Dual tracking PSU regulator from H&H

I've always wanted to build my own dual track PSU, & came across this design in Horowitz & Hill recently. I'd like it to go to +-30volts & probably use a NE5534 instead of the TL071.

Is it good enough for quiet audio work? Is it really this easy?

Any mods to improve the performance?



Many thanks
Peter

I've always wanted to build my own dual track PSU, & came across this design in Horowitz & Hill recently. I'd like it to go to +-30volts & probably use a NE5534 instead of the TL071.

Click to expand...

This regulator requires an op-amp with a common-mode input voltage range that includes its positive supply rail. A TL071 will just cope but not a NE5534.

Any mods to improve the performance?

Click to expand...

Dunno about performance, but I'd include a pair of diodes in anti-parallel from the negative input of the op-amp to ground. Oh, and a resistor/ferrite bead in the base line of Q1.

It's not going to be hyperquiet, but you'll likely outperform a 78xx/79xx combo.

R5 is going to run pretty hot.

JDB.
[btw, in which edition of H&H did you find this regulator ?]

http://www.groupdiy.com/index.php?topic=16039&highlight=discrete+layout

?

:green:

What JDB noticed. Most opamps will need like a +3V rail to do that center-point comparison.

Q1 has no short protection. May very well quit the first time it starts-up into C3.

What is R5 for?

And finally: what the heck do you need a tracking regulator for, in this day and age? Set a pair of regs for +31V and -29V, most audio gear should work happy.

Maybe he wants a variable bipolar supply for the bench...

Hi Peter,

Dunno if you've seen this page yet:

You actually put me onto the Jung Super Reg circuit and here's a guy who's mfging boards:

http://home.swipnet.se/~w-50719/hifi/

I bought PCBs for 2x JSR03 and 2x JSR05 and I recived them pretty quick.
Per-Anders is good to deal with, would recommend.

JDB. [btw, in which edition of H&H did you find this regulator ?]

Click to expand...

2nd Edition (Cambridge edition fig 6-34). Actually, now that I look back at the actual schem, it has a different section for the positive side of the regulator using a voltage ref IC. I subbed in the 317 section....

what the heck do you need a tracking regulator for,

Click to expand...

It is for a bench supply. [go Dave!]

At the moment I have a LM317/337 combo switched with 2 by 6 way switches (coarse & fine) but I am a bit anal about this so I wanted to convert the supply to a tracking design & have it continuously variable.

What JDB noticed. Most opamps will need like a +3V rail to do that center-point comparison.

Click to expand...

I dont understand this part. Where would that rail go in the above circuit?

Thanks all
Peter

they mean input voltage 3v over the intended output voltage of the IC regulator.

[quote author="peterc"]

What JDB noticed. Most opamps will need like a +3V rail to do that center-point comparison.

Click to expand...

I dont understand this part. Where would that rail go in the above circuit?
[/quote]

The positive supply of the opamp.

In the circuit you sketched, the positive supply node of the opamp is hooked to ground. However, the common mode voltage on its input is also at ground. Many opamps (like the 5534) need to have their inputs a few V below their positive supply connection.

JDB.
[EDIT: Embarrassing brain fart deleted. One of these days, I need to learn the difference between NPN and PNP]

Thanks for the reply JD.

So if I went to a Darlington for the negative output pass transistor (instead of the 2N2907) I could use the TL071?

What is the LED-with-resistor trick?

Peter

[quote author="peterc"]So if I went to a Darlington for the negative output pass transistor (instead of the 2N2907) I could use the TL071?
[/quote]
No, that was a gross misreading of the schematic on my part. I thought the opamp had to source current to the base of Q1, not seeing that Q1 is a PNP emitter follower . I edited my post about a minute after submitting hoping noone would see, but you were too fast for me.

Even with a TL071 I would suggest using a positive supply a few V above ground.
[quote author="peterc"]What is the LED-with-resistor trick?[/quote]
Using a LED (or a low voltage zener diode) and a resistor (and some capacitive decoupling) as a low voltage shunt regulator. LED or zener tied to ground, resistor between the LED and +Volts, opamp positive supply from the LED-resistor node.

JDB
[too tired to make much sense ATM]

> It is for a bench supply

An unprotected bench supply is a dead bench supply.

If you want a supply that is blowout-resistant, you can't beat the integrated regulator chips, since they can feel themselves melting. Can't do that in discrete.

What are your real specs? +/-30V is awful close to the +/-35V rating on many processes. And +/-0V looks good in the specs, but isn't useful in high-level audio. And if you are testing little-teeny audio, best to build a separate +/-few-V supply than to waste-away 90% of the power.

This will trim and track from +/-5V to over +/-30V:



There is little point in a "sexy" opamp here; I'd use TL072 because they're everywhere. The -3V supply only needs to supply a few mA for the opamp and the +reg. It can be a 2K 2W resistor from the raw -40V to a 3V Zener.

Hi PRR

An unprotected bench supply is a dead bench supply.

Click to expand...

I had put in diodes, similar to your drawing, & also between the inputs & outputs of the 317 & 2N2955.

My needs are to provide supply voltages from between +- 6v & +-28v for audio work. So approx 200mA max per rail. Nothing too out of the ordinary.

I'll give your circuit a bash over the weekend, thanks for all the effort.

Peter