Author Options:

I need help with short circuit detector / electronic fuse. Answered

So I'm making a variable power supply with an LM317 and I've decided to add a short circuit protection, I know the LM317 has an auto shutdown feature but I thought this would be a nice feature and it would allow me to leanr more about circuits.

The short detector is quite simple, It's a PNP transstor connected to the output. When the output is shorted the LM317 can't supply enough current and the load sucks everything it can, so it sucks the charge of the capacitor, causing the PNP transistor to activate and turn on the relay and the LED. This connects the PNP to ground, making it be permanently oopen until the circuit gets reset.

It works flawlessly from 16 to ~9 volts, but I go under ~9v (I don't know why) the PNP activates and the relay flickers very fast, I tried several things but none fixed the problem.

I have plenty of PNP nad NPN transistors so if you can come with a better and more stable design I would be able to try it, I'll be all the afternoon trying if I can figure this out and posting some updates.

Here's a video of how I want it to work.


Look at the image for more information.

Thanks in advance.



Best Answer 3 years ago

Deleting questions is a selfish act ?

Your ckt only detects a sudden short and then lets the relay go green again because you cannot sustain DC base drive through any capacitor.

The ckt will not trip if a slow increase in current occurs because the PNP base is capacitor coupled.

Deleting questions is a selfish act ?


Sorry, I forgot to include the resistor (red) in the picture, it connects the base to ground when the relay has been turned on, to keep it in that way, until the reset button is pressed, opening the circuit.

I finally made it work for all the range of voltages, but there's a problem, and it seems the transistor is to blame.

If I connect the circuit as shown in the picture it works well, until as I said, I lower the voltage and the relay starts flickering (a transistor problem I'll explain why), I can solve that by increasing the value of the resistor connected to the base (the red one), but as I lower the voltage it's clear the transistor gets progressively active and let's the current flow through it.

Of course, if I don't use the resistor and depend only on the capacitor there's no flickering at any voltage, but there's no way to keep the relay on once a shortcut has been produced. Also, when the voltage is too low, the short is too weak to cause a significant perturbation at the other side of the capacitor, causing it to hint a bit but without reaching the fully active state.

The reason the transistor turns on when it shouldn't ( I guess ) is because the emitter gets 12 volts while the base only gets 1 at it's lowest voltage, if I increase the red resistance the transistor lets pass less and less current. 60 kOhms is the ideal value, any less than that and the transistor will activate and the relay will flicker, any more and the base won't detect the shortcut.

I tried with two transistors, the 547 and the 2904, they are practically the same.

I feel like I'm so close to find the solution but I cant figure it out.

I'm quite obsessed with this circuit, I feel like I'm too close to abandon it now.

Thanks for your suggestion, I already had saved that picture into my circuits folder.


When the voltage is lower then 12 + 3 = 15volts the 7812 stops regulating.

Put another way the 7812 regulator needs an input of 15v to 25 volts to deliver an output of 12vdc

The 7812 is getting contstant 19 volts because it's connected to Vin, the only voltage that varies is the one outputted by the 317, and as far as I know the 7812 doesn't gets affected by it.

The input can supply 2,1 amps so I guess there is no problem related to the circuit draining too much current.

I also thought the 7812 was not doing it's job, but I measured it's output while changing the output of the LM317 and the voltage was always 12 volts.

I'll try connecting the emitter and base of the PNP to the output of the LM317, that way the base and the emitter will have the same voltage. I'll connect the base of an NPN to it's output so the pnp does the detecting job and the NPN the switching.

Picture related. I won't be able to try it until sunday afternoot though.


Most relays are DPDT and that allows you to make it a latching relay.

BTW your new ckt does not need the reset as when the short is cleared the relay will clear on its own.


The relay I currently have has only three pins, not six.

Also, after doing a mental test of your circuit, I think it would stay on because once the short is made the relay supplies itself the current it needs from the 7812 without being affected by the transistors. When the short ends the PNP transistor comes back to its original state because it is not connected to ground, but the relay is still on.

I guess they are basically the same circuit, but yours would be pretty useful in case I needed to switch larger components since the NPN transistor would be free from having all that load passing through it.

I gave up because I realized I could use that space to add another independent output with another LM317.

However I've learnt how to use an OP amp with a small and high power resistor to measure the current.

Here's the picture of it, I also made a short video:



You're the guy who keeps deleting his questions, and erasing the work of other people answering them. Find your own answers.

Sorry if I did wrong I'm new to this part of the webpage.

You've basically asked the same question three times now, people have advised you, and you've deleted their work. I can't see why anyone will want to help you.

The other questions were about the linearity and a question about an electronic fuse which was useful to build this, now I'm troubleshooting why it doesn't works.

I remember your answer about Schmitt trigger and the triac, I simply followed the easiest path.

I have been also looking into OP amps to make a comparator and find the current flowing.

Now no-one ELSE gets the benefit of my comments in context, you see ? Deleting questions is a selfish act. And your "easier" path ? Is it working ? There's a reason I suggested the Schmitt trigger ....