Something Completely Different.

[spuffock]

An interesting problem. I have a pair of solenoids driven from a PIC port with a power fet , or transistor each. The processor is glitched or destroyed with an EMP or HERF weapon. I need to ensure that the fet fails as short circuit, so that the solenoids operate. I thought maybe an SCR in parallel with the FET, and a turn of track around the PCB to pick up enough voltage to break down the SCR. Any ideas?

 

[blueroomelectronics]

Do you live in area where EMP or HERF is a problem? Wouldn't whatever killed the FET kill the rest of the circuit?

 

[spuffock]

It is a security thing. It is possible that some miscreant may use EMP to break into it. A Marx generator into a single turn at a few inches range would turn the processor back to sand. Should this happen I want the solenoids to operate. I don't care what happens afterwards. What I need is a component guaranteed to short circuit permanently if subject to a short overvoltage pulse. Any good ideas?

 

[Nigel Goodwin]

I would suggest you try doing such things yourself first, EMP is far more science fiction than commonplace!. In any case, you would need such a system to check that it works, no point trying to build something if you can't test it!.

 

[picasm]

I don't know anything about protecting from EMP pulses but I suppose you could use a " Protected" type mosfet such as a irsf3021 which gives some protection from ESD voltages.
Perhaps you could have it biased on and linked to the processor output indirectly such that the processor has to send continuous pulses to keep the mosfet turned off.
If the processor pulse output fails then the mosfet turns on.

If the mosfet also fails open circuit, perhaps a lightning protection varistor connected across it's output would be enough to operate the relay when the varistor goes short circuit due to high voltages.

None of this may actually work in practice, but may give you a few ideas.

 

[spuffock]

OK,Nigel, let's call it inductive coupling then. I'm a bit leery of operating the marx in the same workshop as anything else, I lost a TV camera to it last time, at a range of a couple of feet.
Oh well, it's a Branly Coherer then.............

 

[Nigel Goodwin]

OK,Nigel, let's call it inductive coupling then. I'm a bit leery of operating the marx in the same workshop as anything else, I lost a TV camera to it last time, at a range of a couple of feet.
Oh well, it's a Branly Coherer then.............

OK, so you've got something to test it with, so try it on a PIC and see what happens!. However, a PIC is likely to be a GREAT deal more robust than a camera!.

 

[Optikon]

If you want mechanical security in light of electrical attack, then you must use a mechanical system that is immune to electrical effects.

 

[ecerfoglio]

It is a security thing. It is possible that some miscreant may use EMP to break into it.

If you want mechanical security in light of electrical attack, then you must use a mechanical system that is immune to electrical effects.

You can reverse the mechanical action: Let the solenoids open the door (or whatever) and use a spring or weight to close it.

You may add some protection against the solenoids' driver beign short circuited (maybe fuses).

Or let the PIC output a pulse train instead of DC, use a capacitor in series to block DC, rectifier and filter capacitor in paralel with the solenoid. This way, if the pulse train stops the solenoid gets no more current.

 

[blueroomelectronics]

You could build the circuit in Faraday cage.
Just like the err crazy folks here.
http://www.unitedstatesaction.com/emp_and_faraday_cages.htm

Seriously if you want an EMP proof lock then the old mechanical locks still work.
**broken link removed**

 

[dknguyen]

The pencil vs the space pen analogy comes to mind. Apparently the pen does have advantages though- no hazardous broken lead floating around.

 

[philba]

The pencil vs the space pen analogy comes to mind. Apparently the pen does have advantages though- no hazardous broken lead floating around.

nor conductive dust which is probably a bit more dangerous to electronics than a small piece of lead. I love how people criticize NASA over the pen/pencil thing conveniently ignore the graphite issue.

 

[Hero999]

Is your circuit going to be used in a military application or something?

 

[mneary]

You really think your power source and solenoid will survive the same conditions that destroy your PIC?

 

[kchriste]

Ecerfoglio's idea of reversing the mechanical logic will also protect it in the event of total power failure. The best protection against electronic attack is the eliminate the electronics. Those SciFi movies always amuse me where the "super virus" is taking over the base computer and no one thinks to just turn the damn thing off!
Can't you buy the same radiation hardened semiconductors the military uses?
You could also build the entire thing from vacuum tubes!

 

[Hero999]

Those SciFi movies always amuse me where the "super virus" is taking over the base computer and no one thinks to just turn the damn thing off!

Since this base computer is probably in charge of engine management, that wouldn't be a good idea, the obvious solution is to install a real operating system like Linux or BSD on it instead of a Winwoes.

 

[Nigel Goodwin]

Since this base computer is probably in charge of engine management, that wouldn't be a good idea, the obvious solution is to install a real operating system like Linux or BSD on it instead of a Winwoes.

So what makes you think Unix based systems are immune from virus attacks?.

Obviously they are immune from Windows virus attacks, but historically the most 'successful' attacks have probably been on Unix systems?.

 

[Hero999]

Well I wasn't really being serious but UNIX systems are generally less vunerable to virus attacks, but point taken, no system is totally immune.

Apparantly UNIX was the first system to suffer virus attack but that's only because it was the first popular OS by contrast Windows has only been used for servers quite recently.

Many argue that UNIX has more vunerabilities than Windows which is true but the argument in itself is inherrantly flawed. UNIX represents a vast range of various operating systems whilst Windows represents only a few. Obviously that a large group of operating systems will collectively have more vunerabilities than a small group. You can't even compare Windows to Linux in this respect as the latter is only a kernel which is a component of many OSes; you can compare the Windows kernel to Linux kernel but it totally different to comparing a whole OS.

 

[spuffock]

One shot of the marx, 2joules 40kv into 2turns 30cm dia, distance 30cm.
Burnt finger on the PIC, solenoids didn't operate.
WHAT NOW?

 

[Nigel Goodwin]

As suggested by someone else previously, why are you wanting to do this anyway?. If it's a solenoid operated lock, then disabling it in this way will prevent it opening, so it's failed safe anyway?.