Low cost regulator in between Jung and Flea

[00940]

The idea about this reg came in another thread, with a misleading title (Zener regulator with "integrated" pre-regulator ), so I'll post it here.

As said in the title, the sources of inspiration are the Jung regulators article and the "Flea" regulator designed by Martin Clark on PFM. From Jung, I took the ccs for the pass transistor. From Martin Clark, the idea of using a 3pins reg as ccs for the voltage reference.

If anyone sees obvious errors or improvements (keeping in mind that it's supposed to stay cheaper than a full-blown Jung), I'd be glad to hear them before breadbording the reg.

[AndrewT]

Hi,
what would happen if you doubled up D1 or replaced it with a LED?

Is 5.4mA passing R6 enough?
What about some noise attenuation cap across the Zener?
and a big cap on the adj pin?

[00940]

Originally posted by AndrewT
what would happen if you doubled up D1 or replaced it with a LED

The output of the opamp is pushed further away of the supply. So, what comes to mind:
- swap D1 and D2 to reduce the dropout by around 1V but then the input voltage from the reference would be close to the rail, certainly too close (2V).
- or replace D1 by a led and the ne5534 will work in a safer area.

Is 5.4mA passing R6 enough?

Well, it depends on the hfe of the bd139 used. It must be adjusted in order to limit the current that can pass through the pass transistor. I'd set the limit at around 300ma, for a max dissipation of 1W. One could simply set the resistor according to the "worst case scenario". For a bd139-6, the worst beta is 40, according to the datasheet, so you need around 8ma.

What about some noise attenuation cap across the Zener? and a big cap on the adj pin?

The zener's noise is attenuated where it matters most by the RC filter at the input of the ne5534.

A big cap from adjust to ground would certainly improve the performance of the prereg, which is good for the error amp. But then, I need to add protection diodes, to protect the lm317.


So, what needs to be changed:

- change D1 for a led,
- adjust R6 according to the bd139 at hand,
- add a cap at the adjust pin (100uf being ok),
- add a pair of protection diodes around the lm317 (1n4001).

[janneman]

Originally posted by 00940
[snip]- or replace D1 by a led and the ne5534 will work in a safer area.
[snip]

... and then you can supply the opamp from the regulated output for even better performance.

Jan Didden

[lineup]

Originally posted by janneman
... and then you can supply the opamp
from the regulated output

This is of course the ultimate way.
And one of the main features of Jung Super Regulator.

In this schematic, originally drawn for keantoken in this topic:
http://www.diyaudio.com/forums/showt...99#post1551799
.. I use the output to fead the REFERENCE Red LED.

This is a great improvement to many other ways
Feeding the complete op-amp, discrete or chip op,
from the output is even better, of course!

[00940]

Originally posted by janneman
... and then you can supply the opamp from the regulated output for even better performance.

Jan Didden

I'm affraid it won't work without major changes. Since the opamp is working at unity gain, the reference voltage would be equal to the output voltage.

Now, it might work if :
- I use the 5534 with some gain (which means another extra cap to reduce ac gain at unity).
- change D1 for a zener (and I probably must bypass it).
- change D4 for something like a 6.2V zener and D2 by a 12V zener.

I'm affraid of startup problems with such a configuration

[janneman]

Originally posted by 00940


I'm affraid it won't work without major changes. Since the opamp is working at unity gain, the reference voltage would be equal to the output voltage.

Now, it might work if :
- I use the 5534 with some gain (which means another extra cap to reduce ac gain at unity).
- change D1 for a zener (and I probably must bypass it).
- change D4 for something like a 6.2V zener and D2 by a 12V zener.

I'm affraid of startup problems with such a configuration

No it isn't working unity gain. If you replace D1 with a zener of about half the Vout, the opamp output will sit neatly halfway the supply. Not often that you get a chance to increase performance for essentially zero cost

Start-up *may* be an issue, but that can be cured easily. It's worth a try.

Edit: Even if you don't supply the opamp from the regulated output, it is still advisable to bias the opamp output midway of the supply voltage.

Jan Didden

[00940]

Please forgive me but I wasn't clear again. I had not in mind the output voltage of the opamp, but of the regulator. What happen if my opamp is powered by 16.7V and the reference voltage is also 16.7V ??

The answer of LTspice is : it doesn't work (unless doing the changes described above).


Btw, you're right to correct me, the opamp isn't at unity I just meant I'm feeding the inverting input with the whole output voltage and the positive with a high voltage reference.

[00940]

A schematic is better than a lot of words. Those two seem to work in the simulator.

[AndrewT]

I prefer your new circuit.
But change R7 to 510r.