Hi

I just have to share my exciting finds today. I found AliExpress today on this forum and found this

buck regulator:

http://www.aliexpress.com/item/Wholesale-1pcs-DC-Step-Down-Converter-DC-4-0-38V-to-1-25V-36V-5A-75W/32285636716.html

Scroll down and you will find a treasure : specs and documentation (readable)

Boy did I have to get a couple of these, and heat sink, standoffs, documentation

“Danger, Will Robinson!” Those 50V 220uF filter capacitors (in the heavily annotated picture) have a 650mA ripple current spec:

http://www.mouser.in/ProductDetail/Lelon/VZH-221M1HTR-1010/?qs=sXzuR3ywNuc25mceIPLw6w%3d%3d(That says 553mA but the datasheet shows 650mA). With a 5A output current there's a *very* good chance that the input capactitor is going to explode and possibly take somebodies eye out - ask me how I know (it wasn't this particular convertor and it missed my eye btw)!

This shows the equations for calculating the input and output capacitor ripple currents:

http://rohmfs.rohm.com/en/products/databook/applinote/ic/power/switching_regulator/capacitor_calculation_appli-e.pdfLook at the equation on page 3. You would probably find it a bit tricky to determine the inductor ripple current but its not a very significant term anyway. Assuming the inductor ripple current is 1/3 the 5A output current; with an output of 5V and input voltages between 6 and 36V the peak input ripple current is

**2.52A** at Vin = 2 * Vout and a minimum of 1.39A at Vin = 36V.

If you have a low impedance power source and *very* short, low inductance connecting leads then you may be OK as most of that ripple current will come from the source rather than the filter capacitor, but how are you going to measure it to be sure?

Bear in mind that the filter capacitor has a specified impedance of 120m ohms at 100kHz so the total source impedance of your power source and the connections and the PCB traces from the input connector to the switching mosfet needs to be less than approx 120 milli-ohms * .65A/2.52A milli-ohms , or < 31 milli ohms to keep the capacitor's ripple current below 650mA.

At a switching frequency of 300kHz the reactance of an inductor consisting of a single 8mm diameter loop of wire will exceed that, never mind the wire resistance. The convertor's switching freqency could even be 500kHz or higher making the problem worse.

I would replace the input capacitor, at least, with something with a much higher ripple current rating such as:

http://www.digikey.com/product-detail/en/EKZN500ELL681MK25S/565-4071-ND/4843881An aluminium-polymer capacitor is much smaller for the same rating but more expensive:

http://www.mouser.in/ProductDetail/Nichicon/RNU1H470MDN1PH/?qs=sGAEpiMZZMs9%2fSpGH%2fyc7JYN24Rot9XpwJjNTEwdkxU%3dNow you know why those switching convertors are so cheap - they cost reduce them to the bare minimum and then carry on removing essential parts!

Note that ripple current ratings vary enormously with capacitor type and size so you need to check the exact specs.

Alternatively, you could leave the existing capacitor and connect a suitable capacitor as close as possible to the input connections, but beware the inductance again from the leads and PCB traces.

The ripple current is much less in the output capacitor and probably won't be a problem, but worth trying to estimate from the Rohm link.

In the mean time, either limit the current output to not much more than 1A, or a tin-hat and safety vizor might be a good idea!