Motion Failure Alarm

Hi,
I'm trying to build a motion failure alarm to go inside an already complete electro-mechanical system.

Basically there's a piece of metal moving back and forth (linearly at about 70rpm) and I want to sound an alarm if it stops.

I'm working on this idea (see attached). It simply uses an RC network to control some analogue switches.

I imagine the sensor to be a photo-electric one where the output goes high when it senses a presence. This trips SW2 and resets the RC network and starts the cap re-energizing. If it doesn't get this by the time SW1 reaches it's switching voltage, then it trips and sounds the alarm. If it gets a subsequent pulse, it will re-energize and start again.

The supply doesn't have to be 12V, I'm using it currently to fit with a possible sounder, but it'll need to be transformed from mains, so it could be anything really.

Firstly: is this a good system to use? (and will it work)
Secondly, the analogue switches (i'm imagining using a 4066 IC). When do they switch to closed with a rising voltage at the control pin?
Thridly, is this likely to be reliable - ideally, i'd like to leave it running for a few years, non-stop (!!)

Now i haven't finalised R1 and C1, i need to fine tune the timings. I'd like it to start tripping at just less than 60rpm (1hz). Any suggestions, and which kind of cap should i use - i can't get my head around them.

Thanks in advance and I hope you're all well

xephias

However you decide to indicate the motion, the circuit you need is called a "missing pulse detector" - http://ecelab.com/circuit-miss-pulse-det.htm

I do not believe the 4066 is the best choice here. As the difference between 70 ppm & dropout of 60 ppm is relativley small may need a well regulated V supply, some V comparators , pulse shaping network, a few caps & resistors, & a schematic. I tuned in late & must leave now - hot date.

Here is one way: A 555 was used to simulate your photo detector, outputting a 1/2 ms pulse at a rep rate of 70Hz & 60Hz, by shorting R3-33k +-. The pulse stream is integrated by R6 & C2, giving about 3V with ripple @ 70Hz.The peaks of the ripple @ 60 Hz are detected by LM 393 & is filtered by R10 & C3. R5 is adjusted to light LED, representing the alarm, when frequency drops to desired set point,60Hz. R11 helps speed up the response time.

Just re read origional post where it's 60 & 70 RPM== 1s & .857s,change C2 to 5uF & C3 to 50 uF ??

It seems to work best if C3 is left at 1 uF, R 11 @ 1M= 3 s response, 750k = 2s., much lower & have pulses in output.

Hi, sorry - i've been away, back now tho - and that helps, thankyou

If it's a relatively steady 70 Hz movement (t~=.0143 seonds) this could be the ideal situation in which a 555 could be hooked up as the commonly used but often forgotten "mising pulse detector" circuit. First time that piece of moving metal doesn't make its normal cycle the circuit triggers and pin 3 of the 555 changes state - setting off any type of alarm you desire and/or stopping the line completely.

Our missing pulse isn't missing, just delayed by .143 sec. Not knowing what input signal looks like, added a schmitt trigger fpr pulse shaping, followed by sugestions by beenthere & marshallf3 for a missing pulse detector. E- a 555 one shot, is timed for 1 sec. ie; 60 RPM, or as desired. The positive going polse from C-6, causes D to dump the charge on C3 in about 250 μsec.setting E high, still or again. If input falls below 1 sec. E has time to reset, output going neg. for a few μsec to several msec. The resolution seems to be better than +_ 100 μsec. Back to B-4, neg pulse at input rate continusly triggers E. Powersupply used on bread-board was 9V C-Zn, suggest well gegulated supply & filtered. Output E-3, should be fed to another 555 one shot to make alarm long enough to hear. Spare invertor inputs should be grounded. Missing one resistor, 68Ω, from D collector to C3 cap.to limit peak discharge current. How many others??