Load Bank Teardowns Show Danger

[Syonyk] has been acquiring some large load banks to test power supplies and battery packs. These devices consist of a big current sink, a measurement device, and a fan. He picked up two similar-looking boards from the usual Chinese sources, both rated for 150W, both for about $30. Upon closer examination, though, he found that one was really a bargain and the other was likely to blow up.

The loads are rated for 60V and as you can see from the photos, appear virtually identical at a glance. They offer a configurable cut-off voltage and even use 4-wire measurement to avoid problems with voltage drop through the power cables.

Tearing the devices apart, he found an Intel 8051-style microcontroller. The unit that [Syonyk] liked agreed with his bench power supply by about 0.15% which is actually quite good. The other unit — the one he calls a fake — came in at just under 5% off. In some cases, the measurement was almost 10% off.

Ok, so a $30 device is 5% off a bench unit. That’s not a surprise, and you might wonder if he just happened to get a randomly very good example and the 5% number is more typical. But that’s not the case. It turns out that the fake unit also randomly shorts out the load which — mercifully — blows the fuse. It also appears it might be related to trying to use it without the two additional voltage sense wires — something the good until can do, although it reduces accuracy, of course.

Close examination of the board shows the two units are slightly different. The good board has some markings and differences in traces. The fake units also have a different firmware with fewer features. It is possible that the “fake” units aren’t fake at all, but just an older revision with some firmware issue that was later fixed. However, if you are in the market for one of these it would be wise to get the seller to verify you are getting the newer board.

[Syonyk] also did a review of a different load bank that didn’t work very well, either. He calls it the “MOSFET destroyer” which might be a new class of test equipment.


16 thoughts on “Load Bank Teardowns Show Danger

  1. chinese manufacturers surprise me all the time, everything from “how could they have made something so complex with so few parts so cheaply” to “how could they have messed something so simple up so badly?”
    this is definitely a case of the latter
    but in either case its always a fun experience taking apart some dirt cheap ebay finds

    my first experience with an ebay electronic load was a claimed 4kv 100ma load, in reality it used an 800v mosfet with probably 200v clearance and loaded anything put on it to 10 amps with the minimal load i was able to put on anything was 30mA with the slightest jerk of the knob
    but dont worry the mosfets heat was well taken care of, bolted to the side of the case … the plastic case

  2. This has been the worst thing about eBay, AliExpress, Amazon, and others: it’s not even “you get what you pay for”, because sometimes you get better than you pay for, other times you get far less than what you pay for, and you can’t even tell from the price, since all of the suppliers selling similar units will sell them for very close to the same price.

    I ran into this with Li-ion 18650 cells. Of course, you can just assume that all of the ones claiming 5 Amp hours are bullshit, and I just saw one that claims 9.8 A*hr. The problem is, well, if people will lie about 9.8 A*hr, why wouldn’t they lie about 2.5 A*hr cells, just to con the people who know better than to believe ridiculous claims?

    And then there’s power MOSFETs. If they rip off a design using 30 A MOSFETs, they just assume that it’s over-engineered by half, and use MOSFETs labeled with part numbers spec’d at 30 A, but which are only 15 A units.

    And can you trust reviews? No, you can’t, even if the reviewer is both competent and honest, because often you’ll see exactly the same review for the “same” product from multiple suppliers. One of them may be true.

    And as soon as you see an article showing how to identify “real” from “fake”, all you’re seeing are the specifics of one particular fake. One of an unknown number. An example is in the comments to [Syonyk]’s article. In the article, he identifies the “real” unit by having silkscreen showing polarity of the power barrel connector, and by the routing of one wire on the PCB. One of the commenters has found units that have the silkscreen but the “fake” trace routing, so maybe the producer of the fakes has added the silkscreen, or there are two different fakes.

    In this particular case, where there are two nearly-identical looking units, one of which works surprisingly well (more than what you pay for), and another that does incredibly poorly (far less than what you pay for). Most likely, this is because the producer of the fakes was able to copy the PCB, but couldn’t get the firmware, and did just enough development to make it sort-of sometimes act like a “real” board.

    It just doesn’t pay to be cheap any more.

    1. “I ran into this with Li-ion 18650 cells.”

