Cheap 10W LED from eBay – worth it?

LEDs are everywhere. They are getting so popular nowadays that you would find them in every electrical equipment outlet. But what about the loose chip (COB) LEDs available off of eBay for literally just a couple of bucks? What about the huge and cheap 10W LED modules? I decided to find out. I needed some for lighting my lab up with solar energy (read more about going solar!).

But are they good enough?

testing cheap 10w led

The cheap 10W LED chips

It is not really a “chip”, but is commonly called so because of the very common SMD COB package (Chip On Board). These can be mounted directly on aluminum PCBs for proper heat dissipation. The ones I got from eBay are priced at about only about Rs. 30 per 10W LED. That is about half a dollar! Now, for obvious reasons, it is not reasonable to expect the LEDs to be the best available. However, do note that high cost does not mean high quality in case of power LEDs. I have seen sellers selling the same defective LEDs for higher price and people falling for it thinking it is high quality stuff.

By quality of LEDs, you usually mean some indicators like:

  • Leakage current
  • Matched of internal LEDs
  • Lumens per watt
  • Characteristics vs temperature
  • Packaging

Though the packaging is usually standard these days, it matters when some manufacturers make it difficult to mount LEDs by not providing a proper mounting point, which in turn makes heat dissipation a pain in the back.

As for the leakage current and matching of LEDs, the situation is really bad with cheap LEDs. Do note that it is natural for about 10-20 % of LEDs to have internal issues with leakage across some elements, but when over 40% of LEDs have a mismatch internally, you can assume you have been buying the wrong products from the wrong person.

What causes internal mismatch in LED chains? Well, there could be too many reasons from mechanical strain and shock to thermal shock and ESD. But either way, most defects are created from manufacturing and initial handling itself. Lack of strict quality control adds to the number of faulty LEDs.

What “bad” LEDs look like

You can identify “bad” LEDs from a lot of cheap 10W LEDs by supplying a small current and a voltage slightly higher than the forward voltage of the LED. For example, for the 12V, 10W LEDs I had, I applied about 10V with a small current. What do you get? Either a perfectly matched, uniformly lit LED chain or a mismatched, non-uniformly lit LED chain. In the latter case, you know you have leakage across internal LED chips in the 10W high power LED. Examples of “good” and “bad” LEDs are shown here:

defective unmatched unevenly lit cheap 10w led

Evenly lit, matched 10W LED (below) and unevenly lit, poorly matched LED (above)

evenly lit high quality cheap 10w led

Should you use “bad” LEDs?

Yes, you sure can. If all the LEDs light up at an acceptable voltage range of 1-2V, you may use the LED without potential issues. But if all LEDs in a chain have died out, it is just no longer worth the current the LED chain consumes. I will write about what exactly happens in the LEDs internally when they are mismatched. They may light up but they are no longer efficient!

As for series or parallel connection – connecting faulty LEDs in any configuration will affect other LEDs connected with it in series or parallel. If you have individual drivers, it is good to wire one LED per driver if the LED is not properly matched. That will make the best out of the LED and also not spoil the other good ones (which it would, if connected with good ones in series or parallel).

Also make sure you dissipate the heat away properly when you use power LEDs with masmatches or slight leakage current. This is because the leak causes more heating (useless heating). And this needs to be dissipated. Good thermal control will also ensure good LEDs don’t turn bad. This can happen sometimes when they are made to operate at near maximum ratings.

More detailed analysis on this in a new article soon! Feel free to leave any comments below…

Leave a Reply