In renovating the psu for my preamp ive decided to try the cap mx filter post lm317 337 regulation. The gain stage is current hungry class a so the high impedance of the capmx doesnt bother me much.
My question for the moment is what load cap for the lm317 is recommended in this situation. I feel that 317 provides all the lf filtering i need and the capmx is for the hf noise filtering. For that shouldnt the load cap for the 317 337 be 1uf/10uf recommended by the datasheet? How would this affect the input bjt of the capmx? Any stability issues? Would like to know your thoughts. Thanks
My question for the moment is what load cap for the lm317 is recommended in this situation. I feel that 317 provides all the lf filtering i need and the capmx is for the hf noise filtering. For that shouldnt the load cap for the 317 337 be 1uf/10uf recommended by the datasheet? How would this affect the input bjt of the capmx? Any stability issues? Would like to know your thoughts. Thanks
100nf input capacitor and 1uF output capacitor according to datasheet.
I tend to use 47uf on output and 10uf on input too.
Depends how noisy the environment is.
I tend to use 47uf on output and 10uf on input too.
Depends how noisy the environment is.
I use Nichicon upw 10uF on the outputs. I have seen going higher in cap esp using a low esr can cause instability, adding some small series r can solve the stability issue.
Yup i do use sense resistor 20mOhm for the lm regs. What im curious about is how happy the cap multiplier will be with whatever makes the 317 happy
lower noise with LM regs?
did you try LED for reference instead of resistor?
http://www.acoustica.org.uk/t/3pin_reg_notes1.html
did you try LED for reference instead of resistor?
http://www.acoustica.org.uk/t/3pin_reg_notes1.html
Stability of the LM317 is a function of the output capacitor value (and its ESR) AND the load being drawn. Below a few hundred mA you need a much higher value. Both the OnSemi and TexInstruments datasheets don't mention this, but it is easily demonstrated. (The TexInstruments data sheet graph on the effect of Cadj is also incorrect.)
Using Cadj lowers noise by ~15dB. Optimal value is 10uF, above this value the transient response of the regulator suffers.
Best noise reduction is using Elvee's de-noizator https://www.diyaudio.com/community/...-retrofit-upgrade-any-317-based-v-reg.331491/
Graphs below are from Omicron Labs, demonstrate stability under light DC load (25mA)
Attachments
According to late analog great Bob Pease; 10uf as CAdj & a general purpose 47-100uf electro at output works best with lm3xx.