The simplistic Salas low voltage shunt regulator - Page 581 - diyAudio
Go Back   Home > Forums > Amplifiers > Power Supplies

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
Reply
 
Thread Tools Search this Thread
Old 25th March 2013, 07:57 AM   #5801
diyAudio Member
 
Kazuo Ozawa's Avatar
 
Join Date: May 2006
Location: Japan
Default Reflektor +3.3V

Hi, Reflektor lovers.
For last three months, I have tried to optimize the Reflektor, and found some useful knowledge.

1) M2
I guess that lesser the Cxxx of M2, better becomes the sound.
For this reason I replaced STP55NF06L with MTP3055VL.

2) R5
R5 acts important factor. This determines the current of Wilson Current Mirror. And M2 is driven by this. After some tests, 500R is preferred.
I realized 500R by paralleling two 1K VARs.

3) Zobel
As for C1, we must take into account two factors, that is, value and quality. When either factor was missed, I didn't satisfy.
As a result of a lot of testing, I adopt 33nF capacitor made of mica.

Outcome is very remarkable.
IMO there is no peak, no valley no noise but is open, smooth, and clear field.
I am noticed that my previous Reflektor had some weak points comparing to new one.
There is only real music playing soulful. I am satisfied very much.
It may be strange, I feel the volume has become larger than before, approximately 3dB or more.

Again I would like to say that Reflektor is excellent.
Attached Images
File Type: png Reflektor +3.3V 10th.png (34.0 KB, 468 views)
File Type: jpg top+3.3V.jpg (104.1 KB, 461 views)
File Type: jpg bottom+3.3V.jpg (144.3 KB, 421 views)
  Reply With Quote
Old 25th March 2013, 08:35 AM   #5802
canvas is offline canvas  Taiwan
diyAudio Member
 
Join Date: Nov 2003
Location: Taiwan
Hi Kazuo-san,

Thanks for sharing this. I've noticed you got 2 820uf Nichicon LE polymer caps on Vref. Have you tried others? (Polymer caps usually have high leakage current which result in more noise on Vref) For the zobel, I got good result with C0G/NP0 ceramic, but I like your idea of silver mica. I guess HF is important for digital power here.
  Reply With Quote
Old 25th March 2013, 09:39 AM   #5803
diyAudio Member
 
pietjers's Avatar
 
Join Date: Oct 2006
Intersting stuff Kazuo! Thank you for sharing!

I replaced the potentiometer / led (or diode) combination in the reference leg by a voltage reference diode from the LM4040 range. I did this in order to make the reference leg as low-ohmic as possible. In that way the Reflektor will give the largest possible reaction in the regulation leg, and doesnít only rely on capacitor impedance. The disadvantage is that you can not trim the output voltage exactly (unless you add a small pot or resistor anyway), but have to search for the best combination of available reference diodes, taking into account that the two transistors add 1.2V to the output voltage. In this case for 3.3V output voltage you would need a 2.1V reference diode, or a single green LED could do the job to start with. I bypassed this with 2x 1000uF/25V Nichicon Muse KZ capacitors, which are the best I have tried so far. I have bypassed these Nichicons with foil capacitors with small resistors in series, like a zobel circuit (resistors are a few Ohms or less). Experiment with different resistor values. This gives the opportunity to voice it and get the best sonic results. I use multiple bypasses ranging from bigger to small. The results are amazing.

For the zobel across the shunt, I measure with an oscilloscope the output for stability. As long as that there are no oscillations I lower the capacitor value to as small as possible. My application remained stable with no zobel at all. So I donít have a zobel here. But I donít use the original CCS, because of stability issues I ran into. Others seem to have no problems, but I switched to a different CCS design.

Peter
  Reply With Quote
Old 25th March 2013, 09:40 AM   #5804
Salas is offline Salas  Greece
diyAudio Chief Moderator
 
Salas's Avatar
 
Join Date: Oct 2002
Location: Athens-Greece
I will try Kazuo-san's tunings to see if they are preferred on analog circuits too sometime.
We thank him for his love and feedback on fine tuning for his gear powered with the Reflektor.
  Reply With Quote
Old 25th March 2013, 09:47 AM   #5805
Salas is offline Salas  Greece
diyAudio Chief Moderator
 
Salas's Avatar
 
Join Date: Oct 2002
Location: Athens-Greece
Peter there are many CCS to have like depletion cascode from SSHV2 or the LEDS one from 1.1 etc. but your problems could just be needing a faster PNP control BJT in the simple Reflektor CCS plus some input decoupling? Or you use some active pre regulation that interacted?
  Reply With Quote
Old 25th March 2013, 11:44 AM   #5806
diyAudio Member
 
