Rogue resistance between Vout and GND in LDO circuit

Question

I'm currently developing a circuit as seen here:.

I had some suggestions to check my Iout to ensure that I wasn't drawing so much current that I was causing an overheat. After a little checking I believe that my circuit is way under the threshold (~22 mA.)

I've managed to solder up PCBs that have worked, but the vast majority don't. Today when double checking for any strange shorts, I noticed a rogue resistance between Vout on the LDO and GND (~60-100 Ω.) That resistance doesn't exist on the PCBs that do work.

Note: I've tested this circuit with a couple different LDOs and had the same results on each.

Is there a way for me to further troubleshoot this issue? (I don't really know where to start looking for component failure etc.)

Or perhaps any suggestions as to what's causing this strange behaviour?

Is there perhaps an occurrence where the first connection of the circuit is causing something to break?

Answer 1

First, disconnect power to the circuit, then short the IN and OUT pins of the regulator to ground.

Now, check your resistance from OUT to GROUND and see if it still gives you a low resistance.

If you still find that 100 ohms, then start removing components from your PCB.

1. Remove one component that connects to OUT from the regulator.
2. Check the resistance.
3. If the resistance is still low, go to step 1.
4. Repeat until all components are gone or until the resistance changes drastically.
5. The regulator itself is, of course, one of the components you'll remove.

If you remove all components and still have low resistance between OUT and GROUND, then the PCB is bad or you've accidentally made a short circuit on it somewhere.

---

If you determine that it has to be the PCB, then you can destructively find the bad spot.

1. Get a high voltage (50V or so) power supply that can deliver a couple of amperes.
2. Connect the power supply to the OUT and GROUND pin pads for the regulator.
3. If there's a short on the board somewhere, it will either pop and blow out or get hot enough for you to locate it.

Wear safety goggles and keep your hands clear of the board while doing this "test."

---

Now that you've found the culprit, and determined that it dies at poweron, you can start looking for the real cause.

After a quick run through the datasheet, the only thing I can find is that it might be the capacitors in the charge pump.

This chart from the datasheet shows the recommended minimum values for the charge pump capacitors given the supply voltage:

Given that you said that you are converting the circuit to run on 3.3V and that it was designed to run on 5V, maybe the capacitors need to be swapped out.

The datasheet doesn't spell out what will happen if the capacitors are too small. It may be that the charge pump just has to much ripple on the output - but maybe it causes too much load on some switching transistor if it tries to get 5V from 3.3V with too small capacitors.

In any case, the datasheet makes a point of ensuring that you have at least the minimum and that you should allow for aging and temperature effects.

That's all I can come up with. Maybe that's it, or maybe it's something else entirely.

Answer 2

My educated guess is either that:

1. You have damaged (perhaps by ESD) MAX3221 chips that are creating a short across the power rails.

2. You have a bad batch of PCBs (it's been a very long time since I've seen hair shorts on PCBs, but I have seen them from a Taiwan maker many years ago and they were in the 10-100\$\Omega\$ range. If you've purchased claimed 100% tested PCBs from a reliable supplier this should not be true, especially in any number, but it's still possible.

There are other possibilities such as shorted ceramic caps with less probability.

If the MAX3221 is getting hot with a 'stiff' supply, that will be the answer.