Sharp Prox IR sensors read shorted

[adciv]

An FYI to anyone who decides to use Sharp IR Proximity Sensors such as this.

We checked the isolation of our Electrical System yesterday (IAW R37) and found we had a short between our frame our electrical system. After some frantic searching, we found the issue was due to the Sharp Proximity Sensors we are using. With much hair pulling, we discovered the housing is made of a "conductive resin". We verified this using a multimeter with significant pressue to the case (similar to what it would see when screwed down). Our solution was to use nylon washers/standoffs and nylon screws to secure the sensors instead of standard metal screws.

I hope I can save some people some panic if anyone is using these.

[Al Skierkiewicz]

adciv,
Does yours have the circuit board exposed on the back as shown in the picture? If so, I would guess that is the source of the short. A simple lexan sheet behind the sensor would fix that.

[RufflesRidge]

Originally Posted by Al Skierkiewicz

adciv,
Does yours have the circuit board exposed on the back as shown in the picture? If so, I would guess that is the source of the short. A simple lexan sheet behind the sensor would fix that.

Al, the case itself is conductive and will cause a short to ground reading if not completely isolated.

Adciv, you may also find that these sensors are VERY sensitive to supply noise. A small cap across 5v and GND right at the sensor can make a world of difference.

[Gdeaver]

Yes the plastic housing is conductive and the entire sensor must be isolated from the frame.

[Al Skierkiewicz]

Is the circuit board designed to contact the case for this purpose?

[adciv]

A bit more info: We were reading a ~30 ohm impedance between the robot frame and ground, not a hard short. There is a metal screw in the circuit board holding it to the frame. I suspect that is the circuit path. If I hadn't just thrown away one we fried, I'd dissassemble it for examination. Across the full width of the case using a multimeter and sufficient pressure, we'd measure ~100 ohm impedance.

RufflesRidge, thank you. Is there a capacitance value you recomend? We'd noticed the bouncing and have been using software filtering but I'd like to try this as well.

[RufflesRidge]

Originally Posted by adciv

RufflesRidge, thank you. Is there a capacitance value you recomend? We'd noticed the bouncing and have been using software filtering but I'd like to try this as well.

It's been a while but the notes I can find from the last time I used one suggest we settled on a .1uF in parallel with a 10uF both as close to the sensor as possible.