yes, another jung regulator, i know

[geekysuavo]

hello all!

i know the forum has quite a history with the jung super regulator, but i just wanted to see if anybody out there had any helpful comments on my workup. i've read the articles by walt jung and jan didden describing the super regulator and its specifics, as well as an article regarding ccs topologies by mr. jung and the majority of all the posts here regarding the regulator.

attached is my 'take' on the jung-didden (audio amateur ca. 2000) super regulator. i plan to pair it with a negative reg. for +/-15VDC supply for opamp projects that would current-limit around 500mA. the only changes are:
a.) increased pre-regulator voltage to 3.125V.
b.) changed ccs to referenced jfet-cascoded topology.
c.) target output of +15VDC instead of +14VDC.

my (probably n00b-ish) questions follow:
1.) are there any suitable BJTs with 500 > hFE > 150 in a TO-220 package that could replace the D44H11/D45H11? something around hFE=250?
2.) relating to 1, the 2N6284 (complementary PNP is 2N6287) is a TO-3 NPN darlington with an hFE=250. is there any reason it wouldn't satisfy? (speed?)

enjoy,
~ brad.

[geekysuavo]

well, i've modified the design further and worked up a negative compliment as well. the two regulators put out 12V with minimal ripple, as i've proven to myself with spice simulation of the positive supply. as the application of this circuit has changed from variable power supply (no guarantee of stability under complete range of voltages, not to mention pass transistor dissipation!) to a power supply for an opamp audio preamplifier, i'm going to use the D44H series of transistor.

attached are the schematics and layouts. the pcbs have been designed for single-sided application and each have only one wire jumper (red trace).

i should have some time to prototype these things here soon. let me know if you find an error or have a question or suggestion. thanks!

~ brad.

[iko]

There are indeed many "super" regulators to choose from. I am wondering if you made changes to the original design because of parts you had on hand, or because you've improved on it (if so, what are the improvements).

It would help to show some of your sim results, such as psrr, output impedance, transient response, because it would give people an idea how the circuit behaves, at least theoretically.

[aparatusonitus]

IMHO you are going in wrong direction... What you should do is:
- change LM336 to LED, you will get lower noise voltage reference
- change 2k trimer for resistor
- loose 2N5462 cascode and use it instead of 15k resistor as CCS
- or split the 15k resistor in half (7k5+7k5), put an electrolytic cap from middle point to pass transistor collector, you will get higher PSRR in both cases.

[geekysuavo]

hello,

Originally Posted by aparatusonitus

- change LM336 to LED, you will get lower noise voltage reference
- loose 2N5462 cascode and use it instead of 15k resistor as CCS

actually, according to jung's publication on audio current regulators in audioXpress (2007), the noise rejection of the jfet cascoded current source was at or near the noise floor (-140dB) of the test setup, whereas the led-referenced source was at only about -100dB. i'm not sure why you'd ask me to go back...

Originally Posted by aparatusonitus

- change 2k trimer for resistor
- or split the 15k resistor in half (7k5+7k5), put an electrolytic cap from middle point to pass transistor collector, you will get higher PSRR in both cases.

you're looking at the old "drawing" schematic. i posted pdf versions of the new schematics (with fixed resistors throughout) in my second post.

~ brad.

[aparatusonitus]

I've tried to find a better usage of J-fet without increasing the parts number...you can of course keep the cascoded CCS if you like.

[peranders]

The drawback of the voltage reference and the JFET is increased dropout voltage which already rather high.

When it comes to the pcb layout you should be aware of the grounding scheme. Your implementation could lead to instability. You should check the schematics Walt Jung has done and as much as possible let this be the reality. Try to achieve a star ground as much as possible. Try also to have the opamp as close as possible to the output transistor. "janneman" will surely be able to fill in the rest since he did the pcb for Walt's article.

[geekysuavo]

thanks for the tips peranders!

i think i can live with a bit more dropout voltage in the name of better noise rejection...

as for the layout, i'll work on a more stable arrangement. are there any other recommendations that should be made? i'm up for all the criticisms you've got! i would very much like it if jan could give his critiques as well!

thanks,
~ brad.

[jackinnj]

Originally Posted by geekysuavo

actually, according to jung's publication on audio current regulators in audioXpress (2007), the noise rejection of the jfet cascoded current source was at or near the noise floor (-140dB) of the test setup, whereas the led-referenced source was at only about -100dB. i'm not sure why you'd ask me to go back...

The two articles are available on the AX website. (www.audioxpress.com) They're "must read". WJ's articles can be found on his website and on the EDN site. Boards for the original super-regs are available from Old Colony.

As cited, layout is critical, and, if you are going this route you might want to read up on the original 1995 articles and addenda where some bypass is used to lower the odds of oscillation.