A cautionary note about cheap DC-DC converters

Another thread has discussed controlling a cheap DC-DC converters such as this one: http://www.ebay.co.uk/itm/NEW-LM2596-DC-DC-adjustable-power-step-down-module-/130704327784?pt=LH_DefaultDomain_0&hash=item1e6e95c468

As I've been thinking about how to power several 5V Nano/CAN nodes plus LCD etc., I too have looked at these. It appears, however, that their ratings are suspect. I'd surely like to be proven wrong.

The components, particularly the input and output capacitors, used just don't seem to match the claimed specs, at least not for any length of time, and possibly not safely. The LM2596 chip can handle the voltages and current, but not with those capacitors.

This ad specs <=24V inputs and outputs. The input capacitor is, however, 220 mfd, 35V and the output capacitor is 220mfd, 15V. To start with, that 15V output capacitor is going to have a short life if the output is set at 24V. At most, the output should not exceed 10V. The input side is questionable as well. The LM2596 data sheet says that the input capacitor must be rated at least 1.5 times input voltage (though it admits that this is somewhat conservative). A nominal 24V lead-acid battery, for example, will sometimes be nearer, or even over, 30V, especially when hooked to a charger and a 35V-rated capacitor will likely not survive for long; a ca. 50V capacitor is needed. The LM2596 spec sheet also says that the input capacitor should be able to handle transient currents of 1/2 the current draw of the converter, and they give a graph to find the approximate current capacity of typical low ESR electrolytics. From those curves, a high-quality 220mfd, 30V capacitor can handle about 700mA. Checking capacitor data sheets, I have found a few (expensive) that have higher current capacity, and many that have much lower current capacity. Which did the manufacturer of this board choose? One should probably not draw more than 1.4A from this board, despite the 2A (or more with heat sinking) claim.

What happens when an electrolytic capacitor fails. Some times not too much other than causing failures downstream; the electrolyte dries out and it stops working. Other times, however, it overheats, a little hole is burnt in the top of the can, or the circuit board underneath catches fire. Not so frightening when it is the power pack for a laptop. This happened to one of mine two weeks ago, and not in a tolerant environment - we were in hospital. Luckily the unit failed quickly enough that there was just a charred board rather than a conflagration. More frightening if buried inside something like a wheelchair.

So, a more realistic rating for this board, and it may be a convenient choice if one stays within this range, is INPUT: up to 20V, OUTPUT: up to 10V at <1.4A. As one says here in Italy "mi raccomando".
Ciao,
Lenny

LROBBINS:
Another thread has discussed controlling a cheap DC-DC converters such as this one: http://www.ebay.co.uk/itm/NEW-LM2596-DC-DC-adjustable-power-step-down-module-/130704327784?pt=LH_DefaultDomain_0&hash=item1e6e95c468

...
This ad specs <=24V inputs and outputs. The input capacitor is, however, 220 mfd, 35V and the output capacitor is 220mfd, 15V.

?

In the photo on that page they're both 35V...

It's really up to a designer of a project to always select components, modules, and power sources such that they are not operating close to their 100% ratings. I would never select power components that had to operate closer then say 75-80% of their maximum rating to allow for some 'headroom' and possible future increase loads. Device maximum specifications are almost always stated optimistically and assume some conditions that don't always reflect most normal application variations like ambient temperature changes, input voltage variation, and heatsink requirements. Even with simple components one wouldn't install a 2 watt resistor that has to actually operate at 2 watts of heat dissipation. With experience comes the wisdom of how conservative one should be for selecting components in a specific project or design.

Lefty

retrolefty:
I would never select power components that had to operate closer then say 75-80% of their maximum rating to allow for some 'headroom' and possible future increase loads.

Yep.

I bought hundreds of electronic components on eBay, including quite a few of those DC converter boards, a lot of power LEDs, etc. After the first few items that were constantly too hot to touch (some 3W LEDs IIRC) I started running everything at about half advertised power. I've never had a problem since then.

If you design accordingly you won't be disappointed.

retrolefty:
Device maximum specifications are almost always stated optimistically...

My experience is that they're absolute maximums for the device, with best possible heatsinks under ideal conditions. Not a good idea for long term use.

