TL431 pos & neg regulator

[Lovan]

Hello everybody

I'm trying to build a +/- 15V regulator using TL431 but I feel kind of clueless about the negative regulator. The regulator will be driving a JLH headphone amp(circuit 2 at http://www.tcaas.btinternet.co.uk/jlhphones.htm )

My idea is something like this for the positive supply


and something like this for the negative supply


I only have a trafo that will give +-34V regulated at the moment if your wondering about the high supply voltage.
I'm biasing the tl431 with about 10mA since that should be adequate.

Any thoughts?

Thank you for reading this

/Lovan

[peranders]

It might work. One way to find out is to simulate it. You could use the freeware SwitcherCAD (LTSpice) from www.linear.com

[Eva]

The Vbe of the output transistor will be effectively substracted from the 2.5V reference so that negative arrangement is not practical.

You may connect the 5K resistor to the base instead of the emitter and use the TL431 as a plain zener, though. Otherwise, consider a LM337 1.5A negative regulator.

[Lovan]

PA: I have never understood how Spice works but I will try it out sometime.

Eva: Thank you for your reply. If I use a tl431 as a zener how will it compare to a regular zener? better or worse?
Maybe I will just start with zenerregulators for both positive and negative.

I thought that a passregulator would be better than a lm337/317 since I will draw almost constant current?

[peranders]

Eva, you are right here.

You might take a peek at Sonny Andersen's current feedback amp. He uses LM431 for negative voltages.

www.mirand.dk

[Eva]

If I remember properly, TL431 is substantially noisier than any zener above 10V (low voltage zeners are noisier too). However, output impedance is much lower for the TL431 thus yielding much better line and load regulation. Also, TL431 has very good frequency response despite its active nature, producing a low impedance ouput up to at least 100Khz.

If your input voltage does not sag too much and your load current is low, you may consider the TL431 as a plain shunt regulator in each rail (see the datashet if you don't understand). That's a simple yet effective solution.

Consider also the LM317/LM337 pair. These regulators are widely available and perform quite well for their low cost.

[peranders]

Originally posted by Lovan
PA: I have never understood how Spice works but I will try it out sometime.

Draw a schematic. Add a signal generator. Probe the voltages where you are interested. Boom, there you have it. A fast PC is nice to have. My 2 GHz PC has to sweat a bit......

Many people here know how to use this particular software so why don't you start to simulate.

[Lovan]

PA: I had a quick look at his schematic. It's interesting. I will have to think about how it works for a moment. I'm quite slow I know

Eva: Thank you again. I think that I will try tl431 as zeners to begin with.
I thought about the shunt but I think I would need to use it with a transistor becuase of the current needed.

I will try the LM pairs too. Thank you very much for your inputs.

[peranders]

Originally posted by Eva
Consider also the LM317/LM337 pair. These regulators are widely available and perform quite well for their low cost.

This is proably a better solution for you becuase you don't need the precision from a LM431. With LM317 you'll get a less noisy (probably) PS and you have better cooling and mounting and also more protection.

[AndrewT]

Hi,

I thought that a passregulator would be better than a lm337/317 since I will draw almost constant current?

This JLH circuit and most ClassA single ended topologies are not constant current.
All push pull ClassA topologies are not constant current.