I am working on a project where I need to measure lengt of the objects floating on a surface. To perform that kind of sensing I am currently focusing on mouse sensors.

I first tried ADNS 6010; however, 8-bit delta x/y output (as displacement amount registers) of it became so small for the speed I desire which is 2 meter/sec. Even max frequency of reading did not be useful.

Then I tried a ADNS 9500 sensor which has 16 bit delta x/y registers. I get better results than 6010. But the problem is the accuracy of the output. When I measured same lenght with different speeds I get the output with 10% variaton. Also difference surfaces gives different results.

Now I am stuck which method to perform. I am thinking of changing the lense to focus a wide area, or changing technology. Changing lense is difficult to adjust a good focus point. It needs a well designed mechanical equipment. And chaning thechnology is a problem for me, since there is no other technology I can think of. So I would like to ask your recommendations.

enter image description here


closed as not a real question by Olin Lathrop, zebonaut, stevenvh, W5VO Oct 6 '12 at 22:50

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center. If this question can be reworded to fit the rules in the help center, please edit the question.

  • \$\begingroup\$ There is much you left out. What does floating on a surface really mean? You originally asked about measuring length, but then started talking about speed, which makes little sense. Come here with a problem and specify it well, not with some supposed solution. What are the real requirements? How long are these things? How accurate? How fast must each be measured, etc? It should have been clear to you this problem is insufficiently specified. \$\endgroup\$ – Olin Lathrop Jul 21 '12 at 21:34
  • \$\begingroup\$ The floating means floating nylon on a production line. The nylon can float in the machinge with different speeds, and their its measurement should be valid for the sensor for each of those speeds. So the lenght can be hundreds of meters. The accuracy reqirement is 99%. And it can float upto 3m/s. \$\endgroup\$ – Furkan Turan Jul 21 '12 at 21:52
  • \$\begingroup\$ Methods to measure length depend on more accurate requirements. Please define physical constraints of objects and interface, Motion {direction and speed} , time interval { range of measurement }, Height { min/max} , Other factors {reflectivity, transmissibility } . One might consider Interferometry reflection methods using a diffraction grating. Also what mouse are you using LED or Laser? what is spec'd DPI ? what is range of length {min/max}? \$\endgroup\$ – Sunnyskyguy EE75 Jul 21 '12 at 21:55
  • \$\begingroup\$ The old thechnology used is rotating wheel, but it hurt the nylon. Therefore, I am seeking for an image processed solution. Reflection and transmissiblity may vary according to different products of the machine. And as I stated in the question, the sensors I tried gives different results for different nylon types. There is not a limit for sensing range, max 1 meter can be said. For mouse snesors I tried 2 milimeters. The sensors I have used was both Laser, and CPI (Count per inch) can be set. For ADNS it can be set from 90 to 5400. \$\endgroup\$ – Furkan Turan Jul 21 '12 at 22:00
  • 1
    \$\begingroup\$ You were asked what floating on a surface really meant, and you defined it as floating nylon on a production line. Seriously!? You can't see for yourself that someone that doesn't already know what that means isn't going to learn it from your statement? At least half of engineering is properly specifying the problem. You need to learn to do that. In the mean time it is too difficult to understand what is being asked here, and this question should be closed unless you are more forthcoming with real information promptly. \$\endgroup\$ – Olin Lathrop Jul 21 '12 at 22:09

I see a dwg of a roll of material spooled that is driven onto an anti-static roller track conveyor and the issue is measure quality of a bag length cut by a cutter which may be sync'd to the sheet feeder rate. I am not sure if it is continuous feed or stop and go.

This appears to be a measurement error in variability of a cut length of 10% by speed variation rather than a a variation in actual cut length. Pls Confirm.

Examine sources of variability in length; - Variation in material slip to reach feed speed after cut. - resolution of optical sensor for start/end of sheet. 2m/s at 1m max length with 10% error means 10cm error or 10% of 0.5sec = 50mS error in sheet time interval @ 2m/s - so objective is confirm sheet speed is uniform from start of measurement to end < 1% - ensure measurement time interval is less than 1mS ?

I would consider using UV LED or IR that is attenuates optical path through material and use interruptor method for time interval measurement of length.

Consider using LED emitter overhead with heat-shrink tube precisely aimed over detector < 1m apart.. THis method can detect wire size interruptions of light such as a resistor but plastic needs to attenuate light sufficient to trigger detector.

If IR wavelength is OK... I would use Pulsed LED with IRDA I SIR tranceiver rated for 115KHz for Tx path and detector with hysteresis to detect edge of plastic.

Vishay make many IRDA devices in this range that operate ~ 1m or more distance. If this works I can detail path design of optic tube to prevent stray light and get 1mm resolution on edges.

  • \$\begingroup\$ I am thinking of the solutions you proposed. But the problem is not all the machines are similar to the one in the drawing. It is just the one made for cutting. But I am going to study IR technology which I am not familiar with. Thank you for your recommendation. \$\endgroup\$ – Furkan Turan Jul 22 '12 at 13:51

Mouse sensors are interesting technology, but they do not have the accuracy you require.

I've used them before to sense web displacement, and the gain (counts per inch) changes quite a bit with minor changes in distance from lens to web. I suppose you could possibly rig some kind of telecentric lens system to reduce this, but getting better than 10% accuracy with a mouse sensor is just not gonna happen in my experience.


Not the answer you're looking for? Browse other questions tagged or ask your own question.