# How does this current limiter work? - and why doesn't it?

I have a generic Ebay DC-DC buck converter, and the current limiting circuit is not working. I find that such devices often suffer from this issue, so I figured I'd reverse-engineer it in hopes of figuring out the problem.

Below is the schematic I came up with by following traces on the board; and by measuring continuity. I believe it to be accurate.

Unfortunately, the current-limiting (upper) part of the schematic makes no sense to me. Could someone explain what is going on, please?

I see no reason to use a TS431 in full saturation; neither is it clear why the op-amp ground reference isn't connected to ground.

I assume the current-limiter achieves its objective via pin 2 of the XL4015. Perhaps the LED, in addition to indicating, serves to prevent lower voltages from the op-amp reaching the XL4015?

In addition, could anyone offer a possible explanation for the failure of the unit to operate correctly? Where should I begin troubleshooting?

Photos of the board:

EDIT: Fixed schematic and added the 0.05Ω resistor

• Your tracing of the comparators/op-amps is not correct. The TS431 provides a reference that is more stable than the 78L05. – Spehro Pefhany Jun 14 '18 at 17:56
• The op-amps also can't be connected correctly - inputs and outputs may be mixed. – W5VO Jun 14 '18 at 19:00
• Judging by the lonely via we can see on the bottom side just above the 4 stripes at the bottom-left, that board has 1 or more internal layers which you haven't accounted for in your schematic. – brhans Jun 14 '18 at 19:11
• Where is the 0R050 resistor in your schematic? It appears to be in the OUT- path and would provide current measurement feedback to the opamps. You've also drawn VOUT- as connected to ground when it's pretty clear from the pic that the pin on the screw terminal is not connected to that copper pour. – brhans Jun 14 '18 at 19:23
• Your schematic also omits the connection between VOUT- and the pair of connected opamp inputs which you drew on the top-side board photo. – brhans Jun 14 '18 at 19:43

Looks like there are a few connections missing. The low side current must be drawn through the 50m$\Omega$ resistor or through a printed trace that does something similar.