EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: max_torque on October 14, 2021, 07:42:12 pm
-
I need to generate and monitor some form of low voltage analogue or digital signal to enable me to reaosnably robustly determine the condition of a wire "safety loop" ie has that loop been broken or not.
My initial plan was to generate a constant current source, lets say 10mA, using enough input voltage to be able to push that 10mA through the worst case normal impedance of the whole loop (tbc). Feed that current out to the loop and back again, and use some form of common mode current transducer to effectively sum the outgoing and incomming current to zero. If the loop is broken or damaged, that sum would not be zero, and the CT would reach a treshold to trip a comparitor (i need to avoid any s/w in this solution) that then opens a set of force guided relay contacts to signal that open loop
Obvious failure states are the loop shorted to ground, or to some other voltage, so the loop in/out needs suitable protection
I could generate a completely isolated voltage source to drive the loop, so the host systems supply ground / earth is not the reference for the loop, which obvisously helps prevent current draining away somwhere it shouldn't.
What would be the best way to sum the IN/OUT current? A common mode choke with some sort of hall effect sensor to measure the magnetic field in the choke?
Any ideas people?
-
This is getting towards the industrial "safety integrity level" as in IEC 61508 etc. To do it properly needs much deep thinking about all possible failures/threats. Not much point having a monitored fail safe control loop if the controller/IO watching it can have single point failures its self.
Summing currently just sounds like expanding the analysis tree unnecessarily.
-
Alarm systems use simple constant voltage and one resistor on source and another resistor on sensor. Divide by 2 resistive divider.
They sample with A/D converter directly or use window comparator.
That way you can detect 3 states: short circuit (sensor contact activation), open (cut wire) and normal (supervised) state.
You can galvanically isolate after that circuit, if you need .
Many thermocouple front ends have broken wire detection. It is usually simple resistor (high value) connected to input and some voltage. It doesn't have to be precise at all, just to generate obvious "out of range" voltage on input so we know thermocouple connection is broken. RTD /NTC front ends also can detect too high voltage as broken connection(no sensor).
But as Someone said, you need to create (even in a form of informal write down) risk analysis what failure modes are possible and account for all that. Make sure that protection circuit is simple, fail safe, and has safe failure modes..
-
I should stress this is NOT a full SIL/ISO rated safety system.
It's actually a short term solution whilst a proper system is developed, to allow us to run the host system with a lower safety level under very controlled conditions. What i'm looking for is to get a "reasonable" level of robustness, which some basic analogue electronics outght to be able to do!
The idea behind the current source approach was to negate the somewhat variable impedance of the safety loop itself, which passes through lots of connectors and along a long length of wire.
Summing the in and out current basically fairly simply ensures that current that goes out, comes back....
My thoughs with the common mode current transformer mean that any current imbalance could pretty easily be sensed by a simple comparitor style circuit?
-
In which case just run constant DC current of 10 mA trough whole loop and one optocoupler or simply a relay coil. Whenever and whatever breaks the loop signal STOP. No need for complex signal or special provision unless it's explosive atmo or something..
-
With a simple DC current loop it would make sense to generate an isolated current, ie use an isolated power supply (DC/DC), and by that method avoid the obvious failure state where an external short of the loop to a supply voltage makes the loop integrity look good when it actually isn't?
-
With a simple DC current loop it would make sense to generate an isolated current, ie use an isolated power supply (DC/DC), and by that method avoid the obvious failure state where an external short of the loop to a supply voltage makes the loop integrity look good when it actually isn't?
Absolutely isolated current source and then detection as I said. I forgot to point that out, sorry..
But that can be simple isolated DC/DC converter module.. Low current too... 24V