Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Salas SSLV1.3 UltraBiB shunt regulator
Salas SSLV1.3 UltraBiB shunt regulator
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 16th May 2018, 12:19 PM   #91
gwmorgan is offline gwmorgan  US Virgin Islands
diyAudio Member
 
Join Date: Nov 2005
Location: UK
Ok thks
__________________
Geoff.
  Reply With Quote
Old 16th May 2018, 03:18 PM   #92
tmas is offline tmas  United States
diyAudio Member
 
Join Date: Nov 2008
Are the resistors with the exception of Rs, Rf and R1 1/2 watt dissipation?
  Reply With Quote
Old 16th May 2018, 04:05 PM   #93
Salas is online now Salas  Greece
diyAudio Chief Moderator
 
Salas's Avatar
 
Join Date: Oct 2002
Location: Athens-Greece
Salas SSLV1.3 UltraBiB shunt regulator
Yes or 1/4W even. There are 10mm distanced pads for them. Prefer 50ppm. No more than 100ppm anyway. I used Xicon 1/4W 50ppm and Beyschlag 0.4W 50ppm mainly. But whatever good that's stocked and it fits will do. For R9 buy 1kΩ alternative also. Its about when forced to use much spare current because some application changes states of consumption for instance. Not to oscillate in the transition because with say 0.5A spare burned in M2 its transconductance and temperature will go up and its characteristics will change. Most beta testers used 1K that I had designated for safety but they ended up using little spare current which is much better for M2's current noise and temperature (plus smaller sinks). Up to 200-250mA spare current R9 270R is fine and the impedance stays lower up at 1MHZ vs 1k. I don't know yet if there is a subjective benefit in that or to put 1k R9 and forget it. When some beta tester listens between those two gate stopper values where 270R is enough for stability and has a finding I will decide. Pay attention, spare current is what remains to be burned in M2 after load's draw. Not the M1 CCS limiter's R1 mA setting. Spare=CCS mA minus Load's draw mA.

Also for those wanting the reg for no higher than 20V output applications, C2 can be 25V type. And VR1 be 10k for more turns focused in the range of interest. Especially for digital because clocks are supposed to appreciate low 1/f noise C2=470uF/25V Nichicon KZ or 470uF/35V Nichicon ES BP are nice. C3 remains 33uF/50V ES BP or 33uF/50/100V KZ or 100uF/50V Pana FC.

What happens with higher value C2 is the 1/f (subsonic) noise lessens vs the 220uF median (470uF is magenta line when 220uF is red line in post#1 2nd chart) but the voltage rise to nominal is slower at start up until it fully charges with the little mA that are available for that. I don't mean minutes, I mean few seconds "soft start". Some may find lower value C2 caps liked better as "bit faster, more open" in some app or not, that's the user's business to evaluate. Its a low noise reg anyway with no filtering difference from 20Hz up with as low as 100uF C2 even. As that chart also shows. In general with any value C2 its a soft start reg. It does not "spike" to over-voltage at turn on. Its temperature induced output voltage drift to full warm up (1 to 1.5%) is going down also, not drifting up. In other words your digital chips, even if any is directly fed without a local reg in its pcb design, they are safe sir thank you
  Reply With Quote
Old 16th May 2018, 06:30 PM   #94
tmas is offline tmas  United States
diyAudio Member
 
Join Date: Nov 2008
Would the 100ppm PRP resistors be an acceptable solution?
  Reply With Quote
Old 16th May 2018, 08:18 PM   #95
Salas is online now Salas  Greece
diyAudio Chief Moderator
 
Salas's Avatar
 
Join Date: Oct 2002
Location: Athens-Greece
Salas SSLV1.3 UltraBiB shunt regulator
They are alright. But if you can find 50ppm alternatives in comparable pricing go for them.
  Reply With Quote
Old Yesterday, 03:42 PM   #96
chat72 is online now chat72
diyAudio Member
 
Join Date: Dec 2012
This probably be a stupid question.

From your schematic for Neg rail, emiter of q1 is connected to negative side of c1 but in the picture of PCB I can see a trace running from -c1 to collector of q1.

Also an emitter of q1 is connected to r3 and r4. Should it be an opposite or am I missing something?
  Reply With Quote
Old Yesterday, 06:03 PM   #97
arpagon is offline arpagon  Bulgaria
diyAudio Member
 
Join Date: Oct 2011
Schematic is intrinsically unstable in its initial version, see the OLG in the first attached file - the ULG is 13.5 MHz and the phase shift is -183 deg. No more words are needed.
Introducing the RC frequency compensation network proposed by @Selfy (R=100 Ohm, C=1nF) the OLG becomes as shown in the second picture - now the ULG is 2.3 MHz and the phase shift is 96.7 dB which is much, much better - the phase margin is more than 80 deg.
The only thing which scratch my eye is the big phase excursion in the region 1 kHz - 100 kHz.
The both simulations are performed with output capacitor 100 uF with ESR=0.1 Ohm as suggested in the initial document. BTW, finding such capacitors and relying on their constant ESR over time is not an easy task.
And combining with the fact that the capacitance added on the regulator's output via attached load is almost impossible to predict a-priory, the possibility of oscillation becomes bigger and bigger.

