Modern 317/337 alternatives?

[maxw]

Anybody used the LT1964 and LT1761 pair?

They both have lower noise than the LT1963/LT3015, they are pretty cheap and they both come in 5-Lead TSOT-23 package which, although small, should be easy enough to solder as it has 1mm lead spacing.

The pair would only do 100mA on each rail but that would be enough for most preamp/headphone opamp applications?

 

[luvdunhill]

You likely want to add 10 nF capacitors from the ADJ pins to ground: Linear's data sheets indicate noise output is lower and transient response better with those in the circuit.

Where did you note this?

I see the reference to the use of a feedfoward capacitor, Cff, but not an ADJ pin bypass. I was looking at the LT3015.

 

[maxw]

Where did you note this?

I can't see this mentioned anywhere in the Linear datasheets either. Perhaps DSP_Geek was referring to C2 or Cadj mentioned in LM3x7 datasheets.

 

[luvdunhill]

The regulators which just stood out above all else were the Jung, Sjostrom, Burson, NCD and Linear Tech LT1963A/LT3015. They were in a class all by themselves. The LM317/337 were near the bottom.

many LM317 layouts have protection diodes across in and out pins, as well as from out to adj. Does adding these diodes have any negative consequences to the performance of LT1963A/LT3015? Do these parts already have protections against the failure modes that these diodes serve to prevent?

 

[dirkwright]

The LM140L is an excellent regulator on paper at least.
http://www.ti.com/lit/ds/symlink/lm140l.pdf

 

[jackinnj]

many LM317 layouts have protection diodes across in and out pins, as well as from out to adj. Does adding these diodes have any negative consequences to the performance of LT1963A/LT3015? Do these parts already have protections against the failure modes that these diodes serve to prevent?

Belt and suspenders, condom and abstinence.

 

[dirkwright]

Of course, the real answer is just design your circuit to have a decent PSRR. They you don't have to worry about the regulator so much.

 

[dsolodov]

The protection diode is not required for LT1963 as per the datasheet.

I did prototype PCB for LT1963 based PSU. Other folks on amb.org measured the noise, and it was acceptable till ~.5A. Then ripple noise starts to propagate, so the conclusion was LT 1963 is a great pre-regulator.