Boost converter magic smoke caveat


Neil Smith G4DBN
 

A brave experimenter chum of mine reports that they used a neat-looking cheap boost converter (like ebay item 371926878885) entitled "2-24V to 5-28V 12V 9V DC-DC Boost Converter Step Up Module CHIP 225B" to make a 12 to 28V step up supply for some 28V relays.  During testing, the voltage to the board was increased to 15V, which is well within the published range. However, at this point, not only did all of the magic smoke come out of the unit in a hurry, it was pursued in haste by properly hot actual magic flames.  Once the smoke cleared and the fire had subsided somewhat, our hero then tried another board and found that it had naff-all bypassing, and made a decent 3MHz transmitter to boot.

Once tamed with suitable capacitors and a carefully-limited supply, the thing did work and would feed four 28V relays happily. I suspect that the "catching actual fire" aspect of this design might make the devices less than attractive to potential purchasers, and colleagues on the group may wish to consider this caveat when browsing the Bay for Shiny Things.

I take a different approach to this myself.  I use a separate 28V power supply, isolated from ground (but with a TVS diode clamp and resistor to ground on each leg so it doesn't float about too much) and send that up a pair of wires to power the relays. I am SO old-fashioned.

Neil G4DBN


Andy G4JNT
 

I had quite a few crappy-cheap Chinese boost modules, probably the same heritage as these.  The poor decoupling I was prepared to live with, it's easy enough to add more, and you'd probably need to anyway, even if it was reasonable to start with.     But one of my modules did something similar, smoke and a capacitor turning brown and about to burst into flames.  It turned out the output voltage adjustment presets on all of them were poorly soldered.   And if feedback on a boost regulator fails in the wrong direction output voltage shoots up to a theoretical infinite volts.   I guess this is what happened in your case.

Check the connections of the 10-turn preset, that they are properly soldered to the pads and the plated through holes really are properly plated through.    They are 'simple-switcher' designs so you can get data sheets easy enough to see what components should be around the device and the connections.   Those simple-switchers are nice chips, and as they're quite chep are probably not likely to be forgeries.   Can't say the same for the rest of the components on a cheap Chinese  module, just be prepared to replace them.

Oh, and something else I learned the hard way, DO NOT use tantalum caps as extra decoupling where significant ripple current will flow - like the input on a boost regulator.  That poor cap went all brown and started smoking.
In fact, you're not supposed to connect them across low impedance sources anyway - so no tantalums on 12V inputs!  



Virus-free. www.avg.com


On Fri, 27 Nov 2020 at 17:03, Neil Smith G4DBN <neil@...> wrote:
A brave experimenter chum of mine reports that they used a neat-looking
cheap boost converter (like ebay item 371926878885) entitled "2-24V to
5-28V 12V 9V DC-DC Boost Converter Step Up Module CHIP 225B" to make a
12 to 28V step up supply for some 28V relays.  During testing, the
voltage to the board was increased to 15V, which is well within the
published range. However, at this point, not only did all of the magic
smoke come out of the unit in a hurry, it was pursued in haste by
properly hot actual magic flames.  Once the smoke cleared and the fire
had subsided somewhat, our hero then tried another board and found that
it had naff-all bypassing, and made a decent 3MHz transmitter to boot.

Once tamed with suitable capacitors and a carefully-limited supply, the
thing did work and would feed four 28V relays happily. I suspect that
the "catching actual fire" aspect of this design might make the devices
less than attractive to potential purchasers, and colleagues on the
group may wish to consider this caveat when browsing the Bay for Shiny
Things.

I take a different approach to this myself.  I use a separate 28V power
supply, isolated from ground (but with a TVS diode clamp and resistor to
ground on each leg so it doesn't float about too much) and send that up
a pair of wires to power the relays. I am SO old-fashioned.

Neil G4DBN







Brian Flynn GM8BJF
 

I have used these Chinese SMPS boards widely in different projects.  All of the buck types supposedly based on the LM 2596 IC that I have come across used what I took to be a re-marked (not fake) LM2576 chip. This is a cheaper part. The telltale sign is the frequency that they operate at. The real 2596 runs at 150kHz wheras the 2576 runs at 50kHz. Also the markings often wash off with IPA.;-) .It is just a creative way of adding value to the product.  My solution has been to source some LM2596s from a trusted supplier (RS) and just replace the IC. Some simple soldering! While I am at it I usually replace the caps as the ones supplied normally have poor ESR ratings and run hot under a decent load. If the boards are sourced with this in mind it is still cheap for the PCB! I never pay more than 90p for one.
The boost converters using the XL6009 I have found excellent. That (chinese) chip is much more efficient than the NS/TI equivalents as it uses MOS devices as the main switching element with a lower on resistanbce than the aforementioned bipolar parts. That assumes the PCBs are properly soldered, but that is up to "goods inwards" to accertain!

73s
Brian GM8BJF