When do I need to worry about noise from my power supply?

[v0idnull]

So I'm looking into buying a cheap bench power supply from Amazon (I'm specifically looking at this: https://www.amazon.ca/LETOUR-Variable-Regulated-Adjustable-Alligator/dp/B01MQWQTI4 and I'd like it for the weekend). The main complaints I see about cheap power supplies is that the output is noisy. I'm also tempted to try to convert an old 300w ATX PSU into something usable, but that too is a switching power supply.

So when does that become a concern for a simple hobbyist like me? What I'm mostly interested in working on are LED controllers (especially with regards to PWM dimmers for 12v LED strips), and simple motor-related projects (eg I want to make a cord retractor and an automated dolly for my camera), powered by arduinos and/or raspberry pis. I'm just having a hard time trying to ascertain whether the problems these kinds of cheap supplies have, are going to be problems for me and I'm not sure how to evaluate it. Any advice welcome, thanks.

 

[Audioguru]

Cheap usually means poor performance or a short life. It is no-name-brand Chinese so the guarantee is useless.
Its load regulation is good enough but what happens to its output voltage when your mains voltage fluctuates when the air conditioning turns on and off?
Its spec's have no details. It says its ripple is only 10mV but is that with no load then the ripple increases to who-knows-what (10V?) when loaded??
It has no safety labels. Is it certified not to electrocute you and not to burn down your home? Nope.

[tron9000]

What will be a concern is how long do you want it to last. Most of these supplies die within 2 to 3 years?

This one looks like a switching regulated supply. Cheap to produce and MAY be easy to repair if it does pop its clogs. But can you be bothered with that?

Plus side its light and compact.

For switch mode supplies, ripple may become bad when load increases, 100mV? 1V? who knows? only states <10mV ripple, but doesn't state at what load. ask yourself why? if you're powering a 5v project and has 500mV ripple at 2A say, then anything digital in that porject could start behaving incorrectly.

[CJay]

I doubt the 'noise' will be a problem, but if you're bothered about being able to repair it then it's worth hunitng out the schematic now, even if you can't find it you can help yourself by taking a series of good, in focus pictures of the internals so you have component values and wiring layouts.

[tautech]

Coarse and Fine pots will drive you nuts.

It's the bog cheapest solution for what should be an encoder or multi-turn pot.

Get this one instead that has 5 turn pots:
https://www.amazon.ca/HUKOER-Adjustable-Switching-Regulated-MCH-K305D/dp/B07B9TD4TC/ref=sr_1_2?ie=UTF8&qid=1523981487&sr=8-2&keywords=K305D&dpID=51PtvdyQ7tL&preST=_SY300_QL70_&dpSrc=srch

As for output noise, yes all SMPS have some and any remote circuitry should have local onboard capacitance to manage rail noise from either the surroundings or supply.
Digital circuits are the most at risk if switching thresholds get to see the noise but if you've done your bit in the design with enough fast bypass and bulk capacitance you shouldn't have problems.

[IanMacdonald]

Electrical noise is unlikely to be a problem for your type of application, that is more of a concern for audio, radio or precision measurement. Fan noise can be a problem with the cheaper linear PSUs as they often rely on a small heatsink and powerful fan for least cost construction. The SMPS type are considerably more efficient. Safety can be a concern with cheap SMPS though, in terms of input/output isolation. (It is generally not desirable to earth the output of a testbench PSU, so any leakage or short could be an operator hazard)  With linear supplies if the transformer is of decent construction that is the main thing safetywise.

Oh, and the controls often go bad on the very cheap ones, so one with a 5T or 10T voltage pot is probably a better bet as the controls may also be of better quality and last longer.

[Old Printer]

For the last year or so I have been assembling my first lab/bench. A few months ago I decided it was time to get a good power supply. To date I have collected six of the damn things, all used off ebay. I think that well built name brand PS's are a good thing to acquire second hand if you shop carefully. They hold up well and are a very good value. Including shipping I have paid no more than $75 and as little as $30, though I was very close to buying a nice Agilent for about $130, but I passed and still wish I had bought it. With the bankruptcy of the ITT schools there has been a lot of decent equipment around, though it is drying up. The GW Instek 3020's are plentiful for about $50+ shipping (US$). The BK 1651 is a mid-range unit with 2 variable + a fixed 5V and no fan. These have been going for about $30 + ship. I also picked up three of the older HP 6216A's for $40 or so ea. They are a really nice little single output 0-25V 500 ma unit, very small and they stack nicely. Two make a great + & - supply for op-amp circuit development. It does take time though, and maybe a little working on, but service manuals and schematics for all of these are available free or cheap. One of the BK 1651's showed up with a wonky B channel, but after a good bit of study I replaced a zener diode for 20 cents and it is up and running and I learned a bunch. I think this is a great way to get some quality bench supply's for not a lot of money. Sorry, this got longer than I had intended.