Lm317 tracking prereg phase issue

[SonocusEric]

Ive built this circuit and notice improcement in noise but also that 1uf + r filter affects the sound greatly and is not an optimal value for my taste. 1uf+ 1ohm seems to affect the high frequency phase too much to where imaging is collapsed and highs sound 'slow'. Ive tinkered with the values somewhat but so far havent found one thats optimal. How should i tweak this filter?

 

[SonocusEric]

 

[YashN]

I haven't yet tested this configuration as I am still exploring a simpler one (single LM317 in my case, with variations of additional circuitry for lower noise floor and other characteristics), but here are a couple of things you can do:

- Read jbau's thread on the LM317

- Try to sim it so that you can have a good idea for trends with changing the values

Hopefully someone who has implemented it can provide you with more information.

 

[SonocusEric]

i see the first 317 as basically ccs with feedback. please correct me if im wrong.

 

[martin clark]

Did you actually read my article you took the image from?
The first voltage regulator attempts to hold a constant voltage across the second.
As with any LM317-based regulator, you must ensure a minimum load on the output (10mA recommended, 25mA-50mA or more = better)
You have a wide range in which to experiment with R-C2, but the purpose of appropriate R is to ensure that IC1 does not break into oscillation. Guidance on these values is given on part 3.

Try larger values for C2 - 10uF or more....

 

[YashN]

Indeed, Martin's series of articles on the LM317 is a must-read.

 

[timgrindell]

LT1084

Martin,
Do you have any thoughts on using a LT1084 for the 2nd reg (assuming appropriate caps are used)?
I read that they are much "better".
thanks
tim