Ultra compact, quiet, low power, +/-15v PS, needed

Jump to Latest
noviygera

noviygera

Member
Joined 2009
#1
Gents,
As the title states, I am looking for something to replace the now, seemingly unobtanium, SilentSwitcher. I liked those a lot!

I need to power 2 to 4 opamps, so very little power needed, but needs to be very COMPACT and high performance. Are there pcbs or kits available?

thanks!
 
#6
The smallest one I can think of is a capacitor multiplier made with a mix of through hole and surface mount components
It’s slightly too big for you maybe.
 

Attachments

  • E10BE37D-10CB-430B-B4F1-1389660530E3.png
    E10BE37D-10CB-430B-B4F1-1389660530E3.png
    45.5 KB · Views: 231
#11
Happy to share my project!

It is an Lm1875 Gain clone amplifier with an opa1642 op amp in first stage acting as a differential amplifier for balanced inputs. As is, the OPA1642 gets +/-15 V from Zener diodes.

I hooked up a temporary test set up with a dedicated “silent switcher” feeding the opamp. It sounds better this way.

IMG_1673.jpeg


Since the “silence switcher” is no longer available, I would like to get a dedicated power supply for the opa1642. Or clean up the existing +/-15 V supply. I have a bunch of these Gain clone amps, so I need a repeatable solution.
 
#12
Ok, that's better.

Now I can tell you with confidence that the LT3045/LT3094 type regulators would be perfect for that application. The ldovr website sells the little PCB reg miniature boards as well:
https://www.ldovr.com/category-s/124.htm
https://www.ldovr.com/Ultra-Low-Noise-Low-Dropout-Voltage-Regulator-p/lt3045-78xx.htm

If you care to read the 3045 spec sheets... you'll find an extremely useful section (page 19) that talks about how to filter out the very high (MHz range) spikes. You'll probably think that this is simply not applicable;e for low-frequency audio... but it is actually very important... You'll have an ultimate regulator at hand... => you could use little inductors in front of 3045/3094 regs as well... they do not have to be large, but they should attenuate the MHz range rubbish.

Edit: have a look here... I used 4 x 3045 regs with an exceptional result... (this is just one example of LT3045 regs' implementation/mod I did..., i.e. I use them a lot with great results)
https://www.diyaudio.com/community/threads/roland-ua-s10.399385/#post-7352724
 
Last edited:
#15
The LDOVR one has rectifier diodes and filter caps so it is an all-in-one solution. It just needs a 2 x 15V transformer like the good Talema toroid types.

If bricks are SMPS then you would need 2. Inefficient and an ugly solution with regards to mains cabling. What will happen if either one fails?! Besides that many SMPS are RF transmitters and not at all suitable for audio.

! Wait a minute.... if that LM1875 device is the device where the PSU is meant for then you would need a board like you have bought. Then you should be able to tell what raw DC voltages the Zener diodes are fed. Before the series resistors that is.
 
Last edited:
#16
This power supply is needed only for powering the single opamp. Can I use two tiny SMPS bricks and float the Ground or will that create a ground reference problem? If I use two positive V bricks to get and positive/negative supply?

1685984581056.png
 
#17
Why? You have symmetric voltages already there haven’t you? The point where the Zener diodes are connected to via series resistors. It is likely possible to change the Zener diodes to 18V ones and adapt the series resistors to that. Then you could use the LDO modules inputs to the +/-18V….

Of course it would help if we would be informed what the brand and type number of that amplifier is.
 
Last edited:
#18
Jean-Paul,

the amp is Furman Model SP-20A.

I am looking for some explanations about the effects of power supply on the sound.

So far, I have experimented with only changing the PS to the opamp. Opamp only!

TL072+LM1875.jpg


Factory standard: +-15v derived from +-24v with zeners and 1K resistor.

1st change: change TL072 to OPA1642: noticeable improvement in SQ. 3d space opened, sound more detailed but also smoother. like mp3 vs WAV sound.

2nd change: OPA1642 running off the +-15v rails, followed by LT3045/LT3094 type regulated down to 14.5v: Sound a little cleaner but less dynamic and less fun. Better but keep going.

3rd change: OPA1642 powered by SilentSwitcher. This is getting very good; much better than the 2 previous. Unfortunately, SilentSwitcher is unobtanium now. So no go.

The way forward is to understand why powering OPA1642 from amp rails (as in schematic) is inferior in sound to a dedicated power via SilentSwitcher. I feel this is very important for me to understand, to understand audio.

Please explain why this is happening, and what parameters of the PS are responsible for the sonic change. Voltage is same +-15v!
 
#19
Hi I know the phenomenon and used regulated power supplies with equal results. Almost always better than simple Zener diodes and since I found that out at young age I never bothered to find out why exactly. You have +/-24V so you can use regulate power supplies by means of regulators that can have 24V at the inputs. Or you use 18V Zener diodes and feed modern LDOs but I think that is a tad too complex. The Silentswitcher has good linear regulators so it seems good clean rails are important to the end results. No news there 🙂 Probably how the opamps power pins are decoupled is also relevant (also usual) and results may vary and depend on which type and what value caps you use. Very good results here with OPA1642 when correctly decoupled.


Tip: I would replace C102, C202, C201 for 5 mm pitch Wima MKS 10 µF without blinking an eye. D1...D8 could be nice rightly chosen Schottky types.
 
Last edited:
  • Like
Reactions: martin clark and Brijac
You must log in or register to reply here.
Top Bottom