audio power supply optimizaton

[sandyk]

On ‎6‎/‎28‎/‎2018 at 2:38 PM, look&listen said:

1st stage regulator is Audiowind LT1963 PCB, but replace old fast recovery rectifier diodes for Schottky's, wire power status LED at front panel.

 

 Other than replacing the HER303 bridge rectifier diodes with Schottky diodes such as the 1N5822, I would also recommend replacing the 330 ohm resistor with something like a 120 ohm metal film resistor, and the trimpot with a 2Kohm multi turn  trimpot, as the 10K trimpot is too large a value to accurately set voltages normally used by C.A. members. (e,g, 5V to 12V)

 When used as described here, you can also use the DC Input terminals to power a front panel LED with a series resistor whose value is chosen to give the desired brightness level.

[look&listen]

14 hours ago, sandyk said:

I would also recommend replacing

Low return on effort for me. Voltage adjustment has +/- overhang but can get close enough for no trouble with next LDO stage. Absolute voltages not very important with many downstream regulation stages (in PSU and powered device).

LED was 1st parallel to original, but to dim, so clipped out original LED and output OK. Trouble was soldering only 5-6mm leads from tiny device (for small panel hole). Tweezers to sink leads, clamp for wire extensions and light touch. Happy can still do small work with many decades ?

 

Audiowind PCB just cheap ($19) Chinese product, but good for not-demanding role.

 

PS. Schottky diodes = VSB3200S-M73 from Mouser for same basic specs as HER303s at $0.67

[sandyk]

44 minutes ago, look&listen said:

Low return on effort for me. Voltage adjustment has +/- overhang but can get close enough for no trouble with next LDO stage. Absolute voltages not very important with many downstream regulation stages (in PSU and powered device).

 In my case, I am using this with a JLH PSU -add on,(no voltage drop)  so I needed to get the output close to either +5V or +9V.

 I found the adjustment too touchy with the original 10K trim pot, and even lightly touching the trimpot case caused a noticeable voltage variation. You can also use this PCB with an external D.C. voltage and gain further filtering if sufficient voltage overhead , as shown in the attached schematic. Just add the 100uF under the PCB.

 Incidentally, if you added the attached, you would have the makings of a very nice, low noise, adjustable Power Supply.

 

https://www.ebay.com.au/itm/3590S-Precision-Wirewound-Potentiometer-500R1K2K5K10K20K50K100K-Ohm-Pot-10Turn/191964511143?hash=item2cb1f9fba7%3Am%3AmOBP9Laxbe3F69yQwKRRlQA&var=491328143361

 

https://www.ebay.com.au/itm/Red-Blue-Yellow-5-Digit-LED-DC-0-4-3000-33-000V-Digital-Voltage-Panel-Voltmeter-/263002919525?var=&hash=item3d3c321265

 

 

 

[look&listen]

15 minutes ago, sandyk said:

so I needed to get the output close to either +5V or +9V.

Once V is set should be Ok. Not notice any drift with this module after long service, and some resetting V for other purposes.

[look&listen]

On 6/28/2018 at 12:38 AM, look&listen said:

Now needs time for burn-in & listening to many things in many ways. Later maybe try to put hearing experience to feeble words.

Ok, time for short listening report.

Short version; listen for 13 days, not happy, remove Kemet caps post-LT3045, listen again and big SQ+ very happy now.

 

More detail: Had startup trouble costing 2 days of looking for intermittent problem & everything check out OK. ?

 

1) after reinstall some initial casual listening. then checkin with sound various days. Hear good (soundstage, bass, strings & timbral nuance) & bad (fuzzy midrange & lyrics, overall confused), hope bad go away.

 

2) after week+ bad SQ effects still same. Read Alex @Superdad comments on relative merits of LT3045 & Kemet caps & after 13 days frustrated by sound, give up on burn-in trial.

 

3) Take down LPSU stack to remove Kemets from 3 reg. stage 7V PSU & move to Ext-HDD 12V PSU (Sigma 11 variant)*. Similar 12V PSU for server also has same 1120 mfd of Kemet Alu. polymer cap on output bus.

 

4) Good SQ effects remain if not better. Midrange fuzzy gone, & lyrical clarity some better then before changes, where soundstage big solid, stable increase(!), overall sound, clean highs, nuance of small sounds all significant increases. Hear more variation in recorded bass & soundstage quality. More surprising realism in some tracks & instruments. Just tonight listen to some dsf tracks again, and sound better then described so far. Maybe see/hear a few days from now ?!!

