How to SafeGuard against EMP/High Frequency Attacks, Forced System Reset and Force System Failure

[ParmeetGhai]

How to SafeGuard against EMP/High Frequency Attacks, Forced System Reset and Force System Failure

IMPORTANT: I cannot use a microcontroller because of my living conditions whatever electronic that I own which contain firmware/software is vulnerable to re-programming attacks hence I can only work with hardware. no firmware/software.

I have the following circuit designed for a automated lock application.Can anyone think of conditions where the system can be compromised causing a power P-channel MOSFET to conduct and have a high power solenoid activated and also safe guards against EMP attacks, high frequency attacks, forced system resets and force system failure?

Any help will be greatly appreciated.

Note that the timing delay does not require precision. The circuit will be enclosed in stainless steel container hence offering some shielding. cannot use any micro controllers. Due to my living conditions any electronic gadget with software/firmware is vulnerable to re-programming/hacking hence only hardware can be used.

What is needed from the circuit is the following:
1) Manual trigger to activate the system otherwise the system remains idle.
2) A delay of >40 seconds after the system is activated.
3) After the delay is over then system activating a gate driver for a P-channel MOSFET.
4) P-channel MOSFET activates a high power solenoid for a lock and then gets disabled.
5) The initial trigger should never happen by itself unless manually triggered like in step 1.
6) Later an Alarm signal will trigger the power mosfet against
7) A override OFF signal then turns off the mosfet

Besides the solenoid, a stepper motor is also activated which is connected to a Scotch Yoke. it rotates in only one direction. which is why there is a need to turn on and off the mosfet twice. at 180 degree, it will activate the lock. at 360 degree it will unlock the lock.

The circuit should withstand:
1) Whole system resets.
2) Whole system failure.
3) High frequency/EMP attacks.

Preferred condition under such scenarios is that the P-channel MOSFET remains OFF thus preventing the lock from opening under default/reset/error conditions.

Typical use case:
1) After manual activation a delay of >40 seconds.
2) After the delay is over then gate driver activates P-channel MOSFET activating a high power solenoid lock.
3) The initial trigger turns off via optocoupler turning off the MOSFET.
4) The initial trigger never turns on unless its manually activated like in step 1.
5) A alarm activates the gate driver turning on the MOSFET.
6) The alarm is turned off via optocoupler which turns off the MOSFET.

First design:
It only used a Latch for initial trigger and a capacitor at the base of a NPN driver for delay. Some other design flaw made me consider the effects of a system reset and it turns out that the latch would trigger on by default under reset hence its omitted.

Second design:
The second design involved a 555 timer and a CD4017 (LTspice simulation attached). Everything works great except at system reset, the clock output from 555 shows a jitter which if not handled with a 1 µF capacitor leads to the system being triggered by default at system reset. If capacitor goes bad in case of a EMP attack then system is compromised hence this design was scrapped too.

Third/final design:
This one uses a 555 timer as well and a 74HC164 SIPO shift register replacing CD4017 and TC4429 as gate driver instead of a NPN (2N2222). the 74HC164 has a strong pull down at the input A and B thus requiring at least 9 V to be triggered and TC4429 has a built in Schmitt trigger which prevents noise and jitters. PS:I have used LTC1693-5 from analog.com as replacement for TC4429 because the model was giving out too much of a problem.

I have datasheets, LTspice schematics, LTspice libraries and symbols and screenshots of those schematics attached for reference here: https://drive.google.com/drive/folders/1ofFWJBdjE9LCo8oUX_3OSIO0Nte

PBSTD?usp=sharing

 

[dick_freebird]

First off, an "EMP attack" is not going to deliver a malign software load. It's "white-out conditions" with every long harness wire going "X" at once. Probably crowbar a supply (for latchup quench) and forget the other magical thinking about a clean code load meanwhile. You will either survive and reboot, or you won't.

Now next you imagine some "high frequency attack". Again this could cause malfunction, damage if the "victim" has a built-in destructive failure mode that can be triggered by pin electrical stimulus. But getting from "RF on {n} pins" to "demodulate and correctly store code bits, 100.000%"
ain't gonna happen, unless the RF comes looking like your host network and access granted.

A steel cookie can would solve RF when not in use. Your consumer PC and mobile phone have weak ICs no doubt but "antenna" length is very short compared to the systems that care about flying right through nuc EMP, and if you're close enough to the party to see handheld electronics killed, you won't be waking up yourself.

TV gets it wrong to make the story interesting.

If you are serious about physical-layer protection there are folks who try to make parts to do it. You aren't going to retrofit into modern assemblies easily and the price is fear-based.

In these designs you want criticized or improved, I first must question what, or whether you know anything about the piece-part threat response, which might well be more pain caused than prevented.

Probably said too much already in this "international forum".