I think fungus is right and that the output capacitor is also 35V; I think I was fooled by the angle. So this unit should be conservatively OK with input voltages up to say 20V and current drain of about 1A. In my situation, however, a current restriction is OK, but I will have nominal 24V lead-acid batteries. I might be able to find a 50V (and, therefore, physically bigger) replacement cap though. Ciao, Lenny

I own several of these boards. I have a couple like the one in the above Ebay link and I have some that come with 3 potentiometers and are setup as a constant current regulator. The caps on all the ones I have are 220µF 35V for input and output. I have tested them with my oscilloscope and they have peak to peak voltage of 120mV with no load and about 400mV with a 2A load (tested it with some big cement resistors). I personally would not use them to power a micro controller unless the load is very low. When you are drawing +1A from the regulator it will heat up the small PCB and potentiometer. This will change the resistance of the potentiometer resulting in a different output voltage. I prefer a fixed voltage regulator to power a mcu. I have made pcb's with LM2576 and the LM2596 5V versions to power Atmega's. I quite like these old switch mode regulators as they are cheap and easy to find and most importantly much more energy efficient then linear regulators. Something worth considering when you are using batteries.
I have one pcb (constant regulator one) that I use to drive 2 3W leds. I have it setup so it gives about 600mA, has been working for more then 6 months.

As one says here in Italy "mi raccomando".

It is all about disclosure. I see nothing wrong with a product being designed to work within a subset of its capability envelope, as long as those limitations are fully disclosed.

Bajdi:
I have one pcb (constant regulator one) that I use to drive 2 3W leds. I have it setup so it gives about 600mA, has been working for more then 6 months.

What board's that? I've been looking for something similar...

This one: http://www.ebay.co.uk/itm/LM2596S-Power-Supply-DC-Converter-Module-Constant-Current-5V-35V-to-1-25V-30V-/261075646112?pt=LH_DefaultDomain_0&hash=item3cc95236a0

It is all about disclosure.

Yes. Reco and Traco cost quite a bit more, but they do give you data sheets and application notes. With "el cheapo" you're on your own. For example, here are a couple moderately priced replacements for a linear regulator. http://it.rs-online.com/web/p/regolatori-switching/0416874/
http://it.rs-online.com/web/p/regolatori-switching/6664379/
Ciao, Lenny

Bajdi:
This one: http://www.ebay.co.uk/itm/LM2596S-Power-Supply-DC-Converter-Module-Constant-Current-5V-35V-to-1-25V-30V-/261075646112?pt=LH_DefaultDomain_0&hash=item3cc95236a0

I just ordered a couple of these: http://www.ebay.com/itm/140880622800?_trksid=p5197.c0.m619

They look like a rugged version of the board you used but without the secondary voltage adjuster (does that actually limit the output voltage or just light up a LED on the board to let you know your battery is charged?)

I want to power some "10W" LEDs at 6-7W. They look like they can do the job.

I just noticed these when I was on eBay: http://www.ebay.com/itm/271099224017

Buck converter with built-in volt meter (selectable for input/output voltage).

How cool is that? :slight_smile:

After spending more than a month trying to find a decent regulator we have gone with these from rs components Recom Switching Regulator, 9 → 72V Input, 5V Output, 500mA found here --->

English link of the post from LROBBINS above and 2 5mm Pitch Straight PCB Terminal Strip with Screw Termination, Through Hole, 2 Way from the same company.

I will post back once they arrive and we have tested them over a month or two.

chrisom:
I will post back once they arrive and we have tested them over a month or two.

Take your time... This thread is almost 5 years old! what's one or two more months. I remember reading this years ago.

9-72v input is impressive!

LOL,
These seem to be working well for us, not getting over hot. we are running them via 24v and also seem to give a very steady outout.

LROBBINS:
As I've been thinking about how to power several 5V Nano/CAN nodes plus LCD etc., I too have looked at these. It appears, however, that their ratings are suspect. I'd surely like to be proven wrong.

Ciao,
Lenny

Anything cheap from Chinese suppliers should be tested for suitability.

Given the cost of return mail it is often not worth claiming a replacement although many will offer it.
There were articles in the press at one time about clone apple PSU units causing fires.
Some companies have policies in place to prevent the use of employees using their own phone chargers etc for this reason.