 

Must say good SQ results not just because of large upgrade of 7.1V USB LPSU & cable, but little upgrades of other 2 LPSUs, cables & AC wiring simplification (still more to do). Some variables were heard in single or small group clusters, others mixed in big batches. Not time to hear every different thing, but think have isolated important things so far.

 

Things learned:
Thanks for inexpensive low noise parts like R-cores, Schottky diodes, LT3045, low ESR caps, good cable shielding!

Multiple LT3045 ultra low noise regulator modules in singles or series give fine audiophile performance to good gear.
Do not use caps, even fine Kemet Aluminum Polymers, after (or before ?) high bandwidth LT3045 regulators or reduce SQ!
Suspect complete shielding of modules, components & connecting wires important for better sound.
Seems working to millionths and billionths in power supply specs can bring rich rewards!

 

* More problems after moving caps, frustrating hours later track down & fix intermittent short in new DC cable

 

Can write more on SQ effects if desired.
Also write more on all 3 LPSU stack if curious.

 

[lmitche]

This whole DC cable and cap experimentation effort has reaped great SQ here.

 

Per the suggestion from Alex I've taken the kemets off the lt3045 outputs and off the lps1.2s feeding lt3045 inputs. They remain on the lps-1 and sigma 11 feeds into the remaining lt3045 chains. I have done this on two systems with great SQ results.

 

I have four more Gotham cables to build, the most interesting will be the 12 volt 8 pin eps/cpu feed inside the PC. My new crimping tool arrived today, so I am ready to go. This is the highest current cable in the rig, so the results will be interesting.

 

 

[look&listen]

15 minutes ago, lmitche said:

Per the suggestion from Alex I've taken the kemets off the lt3045 outputs and off the lps1.2s feeding lt3045 inputs. They remain on the lps-1 and sigma 11 feeds into the remaining lt3045 chains. I have done this on two systems with great SQ results.

Now count 4 independent listening reports confirming theory of no caps on LT3045 output.

Settled issue for me.

Maybe time for @Cornan to test?

[Cornan]

8 hours ago, look&listen said:

Now count 4 independent listening reports confirming theory of no caps on LT3045 output.

Settled issue for me.

Maybe time for @Cornan to test?

 

I have started the re-evaluations with the Kemets on the output of LT3045, but very slowly since I’m currently on summer holliday with my kids and are not at home that much. You’ve got to remember that I’ve 17 of them spread around my system right now and 11 LT3045s (12 incl LPS-1.2) and it will take some time. So far it is not that consistant as just removing the LT3045 at the output of every LT3045 in my setup. In some spots they still do improve SQ. In some spots it is better without. Still trying to work out were the best spots are, but it is a very time consuming effort in my setup with so many of them.

 

Also remember that I have’nt counted the built-in aluminium polymer caps used at the OUTPUTS of my LS-HPULNs, MS-HPULNs and SD-HPULN. 

 

I also recently got a pair of 18uF Jenzen Audio cross caps that I want to experiment with in series at the safety ground input - ac output ground (cudos to @Abtr) on devices like DC blocker trap filter, balanced isolation transformer and/or balanced ac ps. Let’s see what that comes to in the end? ?

[look&listen]

On 7/14/2018 at 12:12 AM, look&listen said:

* More problems after moving caps, frustrating hours later track down & fix intermittent short in new DC cable

Forgot important point (& happiness);  current limiting in LT3045(s) worked to prevent any component damage! Only warm cable end, suggesting problem.

[look&listen]

6 hours ago, Cornan said:

Also remember that I have’nt counted the built-in aluminium polymer caps used at the OUTPUTS of my LS-HPULNs, MS-HPULNs and SD-HPULN. 

Yes. Understand you got idea from those modules. Also assume from pcb layout that caps connected to output bus - but no confirmation.

Maybe should bring Stammheim into conversation?

[Cornan]

4 hours ago, look&listen said:

Yes. Understand you got idea from those modules. Also assume from pcb layout that caps connected to output bus - but no confirmation.

Maybe should bring Stammheim into conversation?

 

Quite frankly I do not want to through anyone into a infected discussion like this when RB2013/Tubelover2 is around making personal jokes about anything that involves Kemet caps and frankly everything that was’nt his own idea in the first place. Just face it. Kemets do improve SQ in the right spots, but this discussion have scared a lot of people off trying them at all. Anywhere. Input ot output of any linear regulator. 

