LM2576 ON/OFF pin for over voltage protection

Question

I'm trying to shut down an LM2576 regulator in the event that one of its adjusting resistors fails open. I'd like to use the ON/OFF pin on the regulator as this seemed as though it should be easy and cheap, and stops the current rather than having it dissipated as heat. The input is between 5 and 37v, and if this resistor fails, it sends that voltage all the way through to the rest of my circuit, exploding capacitors and killing the processor.

The ON/Off pin needs 1.4v to start shutting down the regulator, which i'm able to feed from the 5.1v zener as in the schematic below (I hope it's below, this is my first post, sorry if i've got it wrong!) But it needs around 5v for full shutdown, which is what I'd like. The schematic is simplified obviously, there are capacitors and whatnot in there as needed by the LM2576

My problem is that by feeding it through the zener, it only limits the output to 6.7v, I assume because this is the voltage the zener needs to give the pin 1.4v. This is enough to stop things exploding, but not really to fully protect the circuit. If i give a direct 5v or 12v or whatever to the pin, it shuts it down completely. I've tried using the zener to feed a transistor, but this just ups the voltage to 7.2ish, and as the pin needs to be pulled high, i can't just use the transistor to go to GND.

I'm sure I'm missing something really obvious, but I can't see it! Can anyone help please?!

^{simulate this circuit – Schematic created using CircuitLab}

^{simulate this circuit}

Answer 1

If you use negative feedback to EN! With a threshold of 1.3V for OVP, it will oscillate from a over-voltage condition. You may want to use a latch for this function with POR and an indicator. Or some other autonomous retry method.

Make sure your step load response is stable with acceptable overshoot.

You might consider a current limiter on the input.

Answer 2

So, from what I understand from Tony's suggestion is this. I possibly just need to use a PNP transistor rather than NPN, as this will give me more voltage to the ON/OFF pin when it's triggered by the zener:

^{simulate this circuit – Schematic created using CircuitLab}

Hmm, drawing it out the way i understand it has made me realise that this will still just oscillate the output at 5.1v again and never actually turn the regulator off. So I think I might be looking in the wrong direction really. So, does anyone have any opinion on this latching version? Its getting a bit big, but I think I've got room:

^{simulate this circuit}

If the output voltage rises above 5.1v, the zener triggers the NPN, pulling the PNP to GND, which in turn sends the unregulated voltage (up to 40v) to the ON/OFF pin and back to the NPN through the bottom resistor, turning off the regulator and keeping it off until the 40v is turned off. I've seen a few of these schematics, but sometimes they omit the cap and R2, and I can't really see any need for them. Transients from the unregulated supply might damage the transistors I suppose, but i can't see how they can trigger them accidentally. obviously they need to be rated for over 40v.

Answer 3

If anyone else is trying to do the same, this works:

^{simulate this circuit – Schematic created using CircuitLab}

It needs the pull-down R2 to stop Q1 accidentally closing, and it needs C3 in order to reset. Otherwise the residual voltage left in the power supply capacitors C2 can hold it closed even when power is removed from the system, making it impossible to reset.

Once the latch has been sent (which it does at 5.3 V), it turns off the regulator, but that still outputs enough current at 2.4 V to keep the latched closed, and the ON/OFF pin high, and so it's possible to use the regulated output from the LM2576 to feed Q2 rather than the unregulated output which could be running at 40 V or whatever.