PSU for High end audio without electrolytic capacitors - diyAudio
Go Back   Home > Forums > Blogs > Creative High end audio

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Version of my blog in English
Description and links to my russian language blog and google translating
Rating: 4 votes, 5.00 average.

PSU for High end audio without electrolytic capacitors

Posted 13th February 2010 at 06:18 PM by Nazar_lv
Updated 1st April 2014 at 10:45 PM by Nazar_lv

From article http://s-audio.systems/diy/psu-without-electrolytic-eng
----------------------------------------------

Electrolytic capacitors are constant part of power supply of audio devices, but in terms of sound he does not bring anything good, dispersion of sound quality-type of electrolytic capacitor is very large. Sonically the best types of electrolytic caps are Nichicon KZ, FG, Panasonic Pureism, Black Gate FK, NX, Elna Cerafine, Silmic II, Silmic,but the attempt to eliminate from PSU electrolytic capacitors was very successful, instead is used "electronic capacitor" (right side of the circuit).
It is global negative feedback free electronic substitutes, that has a low and stable impedance 0,1...0,2Ohm (equivalent to cap 40000u) in the band to megahertz shunting NP0 0,1 microfarads Murata GRM31 in close to chips, it completely detach regulator (317\337) from load, therefore we have the ideal transient response when load "changes", and dont have noise and output impedance modulation. Output capacitors 0.1-10u SMD NP0/X7R is needed for HF bypass (>1Mhz), caps on 317/337 out (not less than 1u) is needed for 317/337 stability. R1-C3, R9-C7 forms LPF that eliminates HF noise produced by swithching of diodes working on cap filter.

Calculation of this schematic is not difficult, according to the desired output voltage and maximum output current calculate values of resistors R4, R6 and R2, R10, respectively, from the formulas below

R4=240*(Uout-Ueb) / 1.25
R2=(1,25-Ueb) / (0.3/Uout)+Iload
where Uout-output voltage V, Ueb-emitter-base voltage VT1, VT4 V, Iload- load current А.

This structure of regulator-"electronic capacitor"(follower) provides in a comparison with the best series audio electrolytic capacitors significantly better results.

-------------------
original article is here
Article, schematic etc. for DIY use only, completely or partially may not be copied, reproduced or published elsewhere without author's permission
Attached Thumbnails
Click image for larger version

Name:	capacitorless.GIF
Views:	10700
Size:	24.8 KB
ID:	113  
Posted in Uncategorized
Views 13947 Comments 14 Email Blog Entry
Main     MASH »
Total Comments 14

Comments

  1. Old Comment
    peranders's Avatar
    Good point but this is nothing new. It's also two different situations, low current and high currents.
    permalink
    Posted 15th February 2010 at 02:50 PM by peranders peranders is offline
  2. Old Comment
    Nazar_lv's Avatar
    what are you talking about?
    permalink
    Posted 15th February 2010 at 03:05 PM by Nazar_lv Nazar_lv is offline
  3. Old Comment
    peranders's Avatar
    Using a regulator is quite normal if the current is rather low and the purpose is to create a low output impedance in an artificial way using feedback and active devices. An another thing is to increase the PSSR. You have maybe heard of the super regulators and this approach make it possible (with fast devices) to get an extremely low output impedance high up in frequency.
    permalink
    Posted 15th February 2010 at 03:16 PM by peranders peranders is offline
  4. Old Comment
    Nazar_lv's Avatar
    All NFB regulators have inductive output impedance, placing electrolytic capacitor especially low-esr types creates a resonanses in output impedance (if the is yet a film or ceramic capacitor we will have more and stronger resonances). Impulse response to change the load current especially with low noise regs (M5230, M5290 etc) is very bad (one solution to this problem is use after regulator CRC filter,but it has its drawbacks), and output impedance and noise is a function of load current, as load changes noise and output impedance will be correlated with load changes (correlated to our music).

    Proposed schematic completely eliminates all problems with regulators and capacitors, and have stable resistive output impedance at frequencies up to 500-1000Khz (then the capacitive), stability of the output impedance at audio frequencies is a very important thing for sound quality and tonal balance.
    permalink
    Posted 15th February 2010 at 05:10 PM by Nazar_lv Nazar_lv is offline
    Updated 15th February 2010 at 05:13 PM by Nazar_lv
  5. Old Comment
    peranders's Avatar
    The LM317/337 has negative feedback so I don't follow you completely.
    permalink
    Posted 16th February 2010 at 01:45 PM by peranders peranders is offline
  6. Old Comment
    Nazar_lv's Avatar
    of course LM317/337 has NFB, but load is not connected to 317 read carefully
    permalink
    Posted 16th February 2010 at 03:25 PM by Nazar_lv Nazar_lv is offline
  7. Old Comment
    peranders's Avatar
    Why do you insert a resistor and then increase the output impedance on purpose? Personally I prefer to have a low output impedance from deep bass and through the whole audio band up to MHz.
    permalink
    Posted 16th February 2010 at 03:33 PM by peranders peranders is offline
  8. Old Comment
    Nazar_lv's Avatar
    that resistor(R2,R10) and 317\337 has no relation to output impedance.
    In real world output impedance of power supply should not be above 5 ohms, 0,2ohm in this design is 20 times lower, it is important that the impedance was stable in the audio band.
    In this scheme is very simple to reduce the output impedance from 0,2ohm to 0,05ohm (also not introducing a global negative feedback)
    permalink
    Posted 16th February 2010 at 03:56 PM by Nazar_lv Nazar_lv is offline
    Updated 16th February 2010 at 04:01 PM by Nazar_lv
  9. Old Comment
    umut1001's Avatar
    Hi.I have 4.7uf 40v ero 5mm mkt1826 4pcs of polyester capacitors.Can i use them?Is it good to use them instead of 100nf caps?Thanks
    permalink
    Posted 22nd October 2011 at 06:10 PM by umut1001 umut1001 is offline
  10. Old Comment
    Nazar_lv's Avatar
    no, 100n is 1206 NP0 close to chips
    permalink
    Posted 23rd October 2011 at 10:53 AM by Nazar_lv Nazar_lv is offline
  11. Old Comment
    Hi Nazar,
    Maj I use these - ECHU1H104JC9 PPS Film Chip Capacitor 0.1uF @

    Regards,
    Ignat

    permalink
    Posted 25th January 2012 at 05:52 PM by niamex niamex is offline
  12. Old Comment
    PCB Power Amplifier try RENKUS HEINZ-model P-1500
    permalink
    Posted 17th November 2012 at 10:53 PM by ricardo rocha ricardo rocha is offline
  13. Old Comment
    afshindiy's Avatar
    Hi
    I think elna silmic is the best capacitor for coupling and power without feedback and Elena silmic II

    Gives you a real voice ... Even for very good DAC
    The filtered voltage + wima mks4 + 10nf ceramic cap...

    tanks
    permalink
    Posted 13th August 2013 at 07:26 AM by afshindiy afshindiy is online now
  14. Old Comment
    Hi Nazar!

    What is the value of C3 and C7?


    Cheers!
    permalink
    Posted 19th April 2014 at 03:09 PM by flyboi flyboi is offline
 
Hide this!Advertise here!

New To Site? Need Help?

All times are GMT. The time now is 10:09 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2