Cheap Sensor Function Checker for 120 RPM with 24V Output

AI Thread Summary
A user seeks a cost-effective device to check a sensor output that operates at 120 RPM and delivers a 24V peak pulse train. The discussion centers around creating a circuit that can monitor the sensor's performance without risking damage to expensive equipment. Key considerations include tolerances for pulse amplitude and frequency, with a preference for a battery-powered solution that lights an LED if pulses are missed. Suggestions include using a timer and a counter IC to trigger an alarm when the pulse count is below a set threshold. The conversation also explores the possibility of drawing power from the sensor's output for the monitoring device.
wolram
Gold Member
Dearly Missed
Messages
4,410
Reaction score
555
I am looking for a cheap device to check sensor function, the sensor is detecting at 120
rpm and should give a 24v output.
If there is a circuit i could build my self that would be great.

Thanks.
 
Engineering news on Phys.org
Are you saying that the output is a 120*60 = 7200Hz sine wave at 24Vrms? Or does it put out a pulse train of 24V peak pulses at 7200Hz? Do you have a DVM with a simple frequency counter in it (my Fluke handheld has a low-frequency counter capability)?
 
berkeman said:
Are you saying that the output is a 120*60 = 7200Hz sine wave at 24Vrms? Or does it put out a pulse train of 24V peak pulses at 7200Hz? Do you have a DVM with a simple frequency counter in it (my Fluke handheld has a low-frequency counter capability)?

Hi berkerman, it is a train of 24v peak pulses, i do not want to leave an expensive meter coupled to this circuit for fear of damage, environmental and human, i need to look at this sensor to rule it out as the cause for a glitch a glich that has gone by the time a meter is set up.
 
Okay, got it. You want to leave the detecting circuit in-place, not just use it for spot checks.

What would be considered not working? What are the tolerances on an okay pulse train, in terms of amplitude and frequency? At what point do you want the detecting circuit to call the output no-good? Are you planning on battery powering this detector thing, or do you have wall power available? Or do you want it self-powered on the 24V pulse train, and it just keeps an LED lit if the sensor is putting out a pulse train within tolerances?
 
As long as the sensor is puting out a 18 to 24v signal and not missing any the sensor would be deemed okay, a battery powered device would be ideal.
 
How about a simple audio output? You could cut off the bottom parts of the pulses and drive a small speaker with the 7.2kHz audio pulses. The ear might be able to hear if there are any consistently missing pulses or whatever the various failure mechanisms are. Don't know if having a continuous audio output would be too obnoxious, though. Are you looking more for an LED indicator?

Also, how many pulses have to be missing before you light a trouble light? Or is it the frequency changing that you want to alarm? This is all do-able, I'm just trying to get an understanding of how complex it will need to be.
 
I think an LED would be best, say it lights if any pulse is missed, this semsor forms a sort of (and) circuit, the logic and hard wear have been in use for years but now it will have a train of (misses) one day and be okay for may be a week.
 
Okay, then the easiest way would be to use a timer to time the period between pulses, and if that timer goes above some set count, it trips the circuit that lights the LED.

So the first part of the circuit would decrease the amplitude of the signal down to logic levels, and make a trigger signal from each pulse which was 18V to 24V in amplitude. That would be fed into the reset input of a free-running counter, so that the counter never reached the alarm count (which it would if a single pulse were missing, or some number of pulses in a row, depending on how you want to configure it). The alarm count would trigger a flip-flop, which would light the LED. The whole circuit can be reset with a single pushbutton, to reset the counter and turn off the LED.

Does that give you enough to build it, or do you want some more details on how you can make it? I'd start with a 74HC4060 counter IC, with maybe a 5MHz crystal on it, to give you a nominal count of about 5MHz/7.2kHz = 695, and you could use the 1024 count output to trigger the alarm LED flip-flop. Divide down your input pulses with a resistor divider, and run that into a comparator circuit that looks for the 18V minimum equivalent voltage level. The comparator output would feed the reset of the 74HC4060. The whole thing should be able to run off of a 9V battery for a fairly long time (use a low-current LED...).
 
I would much appreciate some detail i am not at a level where i can design a circuit,
building will not be a problem though.
Thanks berkerman.
 
  • #10
wolram said:
I would much appreciate some detail i am not at a level where i can design a circuit,
building will not be a problem though.
Thanks berkerman.

I'll sketch something when I get a chance, wolram. It might be the weekend though. I'll also think about simpler circuits that might do the job -- less accurate, and might require turning a knob to tune the sensing frequency or something. Any chance we can draw power from the pulse train? We wouldn't need much. What's the output impedance of the sensor?
 
  • #11
berkeman said:
I'll sketch something when I get a chance, wolram. It might be the weekend though. I'll also think about simpler circuits that might do the job -- less accurate, and might require turning a knob to tune the sensing frequency or something. Any chance we can draw power from the pulse train? We wouldn't need much. What's the output impedance of the sensor?

The sensor can drive a load, from memory it will power a 1 watt solonoid coil.
Thanks for the help berkerman.
 
Back
Top