Kazuo Ozawa's Avatar
 
Join Date: May 2006
Location: Japan
Quote:
Originally Posted by canvas View Post
Hi Kazuo-san,

Thanks for sharing this. I've noticed you got 2 820uf Nichicon LE polymer caps on Vref. Have you tried others? (Polymer caps usually have high leakage current which result in more noise on Vref) For the zobel, I got good result with C0G/NP0 ceramic, but I like your idea of silver mica. I guess HF is important for digital power here.
Thank for your sincerely advice.
In fact I am using BG caps on +5.0V Reflektor. But I did not test difference of changing caps.
It may be worth to try your idea.

Thanks
  Reply With Quote
Old 25th March 2013, 11:53 AM   #5807
RCruz is offline RCruz  Switzerland
diyAudio Member
 
RCruz's Avatar
 
Join Date: Sep 2006
Location: Wallis
Blog Entries: 1
Thank you for your contribution Kazuo
__________________
RC
  Reply With Quote
Old 25th March 2013, 12:03 PM   #5808
diyAudio Member
 
Kazuo Ozawa's Avatar
 
Join Date: May 2006
Location: Japan
Quote:
Originally Posted by pietjers View Post
Intersting stuff Kazuo! Thank you for sharing!

I replaced the potentiometer / led (or diode) combination in the reference leg by a voltage reference diode from the LM4040 range. I did this in order to make the reference leg as low-ohmic as possible. In that way the Reflektor will give the largest possible reaction in the regulation leg, and doesnít only rely on capacitor impedance. The disadvantage is that you can not trim the output voltage exactly (unless you add a small pot or resistor anyway), but have to search for the best combination of available reference diodes, taking into account that the two transistors add 1.2V to the output voltage. In this case for 3.3V output voltage you would need a 2.1V reference diode, or a single green LED could do the job to start with. I bypassed this with 2x 1000uF/25V Nichicon Muse KZ capacitors, which are the best I have tried so far. I have bypassed these Nichicons with foil capacitors with small resistors in series, like a zobel circuit (resistors are a few Ohms or less). Experiment with different resistor values. This gives the opportunity to voice it and get the best sonic results. I use multiple bypasses ranging from bigger to small. The results are amazing.

For the zobel across the shunt, I measure with an oscilloscope the output for stability. As long as that there are no oscillations I lower the capacitor value to as small as possible. My application remained stable with no zobel at all. So I donít have a zobel here. But I donít use the original CCS, because of stability issues I ran into. Others seem to have no problems, but I switched to a different CCS design.

Peter
Thank for your sincerely advice.
I am very glad you have same idea of mine.
Actually I have already realized this idea on 5.1V Reflektor using with two Red LEDs, and 3.9V Reflektor using with one Red Led and one diode.

I am attracted your idea of vref zobel, you bother me
Thanks
  Reply With Quote
Old 25th March 2013, 06:00 PM   #5809
Salas is offline Salas  Greece
diyAudio Chief Moderator
 
Salas's Avatar
 
Join Date: Oct 2002
Location: Athens-Greece
Some notes.

1) M2. Higher gfs Mosfets in higher output voltage Reflektors will have less disadvantage for Cxxx than under 5V operation. The various capacitance phenomena lessen with more voltage. So for users of higher Vout they should compare again the two types mentioned. Higher gfs will lower Zo.

2) R5. Using lower threshold Mosfets lowers the mirror's current with same 1K as it was for normal threshold Mosfets. So 500R looks in the right direction in your build. But other users with higher voltage outputs and normal VgsTh Mosfets should try again and compare.

*About snubber across Vref capacitor it was there in the original posting describing the Reflektor idea. R.C. had played with that back then also. Then dropped as too tweaky, but happy to know such an approach provided for the patient builder in the end.

**About active fixed ref good parts let me remind IR LEDS and LM329 6.9V precision reference low noise also for higher outputs.

Many adequate parts we can't afford not to check in different situations must be in our parts bins.

Click the image to open in full size.
  Reply With Quote
Old 26th March 2013, 05:15 AM   #5810
diyAudio Member
 
Kazuo Ozawa's Avatar
 
Join Date: May 2006
Location: Japan
Quote:
Originally Posted by RCruz View Post
Thank you for your contribution Kazuo
If I did not your previous reports posted on #4451 or 4453, I would not have come up with the idea to try various value of C1.
I thank your effort.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



New To Site? Need Help?

All times are GMT. The time now is 07:15 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