no text

[daywalkerdha]

no text

[NiHaoMike]

For overvoltage protection, just use a SCR crowbar.

[C]

The nice thing about an SCR is that once it's on, it stays on until current drops below ___ current.

Yes, most power supply output protection circuits blow fuses (not the bench ones), but you need to remember that if the power supply was working properly, it's not needed and will never blow a fuse.
That said, try using the idea: turn the SCR on when  _____ to keep the disconnection going until unplugged.

C

[C]

When you look at the good protection circuits, they work to protect what needs the protection, no matter what.

Dont want to switch a fuse every time, so it has to be able to reset. 

I was trying to get you to think, Turn on SCR on Error and Reset by removing it's power source.

C

[NiHaoMike]

Thanks but that's not really a viable solution for my application. Dont want to switch a fuse every time, so it has to be able to reset. Also i don't want the large current through the connected device. + It's not reverse voltage protected or am i missing something?

Use a self resetting circuit breaker instead of a fuse.

[SeanB]

What happens if it fails? Ask what will cause it to fail, and try to design appropriately. The relay needs to energise to enable, not the other way round.

On aircraft there are commonly ground powering connections, where you plug in either a 28VDC supply or a 115VAC 400Hz supply. Both have sockets with a signalling pin, that engages a relay when the plug is inserted ( powered off the battery bus) switching the power onto the internal buses. They both have protection boxes, the DC one is a overvoltage relay ( reverse voltage as well by a big fuse in the line) while the AC one has a 3 phase transformer and monitors, overvolt, undervolt, overfrequency, underfrequency and loss of phase ( 115VAC 400Hz limits are 105-125VAC 380-440Hz) which energises a relay if there is a fault. Problem is the transformer does not provide enough power to actually energise the relay if you plug in 380VAC 50Hz power instead. This being so it was passing this through to the internal buses. Very expensive mistake.

[NiHaoMike]

The MOSFET should be able to handle the reverse current.

[C]

When you are looking at a circuit you are designing, it's easy to get locked in to thinking only on the good side, I got that covered.

One instructor of mine made it a game.

The Doctor Evil test of your circuit.
The designer always gets to go first.

Doctor Evil rules:
Doctor Evil  can see all your circuit, but can not make physical changes to the circuit.
Doctor Evil  can play games with the circuits inputs and outputs.
The scoring system based to how much damage Doctor Evil  could get the circuit to do to itself.
You lost less points for a easy to replace part like a fuse in a fuse holder. On the fail side were blowing pcboard traces off the pcboard, Capacitors exploding, damage to a semiconductor. 

Doctor Evil  won at first with simple sequences like the following.

Turn on power supply, increase power supply to 10% above normal, Turn off power supply for 10 ms., turn on power supply again to normal.

Turn on power supply for ___ time, Turn off power supply, short power supply input for 10 ms., turn on power supply to normal.
 
C

[C]

I do not see anything wrong.

You know the details about where this circuit will be used. So you are also in the best position to think of what could go wrong.

In SeanB's example, someone forgot or did not think of a 400 Hz transformer connected to 50 Hz. I have worked with 400 Hz power, a lot of people have not. I see no transformers in your project, so transformers probably does not apply. In the example, Someone connected "Power" to power. They may have looked at the voltage but something went wrong on the 400 Hz bit. A 400 Hz transformer on 50 hz was before the protection circuit and the protection circuit failed to do it's job. It is not a real surprise to me that someone would miss the "400 Hz". Should the designer of  that 400 Hz powered equipment thought of the 50 Hz/60 Hz world around it and how poor a 400 Hz transformer is on 50 Hz before his protection circuit?         

Your plot shows A  Doctor Evil test of your circuit, and I see nothing wrong in the plot, and unlike some you went well past where some would stop, Good job.

When you get this circuit surrounded by more circuity , look again so you do not get caught by something before or after this circuit like in SeanB's example.

I hope you design more in the future!

C

[jeroen74]

A relay takes quite some time to open and close. Can your circuit survive excessive voltages for 20ms? For short transients I'd include something like a TVS or varistor.

[SeanB]

C, the big mistake was having IDENTICAL socket outlets next to each other, only labelled differently. Then tell the first day apprentice to plug it in, giving him a cable with a plug on the end. As the socket is a common 3 phase mains one, he went and plugged it into a socket on the wall. 50/50 chance and it was wrong.  Needed to replace all the avionics, all the wiring and all the instrumentation on that airframe, along with all the connectors. They turned it on, plugged it in and had the ground cooling running then went to lunch, as the onboard avionics take around 20 minutes to do a cold boot. Came back and saw the smoke pouring out the plane. The sockets were changed the next day on all the walls and on all the cables, along with an urgent signal to all other users to do the same and reply as to compliance state.

[C]

SeanB,

I have been got, I think any one around long enough has.
like,
  The high temp test where you test a circuit and forget that it was built for a bench test and used lead solder instead of the hard to solder silver solder and falls apart during the test. Created a lot of plastic globs that still worked when we needed a quick part test and did not  have the correct package. But It's Hard to find the line that gives good answers the "Will this work at 250C" line in a data sheet.   

 or
  The backup battery for a cesium beam clock that a bunch of us thought the yearly flying clock maintenance team changed and tested,  This was the master, standard procedure was to un coil an extension cord hanging there just for that purpose and get it power, and of course it was important so it;s normal power came from a no break power source. That day everything died for a bit   That old tube in the clock really does not want to get back up to temp and operating with out frying it's self first. Then after hours of constant baby sitting, getting it going, I needed to get it back having a 1 second pulse when it should. That clock was known the be right on, It was un-real the number of calls we were getting. It was a monitor reference for Loran C med. Just to make things more fun the monitor receiver needed a really great 1 second pulse just to let you try to get in sync with Loran C once every 17+ minutes. The next day got a message " never heard of any one getting a tube that old started". After A special flying clock visit, and I missed the second mark by 3.2 micro seconds. Not as much money as your's for the fix, but shouting and screaming was very loud and far ranging. Just for those wondering, A new tube was $25,000 an worked better with a little age.

C