Hey, I have one of these LPS on order from AliExpress:
https://www.aliexpress.com/item/1005004989616614.html
Some of the description:
The materials used are very sophisticated, high-speed and low-dropout 4A Schottky rectifier diode, the output tube is ON tube king 15034, TL072 dual op amp, NOVER 6800uF/35V main filter capacitor. Integrated PCB design with output overcurrent and overload function to improve power supply safety.
I was thinking of upgrading / replacing a few of the components to better quality parts. It's to power a pretty critical device for me, am audio USB PC Card as external power. Thinking caps - but looking at the stuff, the opamp TL072 maybe, not sure, interested in any suggestions.
https://www.aliexpress.com/item/1005004989616614.html
Some of the description:
The materials used are very sophisticated, high-speed and low-dropout 4A Schottky rectifier diode, the output tube is ON tube king 15034, TL072 dual op amp, NOVER 6800uF/35V main filter capacitor. Integrated PCB design with output overcurrent and overload function to improve power supply safety.
I was thinking of upgrading / replacing a few of the components to better quality parts. It's to power a pretty critical device for me, am audio USB PC Card as external power. Thinking caps - but looking at the stuff, the opamp TL072 maybe, not sure, interested in any suggestions.
The seller doesn't supply the schematic on the listing but says its based on Studer 900 PSU so I found a schematic for that. The TL072 opamp is the same, two 1N4148s looks the same but I can only see one zener in the pics.
To make upgrading worthwhile its important to understand where the low-hanging fruit is as regards performance bottlenecks. That's difficult without an understanding of the schematic. Upgrading the TL072 to a lower noise part might improve things but could be the opamp isn't the limit on the noise, rather it could be the zener being used as reference that's noisier than the opamp.
To make upgrading worthwhile its important to understand where the low-hanging fruit is as regards performance bottlenecks. That's difficult without an understanding of the schematic. Upgrading the TL072 to a lower noise part might improve things but could be the opamp isn't the limit on the noise, rather it could be the zener being used as reference that's noisier than the opamp.
Perhaps use your soundcard and spectrum analyser software such as REW to measure the existing noise floor for a signal loopback configuration with standard USB powering (ie. try and identify any mains related signature) - that is then your benchmark for assessment rather than your ears. Then repeat with your new LPS powering the soundcard.
If you identify some noise signature then perhaps make the effort to use a 12V battery and 7805 powering the soundcard, and then with a noisy USB battery brick powering the soundcard (it would likely add non-mains related noise), as a way of cross checking if your LPS has any influence at all over your concerns.
If you do make changes to the LPS then I'd recommend repeating the noise floor measurement after any single changes you make.
The only obvious modification I would make to the LPS would be to test and add a quasimodo snubber to the input LV ac, and identify if a smt cap can be added under the pcb in a useful location to constrain diode commutation noise.
You may end up finding that there are other noise issues of much greater concern than your LPS, and those are where effort should be directed.
If you identify some noise signature then perhaps make the effort to use a 12V battery and 7805 powering the soundcard, and then with a noisy USB battery brick powering the soundcard (it would likely add non-mains related noise), as a way of cross checking if your LPS has any influence at all over your concerns.
If you do make changes to the LPS then I'd recommend repeating the noise floor measurement after any single changes you make.
The only obvious modification I would make to the LPS would be to test and add a quasimodo snubber to the input LV ac, and identify if a smt cap can be added under the pcb in a useful location to constrain diode commutation noise.
You may end up finding that there are other noise issues of much greater concern than your LPS, and those are where effort should be directed.
Alright. I have an OPA1642 with adapter coming. I recapped with it proper parts, took out a cap which was Sam Young brand. All Vishay BC / Nichicon. Replaced the hot run from inlet to switch with solid core silver, replaced some of the pathetic 26awg wire with 18awg kimber. Hopefully the new opamp makes an impact. I have parts for a new DC cable coming, Viborg GX16 on 5.5/2.5 for my USB card, these are on Canare GS-6 cable which is a nice gauge and fabulous looking copper, at least 18 awg conductors. Would placing a small cap across the opamp pins 4&8 be something to consider?
Last edited:
Samyoung's a good brand in my experience, I use them a lot. In my understanding its a Korean company but founded by Nichicon.
From looking at the schematic the most critical cap is going to be the 100uF/25V next to the opamp as that's filtering the zener reference. Its leakage will be critical so I'd search out the lowest leakage part you can find for that.
As regarding adding caps, I echo what @trobbins said about snubbing the transformer and the 4 rectifier diodes. They look to be standard types, not soft recovery so they're really crying out for taming.
From looking at the schematic the most critical cap is going to be the 100uF/25V next to the opamp as that's filtering the zener reference. Its leakage will be critical so I'd search out the lowest leakage part you can find for that.
As regarding adding caps, I echo what @trobbins said about snubbing the transformer and the 4 rectifier diodes. They look to be standard types, not soft recovery so they're really crying out for taming.
The advertising blurb quoted in post #1 indicates schottky rectifier diodes, but the advertising photo, and photo in post #7 indicate standard 1N540x 3A diode (which has no spec for recovery characteristic). Highly likely that UF5404 would be reasonable replacement.
A photo of reverse side of pcb should indicate how well the pcb traces are laid out for noise, and where any additional smt cap(s) may have some benefit.
A photo of reverse side of pcb should indicate how well the pcb traces are laid out for noise, and where any additional smt cap(s) may have some benefit.
Actually I also removed the X2 cap, highs got crispier, but overall sound presentation had an edge to it., put the cap back and still nicely detailed but maybe a slightly better sound to the overall presentation - 100% subjective, just trying things. It might be physiological but I thought the ferrite may have removed some of the sparkle.