I am still not convinced that it is the correct assumption that Kemet does’nt improve at all atthe output if the LT3045. I am not just stubborn about it because I have digged so deep, but because I can clearly hear progress when they are used in the whole DC chain. Despite what you, Alex C, Tubelover2, JS or anyone else have to say about it I am determind to find out about it in my own setup and make the judgement myself. I am only stubborn when I truly hear the possibillities.

[look&listen]

22 minutes ago, Cornan said:

Kemets do improve SQ in the right spots

Yes, point of discussion!

 

23 minutes ago, Cornan said:

but this discussion have scared a lot of people off trying them at all

No. Just to be more precise & effective in application, & peer 'ear' review builds evidence.

 

24 minutes ago, Cornan said:

a infected discussion like this

Perhaps remove testosterone from equation & discussion get more clear for you?

 

[Cornan]

8 minutes ago, look&listen said:

No. Just to be more precise & effective in application, & peer 'ear' review builds evidence.

 

Indeed, but making judgements based on a few isolated LPSU designs does’nt give you the whole picture either IMO. No point being precise or effective until you have seen the whole canvas. That’s the whole purpose of my experiments with the Kemets.

 

14 minutes ago, look&listen said:

Perhaps remove testosterone from equation & discussion get more clear for you?

 

That’s almost a Tubelover2 reply right there. Congrats!

[RickyV]

For anyone’s info who is using these fake sigma 11 power supplies, i use them too, original are from https://www.amb.org/audio/. 

 

The big capacitor on the right is on the output of the power supply and not in the original amb schematic!! Mine came with Hover caps, i removed both and replace the left one with mundorf ag cap. Try it, it is a nice improvement.

[lmitche]

3 hours ago, RickyV said:

For anyone’s info who is using these fake sigma 11 power supplies, i use them too, original are from https://www.amb.org/audio/. 

 

The big capacitor on the right is on the output of the power supply and not in the original amb schematic!! Mine came with Hover caps, i removed both and replace the left one with mundorf ag cap. Try it, it is a nice improvement.

Yes indeed, I use both the original AMB (with Mundorf caps) and the Chinese versions with Nover caps. The latter seem to have lower ripple with both large caps 24 to 10.

 

In any case, the sigma 11s are great for powering the lt3045 DC to DC boards.

[hurka]

I didn’t find sigma11 on taobao.this from other website 

[lmitche]

Here you go!

 

https://www.ebay.com/itm/Low-Noise-S12-Super-regulated-linear-power-supply-board-5-30V-LPS-10000uF-/263610893782?oid=132048787951

[hurka]

15 minutes ago, lmitche said:

Here you go!

 

https://www.ebay.com/itm/Low-Noise-S12-Super-regulated-linear-power-supply-board-5-30V-LPS-10000uF-/263610893782?oid=132048787951

thx!!! 233rmb vs 135

[look&listen]

39 minutes ago, lmitche said:

Here you go!

Another varient.

 

https://www.ebay.com/itm/ZEROZONE-Assembled-Ultra-Low-Noise-linear-Power-supply-board-LPS-board-L158-25/131588996943?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m1438.l2649

 

[hurka]

1 hour ago, look&listen said:

Another varient.

 

https://www.ebay.com/itm/ZEROZONE-Assembled-Ultra-Low-Noise-linear-Power-supply-board-LPS-board-L158-25/131588996943?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m1438.l2649

 

I must leave China,then easy to buy.....

zerozone on Taobao , only have sigma 22 

[tapatrick]

On 5/1/2018 at 11:37 PM, Superdad said:

 

Perfect for we knights on the quest for the holy [audio] grail!  

(Now I can't help but whistle the theme song from the Monty Python movie...)

...this would be good protection against chain-emails too...

[Cornan]

9 hours ago, tapatrick said:

...this would be good protection against chain-emails too...

 

Surely against Kemet too! ?

[look&listen]

20 hours ago, RickyV said:

i removed both and replace the left one with mundorf ag cap. Try it, it is a nice improvement.

Can provide more details please?

Inquiring minds must know!

[lmitche]

2 hours ago, Cornan said:

 

Surely against Kemet too! ?

Hey Cornan,

 

I like the Kemets in place here, just not at the lt3045 board outputs. I am grateful to you for the recommendation, especially when used on the input side.

 

Very helpful,

 

Larry

[RickyV]

2 hours ago, look&listen said:

Can provide more details please?

Inquiring minds must know!

 

Please look&read at the “ISO Regen performance Improvement Cheap!” Threat, my fingers hurt.