

I use a combination of a temp-stabil,
zener-shunt diode and some
Current Amplifying Transistors.
See picture at bottom.

Voltage References:
TS431 (TO-92) 1.24-6.00V 0.06-30mA, low current, min 60uA
TL431 (TO-92,DIL8,SO8) 2.5-36V 1-100mA, minimum 0.500mA
both 50ppm/C temp.
TL1431 (SO8) 2.5-36V 1-100mA 13ppm/C temp, precision.

A LINK to ST products, Referenses.....PDF download
http://www.st.com/stonline/books/
-----------------------------------------------------------
This is an interesting Circuit, used to reduce noise
from regulators with about 20dB (only 1/10 of noise remains)
http://www.wenzel.com/documents/finesse.html
------------------------------------------------------------

See Picture!

You really do not need an OP.
You can use 1 or 2 transistors, depending on
how much Current is drawn.

See to that TL431 gets enough current, 1 mA recommended.
Could manage on .5mA.
TS431 only needs 60uA.

gromanswe

[gromanswe]

gromanswe

[gromanswe]

Here is download of TL431 PDF.
http://rv6llh.rsuh.ru/PDF/TL431.pdf

Of the internal diagram, we can see that this is no diode.
It is a Circuit consisting of 11 transistors.
It has some of these arranged as a bandgap referens.
It is those down at left.
This compensates for tempchanges.
It also uses some small capacitances, to avoid oscillation,
and make it stable.

In PDF is also 13 different applications.
Circuits where TL431 is used in different ways.

gromanswe

[janneman]

Rudolf,

Yes, your asalysis of the start-up sounds correct. Re: output cap: I know a CFO are sensitive to cap loading (who isn't, these days...), but it is another matter with a cap of say 1uF or more. That should not make the circuit unstable. I think it is worthwhile to try it, it will compensate somewhat for the rising Zout with frequency.

Gromanswe,

I agree, a nifty circuit! But it is no match for Rudolfs' circuit. For one thing, your circuit has limited loop gain so inevitably the Zout will be larger leading to much larger output voltage variations with load variations. Also, the rejection of input noise & ripple will be worse. And Rudolfs' circuit can also sink current, which can be an advantage in some applications (high frequency inductive loads as in DACs).
But of course your circuit can source much higher current.


Cheers, Jan Didden

[tvi]

<b>ALW</b> posted this some time ago

"<i><font color=#800040>DIY Hi-Fi myths debunked - no 431.

Here's the noise spectra of a TL431, vs an LM317 and 2 samples of LT1086's.

All are set to the same o/p voltage and properly decoupled.

I'll let you guess which is which </font></i> "

<a href="http://www.diyaudio.com/forums/showthread.php?threadid=985&perpage=15&highlight=t l431s&pagenumber=2"><img src="http://www.diyaudio.com/forums/attachment.php?postid=17984" alt=""></a>

The pictures a link to the thread


James

[Aud_Mot]

Just a thought....

If some one wanted more current, (and maybe lower noise?) couldn't they piggy-back 2 or more amplifiers? Heat disapation might become a problem.

This would not increase the footprint but would increase the verticle dimension only a small amount.

Aud_Mot