False Triggering Alarm

What you like DIY or standard set up?

Long wires pick up induced voltage due to lightning. Shielded cable can help. Perhaps a simpler solution is to insert RC filter closest to the input. Chose RC time constant that is as high as possible relative to the closing of trigger detector duration. My suggestion, time constant of 100~200 ms should be a good value, to swamp out lighitng pulse. RC time constant is Resistance (Ohm) xCapacitance (F) .

What kind of "system" are you triggering?

I am triggering an evacuation system. (non critical use)

You close a contact to trigger the system, that implies that current only flows when the system is triggered. There are two main drawbacks to that setup:

1. External voltage spikes can induce a current in the wire, causing a false positive. (You are already experiencing this.)
2. If the signal wire is cut, the alarm is disabled and will not trigger. (You probably do not want this to happen.)

I would recommend redesigning the system so the trigger will OPEN a contact to trigger the alarm. This will improve the system in two ways:

1. External voltage spikes can still induce a current, but as there already is current, the system should be able to ignore/absorb the extra energy without even a burp at the control panel.
2. If the signal wire is cut, the alarm will go off as if the alarm were triggered. (This will deter the 'hostile actor' who tries to disable the system, by cutting the wire, he exposes himself when he thinks he is concealing himself.

The best method (if the sensor can be set up this way without revealing which wire is which) is to use a 4-wire cable, two twisted pairs, one pair caries the current that goes through the sensor to show the sensor's state, the other pair loop back to provide a status of the cable. Alarm line loses continuity, but cable line is good - alarm is triggered. both likes lose continuity - cable has been cut, alarm could be triggered without sounding bells by the tripped sensor, so security can move in on an unaware perp. cable line loses continuity, alarm line is good - cable is being tampered with, use silent alarm method to summon security.

My reasoning for the silent alarm if the cable is tampered with/cut is such: any moron can try a smash and grab, they're easy to catch, they don't plan ahead. A Trapster(1), on the other hand, WILL plan ahead, including plans for when things go wrong, such as an alarm being triggered. By using the silent alarm method, the Trapster is deluded into thinking that things are still going according to his plan, and having the guards move in while the Trapster believes he is following his plan is all according to *MY* plan. Check and Mate(2).

Notes:

1) A Trapster is someone who focuses on traps and alarms. They are not always thieves and burglars, although Trapster skills are quite useful to those in the thieving and burglary professions. Many Trapsters are law-abiding citizens, either Engineers, locksmiths, gamers, or even just hobyists.

2) When protecting an area with traps and alarms, it helps to be (or at least try to think line) a Chessmaster. Every trap is just a puzzle 'written' in technology, the secret is to hide 'key' pieces needed to solve the puzzle, thus forcing the Trapster to waste time trying to solve a more complex puzzle than you have to deal with, slowing him down for the authorities.

Wow! Everyone already said it all!

Of course.

A) This was a simple circuit to look at, and

B) My coffee had already kicked in.

Hello all, first, A sincere thanks to all who have given me some great ideas, the current loop situation seems like the way to go. Low impedance Vs the voltage sensing. Normally closed contacts rather than open circuit. Secondly, the system has been wired with twin shielded cable (Mic. cable) having plenty of this on hand for my P.A. work. I should have realised from my work that low impedance is critical, also I was hoping that I might have been able to use an op-amp wired for balanced line, 'somehow', must study these op-amps a lot more closely.

Again, thank you all for your input.

Rod. J.

The RFI might be coming through the cable if it is not grounded correctly.

However, RFI will also get into circuitry directly if it is not well designed. You could disconnect the wires and see if it still trips. If so, then you know the problem is susceptibility at the PC board level.

Placing the circuit in a grounded metal box will help, but RF, if the frequency is high enough, can enter into slots and holes, too.

Op-Amps (now called 'differential amps' by some) have, in theory, infinite input impedance. and infinite gain. it is not supposed to draw current on its inputs, and the output should immediately 'pin' itself to the power rail appropriate to which input has the higher voltage.

They are great in 'trigger' situations, since they 'over-react' beautifully to the trigger signal (on the non-inverting input) creeping just above the reference signal (on the inverting-input), but that sensitivity will work against you on a long wire.

