I need to setup a switching device that gets its input from a high speed 3000 - 5000rpm rotating disc. Rotor disc is about 1.5 feet in diameter. Readings are to be made on the outer perimeter.

My fear is mechanical commutator style switch is to slow, and not robust enough, there may be issues with vibration on the rotor.
The rotor has large permanent magnets on it for a generator, so hall effect sensing is out of the question, for there would be major magnetic interference.
So the next thing I thought was opotocoupling...having a light beam directed at the rotor and putting some reflective strips on the rotor disc to trigger switching. Ideally I would want the signal to travel optically for a foot, out of the high magnetic area where the generator action is taking place for fear of interference.

I have been looking at optocouplers online, but having some difficulty making heads or tails out of whats available given what I want to do. Is what I want perhaps, more like a laser detector or optical scanning? Perhaps there are some other creative solutions out there?

 

That is a reflective optical sensor.

 

... so hall effect sensing is out of the question...

Well maybe, but have you looked at the signal coming off? I mean, you might just need to divide by some constant, or maybe set a threshold, or both, and you'll have your result.

 

Good idea wayneh. Don't avoid the magnetic field, measure it!

The idea that there are large magnets, so hall effect sensing is out,
kind of gives me a headache. That's like saying the light blinks on and off so we have to listen for the relay clicking.

 

Well maybe, but have you looked at the signal coming off? I mean, you might just need to divide by some constant, or maybe set a threshold, or both, and you'll have your result.

Hmmmm, that may work. The problem with dividing by some constant is the speed of the rotor will be changing constantly, so timing would be off. With a reflective optical sensor, I have manual control of how much reflective material is on the disc, and "on time" of the sensor will auto-adjust with speed.

Threshold might work, how would I configure that?

 

All I meant is that you might be getting more than one pulse per revolution, so you might need to divide by 8 or whatever to get back to rpm. That ratio won't change.

Setting a threshold to separate signal from noise can be accomplished in circuitry or in software. The latter is easier if you're going into a controller or such anyway.

If you could put an oscilloscope on the output of a hall sensor, it may become obvious what you would need to do to use the signal. Or maybe it would be obvious that you can't use it.

 

Does the number of magnets on your disk change with the speed? You just need to divide by the number of magnets that trip the hall sensor.

 

Placing a smaller or weaker magnet 'behind' the sensor in a configuration that opposes the pole to which it is reactive, reduces the time during which the field is strong enough to create an output in the hall effect device. With a steady field strength from the magnets in the generator a small 'tuning' magnet could be used on the opposite side of the sensor and by making it move closer or farther from the back of the sensor, you control the duration of your output pulses.

 

By the way, the purpose of the sensor is not to measure rpm, but to send a precise control signal for some timing switches in the system. Ultimately I hope to have about 8 sensors to track and switch with precise timing control switches on the output of each generator coil.

Each sensor will send a signal 8x per revolution. I need it to be precise though in terms of alignment with the magnets. Hopefully I can set the optical sensor so that it turns on and off, with mm read precision on the rotating disc.

 

If you were to shine a laser at a reflective label, with a sensor placed where the light would bounce off to, you should get a pretty sharp indication of the label's presence. The laser can be a cheap little thing like a battery-powered pointing device.

 

Instead of reflective strip, could you have a notch or a transparent spot on the disc and use an optical gate? Or a small opaque projection on the disc that would trigger an optical gate? (The former would signal on at that location, and the latter would signal off at that location)

IME, optical gates tend to work slightly better than reflective optical sensors, since reflections often bounce a little bit funnier than you might expect and tend to give noisier reading. Since you said you really wanted high accuracy, this might be something you want to think about.

 

Don't forget that a 1 gram weight rotating at 5000rpm on the edge of a 1.5 ft disc experiences a centripetal force equivalent to 6.4 kg. Balance matters.

But I agree that an optical solution is probably better than a magnetic one, if you need mm precision. That precision could be tough even with optics. Things like temperature changes, loss of alignment, etc. will be headache.

 

If you were to shine a laser at a reflective label, with a sensor placed where the light would bounce off to, you should get a pretty sharp indication of the label's presence. The laser can be a cheap little thing like a battery-powered pointing device.

