FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 

$8 USB to +/-12V "low-noise" PSU
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author $8 USB to +/-12V "low-noise" PSU
I found a cheap USB to +/-12V power supply that seems to be relatively low-noise based on my little scope (<20mVpp ripple?). It's rated at 1.5A for both rails together but drawing that much could be bad news for ripple. My guess is that it could probably handle an asymmetric load of say +12V 800mA and -12V 400mA.

Anybody else want to experiment with this? tive-Regulated-Power-Supply-Module/332677482180?ssPageName=STRK%3AMEBI DX%3AIT&_trksid=p2057872.m2749.l2649

The switching regulator it uses (XL6008E1) seems to have a good reputation - it's also in this decent headphone amp:

I'm also feeding the USB 5V directly to my module:

PLEASE NOTE: If anybody wants to try this, I highly recommend a top-quality USB psu. I measured the ripple of various ones I had lying around and Samsung USB psu's seem to be the best so far.
oh my...

"This is the team finished, the 3 month warranty, PCB double gold, special shielding low loss high-frequency ceramic capacitor inductor, a plurality of parallel power supply resistance is very low. The sound quality of the HIFI equipment is very delicate, clear, and the background is very quiet. Of course, here thanks to the technical support of rice users." oltage-to-positive-and-negative-power-output-DC12V-Regulated-power/328 08889657.html

just ordered one off ebay because why not. got some holidays coming up in july, might try to make a little travel box.

rice: used
sneak-thief wrote:
I found a cheap USB to +/-12V power supply that seems to be relatively low-noise based on my little scope (<20mVpp ripple?). It's rated at 1.5A for both rails together but drawing that much could be bad news for ripple. My guess is that it could probably handle an asymmetric load of say +12V 800mA and -12V 400mA.
Be carful with the current rating, the ideal 1.5A is not achievable under all (or possibly any) conditions. When powered with a 5V USB input you could expect something under 1000mA from the positive rail so 800mA might be close to the limit. They would need to show the specific circuit used to know what the negative rail will provide.
@Mungo: I agree! It's definitely something to classify as "cheap & experimental".

The question is: what does 1.5A even mean here? I would be a little surprised if they specifically mean a symmetric 750mA per rail - my guess is that it's a thermal limit.

In an ideal configuration, an XL6008E1 can deliver 3A to the positive rail and the 4A on the negative from the XL4003, but clearly such a small board doesn't appear to be setup for the kind of heat dissipation required, even though that chips is supposedly up to 92% efficient.

In any case, it's decent so far for lighter loads.

The trimpots have a wide range and so once you set them to +/-12V under full load, I recommend putting a dab of nail polish on the set screws.
Nice find! Usually those buck/boost converter circuits have a fairly large electro on the input and output to filter out ripple? Whats the values of those SMD ceramics on the outputs? Maybe swapping them out for something more reliable might be the go?
@col - the input is a 220uF, the ceramic outputs are chunky, perhaps 4.7uF or 10uF.

I don't have a full-blown scope at the moment so I can't give precise data on the ripple frequency - which could be helpful in choosing extra caps to target specific ripple.

That said, caps aren't a magic bullet against all switching psu ripple.

From what I've skimmed in research papers, the inductors also play a large role in "quenching" ripple.

I looked at a bunch of these cheapo bipolar psus and what sets this one apart is the combination of USB input and the claims of "low-noise" - the others I've seen usually require at least 5.5V, sometimes 14V: -Module-Low-Ripple-Double-Filter/182294373434 -Module-Low-Ripple-Double-Filter/182294373434

Obviously this isn't a Hinton or paults psu, so don't expect anything fabulous here. Or even middling.
Iv'e ordered one to have a play with. thumbs up

Edit: The typical Application Circuit from the XL6008 datasheet shows a 220uF cap on the output:

Clearly something a little bit different is going on as this eBay board uses 470uH and 4.7uH inductors instead of a single 33uH. And no diode on the + rail.

As I mentioned, this need to be scoped under load to see what kind of noise it spits out.
used one of these to make 12v from usb to use some 12v gear while picknicking, works fine! thanks for the Idea
been using one as a test supply for new builds and it's very handy. got another on the way which will go into a portable case at some point.
Still waiting for mine to arrive.
I did a quick load test with one of these. The following is symmetric load, only up to 100mA per rail since I had neither an electronic load nor high power resistors at hand.

As you can see the regulation is not spectacular, but probably ok for light loads. Ripple and noise is around 100mVpp typical (20MHz bandwidth), but fluctuates and sometimes a few times that. The negative rail seems better behaved than the positive rail. This was running from a Nokia usb charger, I later tried with an Ikea charger and the positive rail got a bit better, around 40-50mVpp.

Some observations:
- The postive rail needs a minimum load of several mA, otherwise the switcher doesn't start properly but is oscillating on/off at some 70 Hz or so. Due to lots of output capacitance on this board the ripple is still only 100mV, so it will work fine for many applications, but this is now at 70Hz so could be very much audible in a synth
- At 100mA per rail the switching ICs get hot to the touch. I wouldn't go much higher than a few hundred mA constant load, and in any case keep an eye on temperature in actual use.
- The rails are mostly independent so asymmetric load shouldn't be a problem.
- The unit starts up just fine with 440uF capacitive load per rail (I haven't tested larger capacitances)

All in all seems fine for small loads, and the noise levels are not particularly low but comparable to for example Meanwell or Traco Power DC-DC converters. Once you draw the minimum load the noise is also all very much above audio frequency so many modules will not care about it.
@kassu - Thanks for the analysis! As I mentioned, having a good quality USB charger makes a huge difference. Is this the one you used?

I bet you could get less ripple with an even better quality USB charger.

Also, I highly recommend using a shorter, good quality USB cable with a decent AWG otherwise the 5V can drop a lot, even as low as 4.7V.

FYI I'm using about 220mA @ +12v and 25mA @ -12V. So far it's not getting hotter than a 7812/7912 setup with a 15-17V input.
MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 1
Powered by phpBB © phpBB Group