I would recommend, for your current loop, a low-impedance, no-current-bleed contact for the sensor, a physical make-break switch or relay contacts if possible, and at the receiving end, an optocoupler/opto isolator to 'read' the signal. With a high-impedance 'shunt resistor' across the optocoupler, surges and spikes from line inductance shouldn't trigger the optocoupler, but normal current flow will. (Yes, I took the shunt resistor idea from the standard Christmas Tree Light setup. 'Tis the season.)

Hello adrealser. thank you for your input, wondering if you might have a circuit or a URL for this idea.? I have googled current sensing but all the circuits (and there are many) seem to be not quite what I am looking for. Hope you can help with this.

Rod

Well, to explain the circuit, I'll need to break it into three parts, I'll call them Transmitter, Transmission Line, and Receiver.

For the Transmitter, you'd have the sensor that determines the need for evacuation hooked up to a basic relay, so that when the sensor triggers, it energizes the coil. That's a pretty simple setup, basic series circuit, just make sure the relay coil is rated for the voltage the sensor uses.

For the Transmission Line, we use the Normally Closed contacts on the relay to make the current loop, the wires run from the relay back to the Receiver, where you have the DC power supply, an optocoupler, and a current-limiting resistor. Again, all in simple series. The voltage of the power supply and the size of the current-limiting resistor should be selected to provide a good 'on' current for the optocoupler without overloading it. Remember, LED's have a fixed voltage drop when running, and a maximum current they can handle without breaking down. Subtract the LED voltage drop from the voltage of the power supply and you'll have the voltage drop that will be across the resistor, divide this voltage by the max current the optocoupler can handle and you'll have the minimum size for the resistor. (Use a resistor about 10-20% larger than the formula gives to provide a safety buffer in operation.)

The Receiver end simply connects the output of the optocoupler with the PLC or whatever it is you're triggering for the alarm. The optocoupler's output is basically a control transistor, high impedance when 'off' low impedance when 'on,' and the optocoupler will be 'off' when not alarming, and 'off' when alarming ('Negative logic') so you'll need to take that into account. Think of the optocoupler's output as being an 'I'm good' signal, energized when things are fine, de-energized when there is an alarm.

Hope this helps, I'll try to break it down further if you need it.

Hello again Adreasler, thank you for the explanation. I have worked on electronics for too long, instead of thinking inside the square, I have been wandering down the street of complications !!. I don't mind admitting openly that I fail to think of the easy way of doing things, always opting for the solid state way. DOH!!!! :-(

Nothing wrong with thinking differently. That's what sites like this are for: Someone comes in with a problem, ten people come of with fifteen solutions, and then that gets pared down to the five that will fork for this particular situation. Then the safest, cheapest, or most resilient solution gets used.

I love doing things in solid state too (I've been trying to nudge the plant I work at to slowly shift all equipment that does not NEED 110+VAC to be replaced with equipment that works on 24VDC for 'safety' and ease in automation. Most PLCs/PACs run on 24VDC, even if they take a 110VAC power input.) but sometimes you just need the robustness of a physical make/break connection, even if it means dealing with 'circuit bounce' during a contact activation. (24V/110V 'Cube' style relays typically have the electromagnet and return spring strong enough to prevent circuit bounce, typically an order of magnitude or more higher than required to move the tiny, almost-zero-mass swing arm.)

Yes, you are quite correct there. Also good to realise that the devices you were originally trained on are NOT old hat but still usefull.

I started life as a telecommunications technician, we used masses of relays in those days, (1950's) and the old telephone exchanges were step by step types. Bi-motional switches of the Pre-2000 series SE-50 types and as I was leaving the job, Crossbar. The things were all packed with relays.

Thank you again for the help.

Old-school can be better than new-fangled in the right scenarios.

Anything with silicon chips is likely to simply die a horrible death if hit with a strong electromagnetic pulse (EMP), with the denser ('more powerful'/'newer') chips being more vulnerable They're even in danger when powered down. Vacuum tubes, on the other hand, completely ignore EMPs when off, and may not do more than 'hiccup' if hit while running, like an AM radio giving a moment of static when passing under power lines, but fine afterwards.

Hello anonamous Hero strangely enough I tried that angle. I fed the wires in via a feed through capacitor, it seems quite likely that the problem is via the wiring. I will google for a run down on the current sensing tomorrow. Although it is only 2129 Hours as I type, I battled a couple of large fires yesterday and at present, BED is looking really good.

Rodney J.

Sleep well!

Thank you, I did......luckily, at it again today.