5v SMPS power supply outputting 5.7v

Hi guys, I am new to arduino and electronics so i apologise in advance if my problem sounds dumb to more experienced people.

I am building an animated pixel clock with an arduino mega and a ws2812b led strip turned into a homemade pixel screen. I learned a lot about programming and all went well in prototype phase.

My problem is that when building it for real i switched my homemade bench power supply for a cheap SMPS 5v 10a power supply to feed both the arduino and the led screen but the psu is outputting 5.7v with no load... that made me not trust this psu.

Can i plug the power supply to the led strip directly, will the voltage drop a bit since i am using 256 ws2812b leds? The absolute maximum rating for the ws2812b led is 5.3v so i got worried i might fry my screen instantly after so much work...

Just to be safe i did put two 1n4007 diodes in series to lower the voltage to the arduino USB port (got it to 5.03v) and the mega turned on with no apparent issues. All seems fine, but i did not plug the led screen fearing i might destroy it.

I dont know how i can drop the 5.7v for the screen to at least 5.3v, since i am using 256 individual ws2812b leds so the current requirements are quite high.

Also i am not sure if SMPS psus might have dangerous voltage spikes that might fry my mega or screen in the long run (cant afford an oscilloscope yet). I did put a 1000uf decoupling capacitor in the vcc and ground lines leading to the screen... didnt do that for the lines that feed the arduino though. Should I do that? Im pretty clueless with this one.

I am still quite new when it comes to electronics so i appreciate any help i might get and apologise in advance for my crudeness.

Not knowing what you have I can only venture to guess. I believe (just guessing) all sides but one have a metal shield, the other is probably almost solid. I believe there is a small pot by the terminal strip which will allow you to adjust the 5V output. Connect your voltmeter to the power supply and try to adjust it. I have used several of them and they all had the adjustment. I will assume it has a switch to select 115 or 220. Some of the better designs do this automatically and do not have the switch. More information always helps. Have fun with your project and let us know how it works out.

Thank you for your quick answer. Unfortunately this psu has no trimpot to adjust the output voltage (I opened it up just to be sure). It is similar to a computer power brick, made in Brazil by a generic brand named Kairos, i did Google search the model number, T0599, but nothing came up. Its input is 100v to 240v ac and the output should be 5v dc 10a, but in reality it is outputting 5.7v with no load. I bought this psu because it is quite small and my clock will be hanging on the wall like a painting, so i couldnt fit a big psu.

Connect the 5.7V to the strip in several places with several diodes so that no one part has to have all the current going thru it.
Here are a couple diodes rated for 20A, you may need to use a heatsink with also.

Thank you for your reply Crossroads. That is a good idea. Unfortunately after a couple of hours powering the arduino with the 5.7v psu (with two 1n4007 diodes in series reducing the voltage to 5.03v) the clock started to show nonsense data... that didnt happend while i was using the bench power supply.

Just to be sure i unplugged the new psu and powered the clock through the bench psu again for a few hours and it worked flawlessly. The sketch is working great, no issues there i am certain because i tested it for many hours with zero issues.

My guess is that the new psu is giving voltage spikes or serious ripples... I may have to buy another psu, fixing this might be beyound my current skills but im not sure i can find a psu that is suitable and compact enough for what i need.

Or maybe... just maybe... once i turned the arduino on the load caused the 5.03v to drop bellow the lower limit to power the arduino causing instability... I really dont know, i am guessing... It shouldnt be that because i am using a SMPS psu... I will try measuring voltage with a multimeter while the clock is running anyway just to be sure thats not the case...

It is relatively unlikely that 5.7 v will damage the WS2812s and given the nominal 5 V supply, the "absolute maximum" of 5.3 V is a quite surprising (restrictive) rating as it is only 6% over. Of course I am not going to advise you to exceed it. :grinning:

You really have 256 WS2812s for a Pixel clock? At full brightness (or for any instant that all three colours are on with the PWM cycle) they draw about 56 mA each, so 256 means more than 14 Amps. While you can arguably constrain the maximum brightness in software, it would be preferable to use a power supply capable of more than 15 Amps.

And by the way, the quiescent - "black" - current of these is about 1 mA each, so the minimum power supply draw is a quarter of an Amp.

And if this is a "brick" - sealed - power supply with a fixed output lead and a "barrel" connector on the end, there may be a little voltage drop in the lead also when heavily loaded.


I bought this psu because it is quite small and my clock will be hanging on the wall like a painting, so i couldnt fit a big psu.

Where did you buy it?
A link would help.

Can you post a picture of your power supply?

Thanks.. Tom... :slight_smile:

Hello again, guys.

So i took some pictures of the power supply and of my clock... i am using an arduino uno for now... i would rather fry that instead of my mega.

Paul, thank you for your reply. Yes i am aware that the screen can draw up to 14 amps or so at full brightness but i dont intend to get anything near that since it is a clock and it should be rather discrete. Also the led strip i used is ws2812b, not "ws2812s". I looked at the datasheet and its absolute maximum rating is +3.5v to 5.3v so yeah, my 5.7v psu worries me a bunch.

TomGeorge, I bought it online in an website similar to ebay. Here is a link to the psu specs at the manufacturer's website. Seems like they dont make this particular model anymore. Also the website is in portuguese so not sure it helps much:


I doubt that 5V 10A would be possible from a small chassis like that and through light gauge wiring on the output.
When you project began faulting, what did the 5V supply output measure?

Thanks.. Tom... :slight_smile:

Thank you for the response Tom, i dont know… I will measure it and post the results later. I didnt plug the ws2812b led screen though, just the uno, the rtc and the common max7219 8x8 led matrixes… I am reasonably certain the current drawn was quite small since i can power it all through the USB port of my PC with no issues whatsoever.

Happy new year guys!

So it turns out a had a bad connection to the rtc module, thats why the arduino was showing nonsense data after a couple of hours. Strangely that didnt happen when i switched back to my bench power supply, but yeah, once i took the rtc cables out of the breadboard and soldered them to the arduino i had no more issues. Well, thats a newbies mistake, but i have to say i am glad the project fixed itself in a sense. :wink:

I will try Crossroad's suggestion later today (putting a power cable and a diode in each row of the ws2812b led screen) and post the results and photos here. I believe it will work fine.

Thanks again to all of you guys and a special thank you to Crossroads for his insight.

Cheers guys.

Thanks for getting back and telling us your solution, it helps with the thread.
You can edit the title of this thread and add [SOLVED], this will indicate that the thread is complete with a solution.

Happy New Year mate… Tom… :slight_smile: