360 degree non-contact rotational indicator 7

[whistlemeup]

While there are many rotational function generators that give 360 ppr, they rely on direct coupling to a rotating shaft. What I am looking for is a way to place 360 marks, painted bars, reflective strips, etc. on one section of a flat 12 inch diameter coupling, and using something like a bar code reader to generate 360 pulses directly for a TTL logic device. This is for a 1200 rpm 200 hp electric motor driven compressor that has no free shaft for direct coupling. A 1/rev pulse is not a good option as we need to stay syncronized to the crank for research purposes. I have searched this site and many others without finding a source. Any ideas would be appreciated.

 

[VEBill]

I can't help you with suggested hardware, but it's common for such optical rotary encoders to have both ACP and ARP outputs. The Azimuth Clock Pulse would correspond to your 360 pulses per revolution, and the Azimuth Reference Pulse would be another pulse to mark the reference position ('top dead center' sort of thing).

There are also absolute encoders that use various multi-bit codes (such as Gray Codes) that define the position directly.

If you can only find the sensor, then the rest could be done in fairly straightforward software.

Think about latency. "It's exactly here some time ago."

 

[VEBill]

Speaking of "top dead center", I think that many modern cars have 'crank position sensors' that address the exact same problem. That is, knowing exactly the exact angle of the shaft at any point in time and at a very high rate.

 

[MacGyverS2000]

Why not etch the shaft? It's about as permanent as you can get unless the shaft shreds. Two rows of markings offset by a mark's width gives you phase information for direction. A third row with a single mark will give you absolute position from last full-rotation pulse.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[ScottyUK]

It is quite common on rotating plant to fit a toothed wheel, or perhaps use an existing gear, with a proximity detector. This method has better immumity to dirt than a surface etching, but whether that matters depends on whether you have a dirty or clean process.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[MikeHalloran]

TI used to sell a 'retroreflective' optosensor that was sort of folded up, and intended to be used with black/white stripes a few mm away from its pointy end.

I used one to sense liquid level with a 'Delco Eye' type extension. It worked fine. That was a long time ago...



Mike Halloran
Pembroke Pines, FL, USA

 

[roydm]

The toothed wheel Scotty mentions would be a good way, use a second sensor to give you the zero point, once per rotation.
Another option would be to use a Grey coded strip with several sensors, you could print it on paper and laminate it.
Let us know what you end up with.
Roy

 

[Compositepro]

Red Lion sell kits for motor speed measurement that will mount on standard motor faces like "c" face. It will fit between a motor and gear box without modification. It consists of a mag pick-up and a gear.

http://www.redlion.net/Products/DigitalandAnalog/Sensors/MotorsGears.html

I see a 120 tooth gear at 5" diameter. You will probably have to find another supplier for a 360 tooth gear or sprocket. But it seems you have plenty of room. Their sensors and meters will work with any gear that has sufficiently large teeth.

There are meters that count pulses and there are others that calculate speed from period between pulses.

 

[sreid]

I'd suggest using Omron Photoelectric [reflective] sensors [eight sensors] and paint a black and white eight bit Gray Code around the coupling. You then always have absolute position.

http://www.ia.omron.com/product/18.html

 

[ScottyUK]

"You then always have absolute position."

Until it gets dirty!

Has the OP abandoned us? Some more information about the application would be helpful in identifying the best solution.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[CharlieGill]

The suggestion to use a sprocket gear and reluctance pickup is probably the best. You should not need 360 teeth. I cannot think of any reasonable application (particularly on a compressor) that needs an absolute position reference at each degree. I do not know what hardware you are feeding this to but surely it can interpolate whatever pulse/degree information it gets and extrapolate position information until the next pulse. Essentially a software phase locked loop. For the zero reference you can weld a small stud to the side of one of the teeth and add a second proximity pickup to detect just the stud. Another option is to grind off one tooth, the software PLL can easily detect the missing pulse to sync pulse train to an absolute reference.

Charlie

 

[whistlemeup]

Wow! I got a lot of good suggestions. The use of gears, mag pickups, and proximitors,.... I have used all of them over the years. Staying in sinc with the rotation of the compressor crankangle is important as we are generating PV cards for research. We have a two throw motor driven compressor, one compressor active and the second deactivated. Analysis shows rod loads will allow this but we may have some non-linear rotation. Also, to apply the gear, you would need access to a shaft (unless using a split gear). No shaft is available. Only the flat surface on the coupling. I was thinking along the lines of a pre-printed tape with an optical readout. Possibly a stamped tape with a proximitor. I still have not brought the correct hardware together.

 

[BobM3]

Looks like a tough environment. Optical may be problematic. I'd favor magnetic.

 

[VEBill]

For a short term test you should be able to keep it clean enough for optical to work. Maybe obtain several extra tapes for backup.

Another "optical" option might be a high speed vision system programmed to read the code you've printed on the circumferential tape.

 

[ext123]

We looked into using Renshaw product. RESR Rings. You may be able to press one onto coupling. Our project fell through. Good luck.

http://www.renishaw.com/en/6453.aspx

 

[ScottyUK]

Put the coupling onto a dividing head. Using a slot drill machine a series of small - 1/8" or so - equi-distant indentations around the periphery to a depth of about 1/8". A prox should be able to detect the passage of the indentations as the coupling rotates. You won't remove enough metal to significantly affect the integrity of the coupling.


----------------------------------

If we learn from our mistakes I'm getting a great education!