      As you mentioned, ridiculous mAh claims are a dead giveaway, but I have NEVER bought a new 18650 that was actually more than half of its entirely reasonable capacity claim. They can get away with this because the vast majority of people don’t have a way to measure capacity. Short of paying a ridiculous price per cell and then MAYBE getting a actual name brand cell and not a counterfeit, the only way I’ve found to make sure to get a name brand cell in halfway decent condition is to remove them from name brand laptop batteries.

      1. I have. But they were major brand names, bought either from the manufacturer or their authorised rep.

        A123 made some absolutely phenomenal 26650 cells which would deliver surge currents of something like 35A…and they did!

        But I doubt you’d find them on Amazon.

      2. It’s been a problem for a long time. I bought some Tenergy 2500 mAh AA NiMh cells from their ebay store back in 2005 or so. The best measured about 1900 mAh and most were barely over 1200 mAh.

      3. 18650s on eBay are frequently fakes or depleted old cells from laptop batteries that don’t meet the specs stated in listing description. In such situations just make sure to leave negative feedback to seller and ask for a refund. That way you will warn other people that don’t have testing equipment and you’ll discourage seller from selling crap that can put people in danger.

  3. Old timers will note that this sort of thing is not new. I can remember scandals regarding the 2N3055 power transistor going back to the early 1980s.

    The idiots perpetrating this know damned well what they’re doing. They’re sacrificing long term trust for short term profit. As a result, nobody will pay them good money for a good product until they prove that they’ve cleaned up their act. Reputations don’t clean up that easily.

    With notable exceptions like Japan, small purchasers can’t know for certain what they’re going to when buying something from the Far East. The only way to deal with this is to purchase regularly and in sufficient quantity that the vendors realize that if they rip off the client they’ll ruin a long gravy train of future work.

    It favors big companies and large quantities. What the sales jerks don’t realize is that small companies often grow in to big companies and they remember how they were treated…

    1. This is why randomized testing is absolutely necessary, not only of your suppliers but also your finished product. Finding (and fixing) problems before the customer does is a forgotten part of most business plans.

      And the first department to get cut is usually test, often because bulshitters infiltrate it and don’t actually do the work that they’re supposed to…

  4. “Quality Chinese Engineering”

    They can make some wonderful things, but since there is no reputation attached to most Chinese brand names, they mean nothing. And there’s the rampant “do anything you can to make money” attitude that puts masses of nonfunctional junk on the market, along with the quality stuff.

    You just can’t tell. And you can’t go by brand names. Or UL/CE markings (because they counterfeit those, too)
    Even Amazon Prime stuff can be junk (but, at least you can return it)

    1. You are right. However for some products I find that they can be easily build with more certain components, decently fast but for a better quality. This is one of those. Of course, you are exchanging the ease of purchasing for some time and learning maybe.
      I built mine, but only because I wanted a dynamic load. Otherwise, I was almost set on purchasing the exact same (hopefully not fake) one… http://www.electrobob.com/psu-burner/

  5. The problem is also often times compounded. Take cloned device that should have a 30A mosfet. The outfit cloning it figures it will only be pushing that rating on rare occasion so they sub a 20A part. Only they sub it with a 20A imported part. The outfit that made that part figured that you would only run it near the rated 20A on rare occasions so they really made a 15A part and marked it as a 20A part. No doubt the raw materials the parts are made out of are similarly misrepresented.

    The problem is that the devices look very much like they should for small signals and most in pre insertion component testing is based on small signal parameters. Even companies who “test” 100% of the parts that are placed are only checking the diode junctions and not loading the parts down.

    Random full power testing is essential especially for the small scale builder who often times needs source parts through alternate channels.

    1. Yeah, the thing is that depending on how far you push some limits, you could only see a degradation in lifetime.
      But you could be pushing a component from a lifetime of 10.000 hours to 10 and all your devices will pass because you only did a 10 minute test.

      That is one of my experience with crappy components/modules. I cannot count how many switching power modules failed at low loads like 1/3 or 1/4 within months.

  6. I feel that there is an untapped market for grid-tie electronic loads, that dump the load back into the electrical supply just as if they’re a solar cell. Nearly infinite power sink, no heatsinks, and if the grid power goes out, the unit shuts down so you don’t have to worry about electrocuting a power line worker somewhere.

    1. That does exist:

      and as four quadrant regenerative power supplies, too:

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.