However, the other advantage of the proposed compensation network is that it permits on the regulator's output to be used bigger capacitors with more "normal" ESR of, let say 50-60 milliOhm.
On the third picture you can see the OLG with attached capacitor 1000 uF, ESR=0.06 Ohm on the output. The nasty belly is compensated by 20 dB which is not bad, isn't it?
And on the fourth picture is plotted the schematic's output impedance - 38 uOhm on 1kHz. This proves that the insertion of the compensation network does not deteriorate the regulator's output impedance.

In conclusion - very good results for such a simple schematic.
Two shortcomings though - the poor phase linearity and the poor thermal stability of the output voltage.
Attached Images
File Type: png SSL_V1.3-OLG-orig.png (57.0 KB, 87 views)
File Type: png SSL_V1.3-OLG-compensated.png (58.1 KB, 89 views)
File Type: png SSL_V1.3-OLG-compensated-high_out_cap.png (61.2 KB, 79 views)
File Type: png SSL_V1.3-Zout.png (58.9 KB, 74 views)
  Reply With Quote
Old Yesterday, 06:59 PM   #98
Salas is online now Salas  Greece
diyAudio Chief Moderator
 
Salas's Avatar
 
Join Date: Oct 2002
Location: Athens-Greece
Salas SSLV1.3 UltraBiB shunt regulator
Quote:
Originally Posted by chat72 View Post
This probably be a stupid question.

From your schematic for Neg rail, emiter of q1 is connected to negative side of c1 but in the picture of PCB I can see a trace running from -c1 to collector of q1.

Also an emitter of q1 is connected to r3 and r4. Should it be an opposite or am I missing something?
No you are not missing something. Crt missed something and I did not catch it. Q1's orientation print should have been reversed. Tricky, it works all the same in test for steady circa -0.6V across R1 following well with the positive section's R1 voltage. That is why the glitch passed unnoticed. The negative section's Q1 can be positioned the other way around than its now printed of course because the base pin is the middle one in those types. To be typical and have closer base current to Q1 positive.

I will put this corrected orientation down in the build guide and will see to reverse the solder mask print in the future pcb runs.

Thanks for noticing.
Attached Images
File Type: jpg OrientQ1neg.jpg (452.4 KB, 72 views)
  Reply With Quote
Old Yesterday, 07:09 PM   #99
Salas is online now Salas  Greece
diyAudio Chief Moderator
 
Salas's Avatar
 
Join Date: Oct 2002
Location: Athens-Greece
Salas SSLV1.3 UltraBiB shunt regulator
Quote:
Originally Posted by arpagon View Post
Schematic is intrinsically unstable in its initial version, see the OLG in the first attached file - the ULG is 13.5 MHz and the phase shift is -183 deg. No more words are needed.
Introducing the RC frequency compensation network proposed by @Selfy (R=100 Ohm, C=1nF) the OLG becomes as shown in the second picture - now the ULG is 2.3 MHz and the phase shift is 96.7 dB which is much, much better - the phase margin is more than 80 deg.
The only thing which scratch my eye is the big phase excursion in the region 1 kHz - 100 kHz.
The both simulations are performed with output capacitor 100 uF with ESR=0.1 Ohm as suggested in the initial document. BTW, finding such capacitors and relying on their constant ESR over time is not an easy task.
And combining with the fact that the capacitance added on the regulator's output via attached load is almost impossible to predict a-priory, the possibility of oscillation becomes bigger and bigger.

However, the other advantage of the proposed compensation network is that it permits on the regulator's output to be used bigger capacitors with more "normal" ESR of, let say 50-60 milliOhm.
On the third picture you can see the OLG with attached capacitor 1000 uF, ESR=0.06 Ohm on the output. The nasty belly is compensated by 20 dB which is not bad, isn't it?
And on the fourth picture is plotted the schematic's output impedance - 38 uOhm on 1kHz. This proves that the insertion of the compensation network does not deteriorate the regulator's output impedance.

In conclusion - very good results for such a simple schematic.
Two shortcomings though - the poor phase linearity and the poor thermal stability of the output voltage.
Are you talking about Selfy's initial and then the compensated?

Because the UltraBiB simulates like this for OLG:
Attached Images
File Type: png UBiB_OL.png (17.6 KB, 64 views)
  Reply With Quote
Old Yesterday, 07:42 PM   #100
selfy is offline selfy  Bulgaria
diyAudio Member
 
selfy's Avatar
 
Join Date: May 2009
Location: Thracian Valley
Salas, where/how do you “break” the closed loop to simulate the open loop?
__________________
"Music is enough for a lifetime, but a lifetime is not enough for music" — Rachmaninoff
  Reply With Quote

Reply


Salas SSLV1.3 UltraBiB shunt regulatorHide 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

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
GB Thread for Salas SSLV1.1 BiB shunt reg Tea-Bag Group Buys 1662 13th May 2018 06:52 PM
FS: Salas SSLV1.1 BiB shunt regulators Rudi_Ratlos Swap Meet 2 2nd April 2018 03:20 PM
WTB: Salas SSLV1.1 BiB shunt reg KevinHeem Swap Meet 1 12th February 2015 05:04 PM
WTB: Salas SSLV1.1 BiB shunt reg positive board sbelyo Swap Meet 1 29th December 2014 03:17 AM
WTB : Salas SSLV1.1 BiB shunt CaNNA3IS Swap Meet 1 24th November 2013 09:43 AM


New To Site? Need Help?

All times are GMT. The time now is 02:36 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki