TL431 comparison

[borges]

I just saw that Diodes is out with a TL431. In an earlier post, ONsemi's TL431 took the low-noise throne. But judging from the data sheets, the Diodes part may be even better at higher frequencies. What is your interpretation of this?

ONsemi: figure 13, p. 6 on
http://www.onsemi.com/pub_link/Collateral/TL431-D.PDF

Diodes: p. 6 on
http://www.diodes.com/datasheets/TL431_432.pdf

True, the Diodes part has way more noise at lower frequencies but much less at higher. The two curves seem to meet at around 50Hz. But with decent decoupling, which regulator should be the least noisy? I'm thinking at least 10uF in X7R close to the intersection of load and regulator, and then some 0603 NP0s close to the load.

I'm interested in using more shunts. One good thing (at least for me) is that the Diodes part comes in a smaller SMD package than the SOIC-8 from ONsemi. I'm lucky enough to have access to decent soldering tools.


Cheers,
Borge

[ekaerin]

Borge,

Depending on your application Higher frequency noise might be easier
to attenuate. With some stray inductance and lower capacitance.

In any case 10uF fairly close to the 431 should keep it stable.

My experience is that if you are using it as a "global" ref for many circuits the stray cap may cause instability. So if there are no stringent requirement for fast start up, 10uF definitely keeps it quiet.

regards / Mattias

[ekaerin]

When speaking about the 431. Sometimes you want to use it at higher voltage than the maximum nominal allowable cathode voltage
of 36V. All you need to do is cascode it. Beware of the collector leakage if used as a comparator though.

/ Mattias

[peranders]

The noise curve of the Diode 431 looks very suspicious. Zero noise at 1 kHz?

[zapnspark]

Here's a circuit suggestion you wont find in the data sheets:
http://tinyurl.com/24tzuet
R2 and R3 are larger values than is typical for the TL431.
The LM317L is used as a constant current source.
C2 is the part that reduces noise, both the flicker noise from the TL431 and the noise from the LM317L current source.
A quiet voltage supply for some audio applications?
Looks good on the my bench check of the prototype.
I'm just throwing out the idea.

Cheers.

ZAP

[borges]

Originally Posted by peranders

The noise curve of the Diode 431 looks very suspicious. Zero noise at 1 kHz?

That's what I thought, too. Maybe they have fitted it with a large capacitor before doing the measurements?

Borge

[borges]

Originally Posted by zapnspark

Here's a circuit suggestion you wont find in the data sheets:
http://tinyurl.com/24tzuet

Thanks! I'm inspired by the ideas here to make a local power regulator for a VCXO:

The Extremist DAC

I'll make sure there is room to decouple the upper resistor to the output voltage.

Speaking of decoupling, how do you guys feel about using a 10uF X7R 0805 in paralell with the largest possible NP0 0603? No fancy logos, just pure ceramic caps in small packages.

Borge