Yes this is exactly the kind of thing I was imagining might give me the best accuracy. Instead of a battery-powered pointing device, any idea where I might get a more permanent high precision laser (very small diameter laser point <1mm) and reflective material, plus light sensor? A kit would be ideal....

Instead of reflective strip, could you have a notch or a transparent spot on the disc and use an optical gate? Or a small opaque projection on the disc that would trigger an optical gate? (The former would signal on at that location, and the latter would signal off at that location)

IME, optical gates tend to work slightly better than reflective optical sensors, since reflections often bounce a little bit funnier than you might expect and tend to give noisier reading. Since you said you really wanted high accuracy, this might be something you want to think about.

I agree with what you said about noisy readings from a reflective optical sensor hence why I am more interested in a pinpoint laser style approach. By optical gate, do you mean having a hole in the rotor, whereby a light shines through to a sensor on the other side? With the optical gating you speak of, how easy would it be for me to adjust the on/off time? Do I have to start drilling/filling adjusting the size of the window/hole?

Don't forget that a 1 gram weight rotating at 5000rpm on the edge of a 1.5 ft disc experiences a centripetal force equivalent to 6.4 kg. Balance matters.

But I agree that an optical solution is probably better than a magnetic one, if you need mm precision. That precision could be tough even with optics. Things like temperature changes, loss of alignment, etc. will be headache.

Yes I am more inclined to optics, however I am wondering if I did use a hall effect, could I perhaps adjust offset (on/off time) to mimic precise millimeter like timing via a micro-controller/chip or a simple pot?

Thanks for the help so far guys. I really appreciate the creative ideas.

 

Is there a reason that the measurement must be taken from the outside of the rotor?

Do you have access to the shaft the rotor is mounted to? If so you could make a smaller diameter disk to mount to the shaft and then a optical or hall sensor could be used with it. This would also create less of an unbalance at speed, being a smaller diameter.

 

Yes this is exactly the kind of thing I was imagining might give me the best accuracy. Instead of a battery-powered pointing device, any idea where I might get a more permanent high precision laser (very small diameter laser point <1mm) and reflective material, plus light sensor? A kit would be ideal....

Maybe this would be adaptable. http://www.ramseyelectronics.com/cgi-bin/commerce.exe?preadd=action&key=LTS1

 

I got myself the Optek - OPB740WZ Reflective Optical Sensor and am new to reading these data sheets. What is the min/max/ideal voltage to power the LED? And what is the min/max/ideal voltage for the transistor?

Link to Data Sheet

If I am reading it correct....max on the LED is 2V with an ideal at 1.8V...and the transistor is anywhere between 5V and 30V?

 

Common mistake. You don't feed LEDs voltage, you feed them current. You have to use a resistor between the power supply to limit the current to 20ma after the LED uses up 2 of your volts.

No, I think it's 40ma.

The 5 volts is a reverse breakdown. You aren't going to use that.
The 30 volts is the high limit. It's guaranteed to survive anything under 30 volts.
Don't push it. Make your power voltage considerably less than 30 volts, like 20 volts or 12 volts.
It will even work down to 5 volts if you're running this to a microcontroller.

 

There's no need to run the sensor on anything but 5V. Put the LED in series with a 220 ohm resistor to start with. And connect the emitter of the phototransistor to Gnd, and the collector to +5 via 10K ohms (maybe 3.3K if you have the Darlington version). See how it works and change the resistors if you need to. But be warned that units like this are very sensitive to ambient light, especially daylight or incandescent--fluorescent, not so much. You'll think that the sensor is never seeing a dark stripe (or it's always seeing a reflective stripe) when actually what's happening is it's always seeing the desk lamp.

 

The next thing I want to do is control a IRFP3710 MOSFET with the OPB740WZ Optical Reflector.

Is this how I would wire it up? I want the optical reflector controlling the MOSFET like a switch.

 

The gate is like a capacitor and will hold a charge (and voltage) unless you allow for a drain-off path. A resistor to source will do the job. You'll need the lowest value you can tolerate (to reduce turn-off time) that will still allow the gate to come fully on when the transistor comes on.