Best way to sense a moving noodle?

[nomorebots]

I am working on a project that needs to sense if a noodle is being pulled through or not. Currently I am thinking about getting a ATTiny85 board with a rotary encoder like this (http://www.hobbytronics.co.uk/image/cache/data/sparkfun/com_09117_rotary_encoder-500x500.jpg) that the shaft is butted up to the noodle. My idea is to push the noodle against the shaft. When the noodle is being pulled away, it will turn the shaft. I will just see if the shaft is turning or not to see if the noodle is moving or not.

A couple of questions hoping to get answered:

Is the encoder 'smooth' like those you can find on amplifiers? In other words, no clicking. Best if there is no damping so the encoder can be turned easily on contact with the noodle. What key word should I be looking for?
are there any other good and cheap way to do the same job? Best if the package can stay extremely small. I don't need fine resolution and a rotary encoder has everything (encoder, pin-out, shaft that the noodle can roll on)
thought about using those old mouse optical encoders. But they are kind of big (smallese wheel diameter is about 22mm). Would any one recommend to use the optical mouse sensor instead?

[Rerouter]

the clicky style is an option called detent, without it most are smooth scrolling and low torque between clicks,

do you in any way need to know the rate or amount of noodle that has come through, or just that it has moved, from your image, it seems your noodle is more like steel wire, if such is the case, then i agree i cannot think of much simpler than an encoder with a rubber wheel or similar...

[jimon]

Couple of days ago I've fixed my mouse and in mean-time checked out a wheel assembly : http://www.ebay.com/itm/Original-Logitech-G700-G500-M705-MX1100-Mouse-pulley-scroll-Wheel-MOUSEWHEEL-/281178770516?pt=LH_DefaultDomain_0&hash=item41778f8c54 It's a nice wheel with two modes - clicking and free running. It doesn't have proper bearing so lifetime is limited to couple years  But a good thing is a rubber on a wheel - it should be able to grab the noodle.

Maybe there are similar free running optical encoders with a rubber band ?

[Rerouter]

yes should have stressed that, optical or magnetic encoder is a must, 10 million steps may sound like a lot, but when you do the math, in how you will be using it, its about what you should aim for

[Kalvin]

How about using optical mouse? I mean, not the mechanical roller, but the optical sensor use to track the mouse movement. Should be accurate enough. Best feature: no mechanical contact required.

[Kalvin]

Another simple solution - with a mechanical contact - is to use a "pulley" or a wheel. The noodle will turn the wheel. The wheel has a magnet on it, and a magnet sensor will detect whether the wheel is moving. You may also get an accurate measurement of the noodle length.

[SL4P]

Continuing the theme of a wheel - consider pulling the noodle over the relatively small diameter shaft will probably slip - giving variable +/- results.

Increasing the diameter of the rolling part (wheel) will reduce the rolling resistance, and carefully designing the noodle path (using gravity or tensioning idler/roller)  - can also increase the friction / contact area of the noodle & wheel.

[AndyC_772]

Noodles are food, and the fewer mechanical parts they come into contact with, the less the chance of contamination.

Dismantle a mouse, mount the optical scanner where it focuses on the noodle, and write an application to record the amount of pointer movement. It's cheap, uses readily available consumer technology, and it's non-contact so there's nothing to wear out or contaminate the noodle with crud.

[speleodox]

I think using an optical or laser mouse is the way to go, but if that doesn't work another option, probably more expensive, is to use an ultrasound transducer. There are consumer fetal heartbeat monitors that I am pretty sure use doppler for around $40 all up. The hard part would be finding the right resin to encase it in so there isn't too much impedance mismatch, but that is definitely a solvable problem. The ones we use at work are just 2 piezo films encased in epoxy. The noodle would have to be coupled to the resin, embedded in the wall of the noodle channel, by some liquid, presumably water. I think all you need after that is filters, amps, and the signal processing to pick out the doppler shift.

[Teledog]

How about a series of photo interrupters? Cheap & effective
They're often used for flow measurement -different designs than the usual "off the shelf" units, mind you.

[babysitter]

Get a heat source like a halogen light and run it intermittent, so you get a hot-cold-hot pattern on the noodle. Use a PIR Sensor, Thermopile, Bolometer or similar downstream (downnoodly) to detect the hot-cold-hot pattern.

[SL4P]

Get a heat source like a halogen light and run it intermittent, so you get a hot-cold-hot pattern on the noodle. Use a PIR Sensor, Thermopile, Bolometer or similar downstream (downnoodly) to detect the hot-cold-hot pattern.

Clever. Only weakness if the feed stops, you have to wait for the restarted pattern reach the sensor.  Nice.

[Richard Crowley]

Is this "noodle" as in human-edible pasta?  Dried/hard or new/wet/cooked?
Even if contact is allowed with a hard piece of pasta, those little rotary encoders need MUCH more torque to turn than than you can expect from the circumference of the shaft of the encoder.
Even with a larger wheel and "pinch-roller" (as in old tape recorders) there wouldn't be the expectation of much reliability.
And if the pasta is wet/limp, that solution seems completely unworkable.

Also not clear if it is discrete pieces/lengths of noodle that need measuring, or a continuous "strand"?
It would be easy enough to count and even measure discrete lengths. 
But measuring a continuous strand seems much more problematic.

As others have already noted, none of those "user-interface", panel-mount rotary encoders are compliant enough for your application.
Probably the most "compliant" kind of rotary encoder is the optical-sensed type because there is no friction in the sensor (only in the rotating part).

If you can devise some sort of rollers/pulleys, etc. that the noodle can reliably turn, then you can simply apply a black/white pattern to the rotating part and use an optical sensor to detect when it is turning. With a pair of sensors, you could detect the direction of travel, and with the proper diameter and pattern, you could actually measure length/distance.

[SL4P]

... those little rotary encoders need MUCH more torque to turn than than you can expect from the circumference of the shaft of the encoder.
Even with a larger wheel and "pinch-roller" (as in old tape recorders) there wouldn't be the expectation of much reliability....

Richard - normally I agree with your posts but 'proper' rotary encoders - not the 'user-dial, knob and push' type - are extremely easy to turn... possibly single digit grams of torque without the help of a pulley. The wheel and idler was suggested to increase the traction footprint - less slippage.
(Think of the free turning optical encoders inside printer carriages etc.   Very lightweight to rotate.)

Cheers

[Richard Crowley]

Yeah, I was refering to those inexpensive user-control things like the OP was suggesting.
Certainly the more "industrial" encoders take much less torque.
But trying to sense the movement of a wet, flaccid noodle seems like asking a lot even from the very low-torque industrial encoders.

But there are a whole lot of other issues around the problem that are not revealed here.
Like what is causing the noodle to move?  Is it being extruded, is it being pulled? is it falling from gravity?

[IanB]

I have assumed all along that "noodle" was a technical term for some kind of positioning rod. If it is literally a food noodle, like a piece of spaghetti, then...what the heck? 

[babysitter]

[Patented super method]

Clever. Only weakness if the feed stops, you have to wait for the restarted pattern reach the sensor.  Nice.

Its not a weakness - stopping feed is the condition which is to be detected, as there is no more AC from the sensor 

The real weakness was that I didn't talk about stopping the heat source if the noodle stops to make a too hot spot and as a result SWMBO complaining to hubby because he serves her pre-burned noodles on that one day per week when he cooks.