I'm currently using a buck-boost converter (generic XL6009 buck-boost like this one) to power a router and fiber terminal from a 12V battery backup system. The reason for using the buck-boost converter is that the 12V battery backup system doesn't regulate the output, so the output can vary between 11V-13.8V depending on whether the battery is charging etc. and I'd rather err on the safe side than assume the equipment has regulators that are OK with that voltage range.

The problem is, the buck-boost converter's inductors and caps are getting very hot, while the driver is running much cooler at a slightly warm temp. In the image below, the circled parts are the ones getting very hot. I measured the current draw on the battery side while in use and it only ran at around 1A @ 12V output, which according the the converter's specs shouldn't be an issue. I've been running the converter as is for around 3 weeks now without any issues, but given that this is a battery backup system I'm worried that the increased heat will further decrease the converter's efficiency and I'm also worried that this could damage the converter prematurely in the long run. Is this normal for a buck-boost converter at these power levels or is the unit possibly defective/out of spec?

Parts of the buck-boost converter getting hot

  • \$\begingroup\$ I use these red ones qskj.cc/shop/product/… they are basically the same thing, and they also get hot above 10W \$\endgroup\$ Jul 11, 2020 at 9:20
  • \$\begingroup\$ Where's the schematic for the little switching board? Where are the data sheet links to the inductor? \$\endgroup\$
    – Andy aka
    Jul 11, 2020 at 9:40
  • \$\begingroup\$ @Jasen - is it also the caps and inductors on your board that gets hot? I assumed that it would be the driver rather than the caps/inductors getting hot, hence the question. If yours is getting hot as well then that's probably good news. \$\endgroup\$ Jul 11, 2020 at 10:13
  • \$\begingroup\$ Yes, capacitors ahd inductsr get hot, diodes too. after several iterations of upgrading the part that's wasting the most energy all parts get equally hot. \$\endgroup\$ Jul 11, 2020 at 10:39

2 Answers 2


I've tried some cheap "LM2596" modules and their main problems were: caps' ESR was too high, inductor was saturating, and "LM2596" was counterfeit. So, basically, everything was wrong. Sure your module uses a different chip but considering the price these are usually sold at, I don't see any reason why they wouldn't be the same kind of garbage.

High-ESR capacitors used above their ripple current rating get very hot, which is a problem because they will eventually die, which will cause large output voltage ripple which could harm whatever device it is powering.

Since your router takes 12V input, here's what I would do.

First, check if any parts on the board actually use 12V. You can do this by looking at datasheets, capacitor voltage ratings, or looking at the traces and continuity testing.

If it is a DSL modem/router the analog circuitry could use 12V, I also see a 12V relay on the photo, but these will work fine between 11V-13.8V.

However, most likely there will be a buck converter to step down 12V to whatever voltage the logic chips uses, like 3.3V and some lower voltages for the CPUs and big logic chips.

If this is the case, you don't need to feed it accurate 12V. Just check the buck converter and caps can handle your battery max voltage, and power it directly from the battery with a fuse.

  • \$\begingroup\$ Thanks for the info on the modules, if they tend to be that bad I'll look into another solution - I've contemplated opening up the routers to check if they'll be fine with the increased voltage but I didn't get around to it, guess its worth a shot. +1 \$\endgroup\$ Jul 12, 2020 at 10:33

The problem with your system is you are using a sepic converter (2 inductor system) & disadvantage of sepic converter is that it is problematic at current higher than 1A. For higher currents a 4 switch buck-boost will always be the better solution. SEPIC (as well as inverting CUK type ) switches VIN+Vout on its switch node and carries the large du/dt EMI problem. And you switch more current than Iout. That to a much larger Hot loop which includes the SEPIC coupling capacitor. In essence, you store the complete transfer energy in the inductive components so those inductors have to be sized that way.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.