TL431, LM431, KA431, TL1431 like stability

That is the difference between Mill spec. and home entertainment components.

The price also reflects that, too.

P.S.

When designing for critical operations or customers, you normally fully test each batch and cat it. Also, make previsions in the design and assemblies to compensate for each unit in the final test.

Jamie

That's a good reminder that just because two chips have the same part number doesn't mean they behave the same way, especially under unusual conditions.

The TL431 is pretty slow, so ESL shouldn't be a big worry. (You're going to be using SMT ceramics, I assume.) The simple method is to bomb it flat with a 10 uF X7R ceramic and not worry about it any more. (The 431 is a nasty noisy beast anyway, so filtering hard is all to the good.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

On 29/04/12 15.14, Phil Hobbs wrote: ...

Thanks for your answers.

I have compared TL431 with some other ballast-regulators, but they do not have much lower noise?

How low "can" the noise reasonably be?

How low will a "standard" ca.0.4W 9.1V zener diode at 10mA e.g. be for comparison?

Is this the state of the art ballast-regulators regarding long-term stability (pretty expensive)? No noise level is given as far as I can read: LT1019A 5V:

LT1019A 2.5V:

Quote from datasheet: "... If output bypassing is desired to reduce high frequency output impedance, keep in mind that loop phase margin is significantly reduced for output capacitors between 500pF and 1?F if the capacitor has low ESR (Effective Series Resistance). This can make the output ?ring? with transient loads. The best transient load response is obtained by deliberately adding a resistor to increase ESR as shown in Figure 1. ..."

-

I have found this noise cancelling circuit, but it will ruin the voltage stability:

Finesse Voltage Regulator Noise!:

br

Glenn

If you don't mind spending a bunch of supply current, as in the shunt post-filter case, you can just use a better voltage reference such as an LM329, plus an op amp and a transistor. That's what's inside the 431, except that the reference is a bandgap, and those are always noisy. They work by adding a diode drop to a scaled version of the deltaV_BE between two transistors running at different current densities, and to make it work the scaling has to be about a gain of 10. Thus in round figures, even a good bandgap is at least 20 dB noisier than a V_BE drop.

There are tricks to getting reference noise down to lower levels, e.g. charging up a cap to the reference level between measurements and using a buffered copy of that voltage instead of the jiggly reference. You still have the scatter between baselines, but that can often be removed afterwards.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net