 

[D.A.(Tony)Stewart]

A typical nuclear EMP has a peak electric field strength of 50 kV/m
Lead foil wrapped Sn plated brass cases for RF circuits can effectively attenuate RF (not Gamma) but the I/O cables need special protection like "ESD -proof on steroids."

Any gaps can be penetrated in lids unless they are tight but the I/O cables are the vulnerable conductors all around for Gamma Radiation from a major solar flare (1 per century ? ) or a nearby nuclear explosion. The energy of the gamma rays (typically 0.1 to 10 MeV for nuclear events).

Your assumptions might be invalid if discrete H/W can be modified much easier than a fuselink PROM in an MPU. Although analog won't latch-up, it may melt.

AT89C51RD2 (an 8-bit microcontroller), has shown that they can operate up to a total absorbed dose of about 0.6 kGy (600 Gy) of gamma radiation without significant degradation.

A 3 cm thick lead shield provides a good balance of protection and practicality, reducing the dose well below the failure threshold for most gamma ray energies encountered in a nuclear event.

 

[ParmeetGhai]

Thank you for your reply. Well , let me first describe what the circuit actually does and what its intended for.
The circuit along with a stepper motor and a solenoid will be placed inside a steel container. The circuit will require a manual trigger first and then it will provide a delay of 40 seconds. this will allow me to close the lid on top. then the circuit will close a lock on inside. the circuit will then again get triggered by an alarm clock (mechanical) which will cause the lock to open. the circuit will then stop leaving the lock open.
basically I am skipping any mechanical keys, passwords etc which are on me and can be accessed by some one in my sleep.
I was afraid that someone could drill a small hole and have a high frequency-high voltage probe place near the circuit causing the circuit to fail and getting the lock to open
like how a solenoid would fry with a high powered/voltage noise/EMP which would cause the solenoid to break and have the lock to open.
from all the comments that I have been getting from different forums. it seems that that kind of EMP might not be plausible but then again in theory it could work. I believe police in the states use some sort of EMP attack to stop car electronics and military does it too but I could be wrong
I will proceed further with my design and post detailed workings of it later on
Thank you for your feedback

A typical nuclear EMP has a peak electric field strength of 50 kV/m
Lead foil wrapped Sn plated brass cases for RF circuits can effectively attenuate RF (not Gamma) but the I/O cables need special protection like "ESD -proof on steroids."

Any gaps can be penetrated in lids unless they are tight but the I/O cables are the vulnerable conductors all around for Gamma Radiation from a major solar flare (1 per century ? ) or a nearby nuclear explosion. The energy of the gamma rays (typically 0.1 to 10 MeV for nuclear events).

Your assumptions might be invalid if discrete H/W can be modified much easier than a fuselink PROM in an MPU. Although analog won't latch-up, it may melt.

AT89C51RD2 (an 8-bit microcontroller), has shown that they can operate up to a total absorbed dose of about 0.6 kGy (600 Gy) of gamma radiation without significant degradation.

A 3 cm thick lead shield provides a good balance of protection and practicality, reducing the dose well below the failure threshold for most gamma ray energies encountered in a nuclear event.

 

[kd502]

MOSFET driver with 1MOhm resistor doesn't make sense. Drivers are used to make switching fast. With high power load MOSFET can burn out when it is switched slowly.
Delay should be before the driver.

Due to my living conditions any electronic gadget with software/firmware is vulnerable to re-programming/hacking hence only hardware can be used.

Hacking without any knowledge of your circuit is impossible. If someone has access to the PCB of your device, and plenty of time to reverse-engineer it, he can hack circuits without microcontrollers too. He can make hardware backdoor for future use, or switch-on solenoid directly.

 

[ParmeetGhai]

Thank you for the reply. The 1 meg resistor to cause delay with on time of the p channel mosfet . the diode connected across it bypasses the 1 meg resistor when driver pulls its output low causing the mosfet to turn off fast. the driver has a built in schmitt trigger which prevents noise and high frequency attacks to a large extend otherwise switching speed is not needed. the delay with the on time is needed to act as "soft start" for some other module which I have not mentioned in my design.
a microcontroller can be re-programmed within seconds with physical access but you cannot make a hardware do something that its not designed to . plus I can see the wiring in the circuits and the device ICs so I can tell is something has been changed. with a microcontroller you would have to go through the entire code to see what was changed. I have a workaround for that as well but its implementation will take time.
Thanks cheers

 

[kd502]

When you slowly turn on the MOSFET power dissipation will be huge for long time. You can add heatsink but SOA of this transistor doesn't allow for linear operation over 10ms. It can burn.

a microcontroller can be re-programmed within seconds with physical access

Yo have to know type of microcontroler, how circiuts are connected to GPIO, have hardware and software to program it. Buy similar circiut and try to hack it within seconds. Hackers are not magicians.
Hacking is older, than programming, hardware hacking is possible. I saw it someone disabled alarm system without visible change. Electronic components can be damaged inside with some overvoltage spike.