Ping sensor

scotta

Does anyone have first hand experience with the ping sensor ?

I would like to know how accurate the measurements are from
1/2 inch to 1 foot.

I am trying to determine the diameter of a roll of paper as it
winds up.

Scott

SRLM

No way the ping is good enough for that. First off, a paper towel is soft, so it absorbs sound. Second, you'd have to do some compensation for the temperature. See the top of page four in the Ping))) documentation. Plus, the size of the roll increases very gradually, so you will have measuments that for one reason or another fluctuate around the actual measurement. The ping might be able to determine if the roll is more on the full side or more on the empty side, but it'd cannot do what your are looking for.

A possible solution would be to use several IR devices. Make a row or IR LEDS with some recievers between them, then have the line follow the radius of the paper towel out there. Since the towels are white, they'll reflect the light fine. Then just use them to detect the presence or absence of an object, and you'll get how full the paper towel dispenser is.

The most accurate solution would probably be a mechanical arm that rubs along the roll has has a potentiometer attached.

Dgswaner

how about putting an encoder on the roll with the roll of paper, that way you could keep track of ever mm. or a less accurate but easier method could be to have a spring loaded arm attached to a Pot, that hits against the roll. the different values of the pot would indicate the amount of paper that is there..... or not there.

I'd agree the Ping not the right tool for this job.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"A complex design is the sign of an inferior designer." - Jamie Hyneman, Myth Buster

DGSwaner

scotta

To both of you: Thanks for your input.

SRLM: The machine is designed for a range of products, from damp
toilet paper to stainless steel ship stock. It is being designed this way
because the company that is buying them specializes in developing
product brands for other companies, and the product is usually based
on a roll to roll process.

The idea of a mechanical dancer arm has been eliminated by their
development engineers. They are specifying a non-contact sensor
to eliminate tension on the web. I thin your correct that the ping
would be a little noisy. We just have to try every idea so we know
we have the correct solution.

Dgswaner: The product line is a continuous, system. They have
splicers in-line so the process never stops. The product rolls may,
but not always, vary in length. I wish I could slap an encoder on
the rollers and be done with it !

I suppose we are looking at an optical solution after all. There are
off the shelf solutions that would work for this application, but we
are always trying to develop new products ourselves.


Thanks,

Scott

SRLM

The encoder seems to be the best non-contact idea so far. It sounds like you are using this in the manufacturing part of the production: to know when to stop putting on more material. The encoder could be permanently mounted on the loading arm. This way you could get an exact length per roll. You'll have to find the formula to convert a spiral distance to linear, but I'm sure there is one out there somewhere. So why can't you use encoders?

agodwin

Shine a light (probably a laser) from one edge of the roll to the other, with a slight tilt so that when the roll is empty the spot will hit the roll far from the source, and when it's full it will be near the source. Then focus an image of the roll's width on a photosensor (you could use a camera, but a ccd line sensor would do and is a lot less data to process). You'll get a bright spot on the sensor that corresponds to how full the roll is.

This system is used commercially to measure the ride height of motor racing cars.

ElectricAye

Scott,

You might consider putting a sensor on a traveling arm. The sensor/arm/motor system is programmed to always keep the sensor within a certain distance of the toilet paper. As the paper moves up or down, so does the traveling sensor arm. An encoder keeps track of how many revolutions the stepper motor has made and in what direction, and these encoder counts are translated (no pun intended) into a distance (position) measurement by your computerized processor. I've attached a sketch to illustrate the concept.



Hope that helps,
Mark

SRLM

The real question here is "How do you flip the screen?" I thought it was Ctrl+Alt+[noparse][[/noparse]arrow key], but that didn't work. BTW, in case you didn't look, the previous image is upside down. I've flipped it and reattached it.

Anyway, I think the general idea is a good one. But why not have the encoder look dirrectly at the roll? This way it could be hidden in back and not get in the way of whatever input/output there is. The idea of a touchlike non-contact sensor seems to be the easiest here so far...

ElectricAye

Praise be to SRLM for flipping my image. If you want to make the system simpler, you could eliminate the encoder and keep track of pulses sent to the stepper motor, but then you'd have to add a homing switch for the stepper system to zero itself upon. Losing the encoder might make the system a little less precise, maybe, but might be worth the savings in money and complexity.
As for sensors: consider optical or, if you like player pianos, consider pneumatics.

Placing an encoder directly on the uptake roller is probably too indirect a method if you are processing many different kinds of materials. Even a change in humidity might change the thickness of a sheet material, so your spiral formula might be adversely affected.

Mark

SRLM

Question about manufacturing in general: are mills that process metals temperature controlled? Seems a customer might get less material in the summer that the winter, since the metal expands with heat. Of course, this only works if it's measured by volume, not weight. Just a thought about the rollers.

ElectricAye brings up a good point though: for some things you'll want to measure by weight (sheet metal, wire namely), others it length on the roll (Paper towels, toilet paper, tape, carpet, etc.) So it seems that you'll really need to come up with two systems, and choose the right one based on the material. None of the ideas here would work so far since they are all temperature/humidity dependent. What really needs to be measured is length or weight. The weight is simple enough conceptually, as is the length. If your rolls are on an arm, you could rest the arms on springs, and have a linear pot in an RC circuit. For the length, you could possibly read the material itself. For example, if the item is postage stamps, you could put a IR device that reads between the white of the edges and the colors of the picture. You could then deduce the length of the roll by the pass by count.

TJHJ

Why not compensate it? There are a ton of temperature sensors and or humidity sensors that can be easily interfaced to the prop. And knowing the thermal coefficient of expansion of the metal it would not be too bad to calculate it out, what the density change of the material is. If you really wanted to be accurate, you could read the temperature of the metal coming off the production line using an ultrasonic or similar surface temperature measurement device. ·
·
Just a thought here, but when I buy any roll based material; its spec is in length only, and material. To be concerned about the change related to temperature doesn’t make much sense because what happens when the product it purchased in Texas, versus Alaska. The resulting temperature change and expansion will continue after manufacturing. Also most manufacturing process result in a stable density output of the material in the method it is created. Look per example at a drawn sheet metal rolls. The drawing process creates heat, but the machines tend to be in a climate controlled warehouse. So the start temperature is the same always and the process of drawing the metal creates a certain amount of heat depending on how much of a change the metal is undergoing. ·Now the metal travels to the rolling/packaging process a given distance, at which it is exposed to the same climate controlled area it started in, so the cooling is constant, it may not have cooled all the way down, but it will have cooled the same amount as it travels. So knowing these relationships you could figure how much cooling is left and resulting contraction of the material, so that after the heat has dissipated from the manufacturing process it would be the same constant length.·
Last thought, You could measure the weight of the start material entering the manufacturing process and relate that to distance for density. I am not really sure how this would work for paper, or textiles, I am a metal person. For paper·drying·measure the amount of water being pumped into and drained off using an impeller system. Then you would know how much water is in the paper, I think.

For the main question, I get measurements in centimeters, I have never tried to split it again, but it would all depend on your counting system, using 32bit I think you could get even more accurate.· In the range of 10x or more.

Post Edited (TJHJ) : 8/17/2008 5:42:25 PM GMT

ElectricAye

TJHJ,
to compensate for environmentals, one would need to do a HUGE number of experiments ahead of time to nail down the behavior of the material. Non-linear materials like paper, rubber, etc. would be much worse than metals. It's best to stick to a direct system of measurement for any manufacturing process. While it's true one could measure oxygen intake and weigh sweat, water vapor, etc. coming from a person, it's often better to poke a thermometer in his/her mouth when all you want to know is their body temperature.

Mark