0
\$\begingroup\$

I found this sensor VCNL4040 and it has the emitter angle and the sensor angle.

enter image description here

Question:

How good is a Time-Of-Flight (light) sensor to measure small objects that are 100mm away and about 20x20mm in square? The material is metallic, gray with a small film of transparent yellow oil onto. So it quite shiny.

The accuracy must be in ±1 mm.

If you have a better one, please tell me.

\$\endgroup\$
6
  • \$\begingroup\$ Does the target rotate or only translate? Do you care about motion perpendicular to the ranging axis? Micron class accuracy is pretty doable with a laser pointer and a webcam if the object can't rotate (just look at which pixel the laser beam hits and do the trigonometry). \$\endgroup\$
    – user1850479
    Commented May 25, 2021 at 19:12
  • \$\begingroup\$ @user1850479 The target only translate. No, this will be a very slow motion. I need to measure and get a least an analog value. \$\endgroup\$
    – euraad
    Commented May 25, 2021 at 19:32
  • \$\begingroup\$ I'd probably do the laser+trigonometry approach then. You can possibly find chips that integrate the laser and sensor or make your own. I actually built something similar for a related project where the object couldn't translate but it could tilt and it works quite well. Can easily back out two axis of tilt (by assuming no translation) to tiny fractions of a degree. You could do the reverse and solve for translation. \$\endgroup\$
    – user1850479
    Commented May 25, 2021 at 20:52
  • 1
    \$\begingroup\$ @user1850479 Thank you! What laser did you use? Cheap components or very expensive? Do you have a article number? \$\endgroup\$
    – euraad
    Commented May 25, 2021 at 20:56
  • \$\begingroup\$ I used a cheap laser diode module with adjustable focus (few dollars on Amazon) and a webcam with the lens removed. Getting the laser intensity low enough to see on a CMOS sensor without saturation required an reflective attenuating filter for attenuation at the lowest current the diode would lase, which was more expensive. \$\endgroup\$
    – user1850479
    Commented May 25, 2021 at 22:28

2 Answers 2

Reset to default
1
\$\begingroup\$

This is an IR proximity sensor, not a time-of-flight sensor. It basically shoots out an IR cone and measures how much light is backscattered. While this works for presence detection, it is not suitable for ranging.

2 mm range resolution is very tough ask for a time-of-flight light sensor. Think about it: the speed of light is \$3 \times 10^8\ m/s\$. Range resolution for an unmodulated pulsed TOF laser system is given as \$\Delta R = c\frac{\Delta t}{2}\$, where \$ \Delta t\$ is the pusle length. If you do the math, you find that for 2 mm range resolution, you need \$t=26\ ps\$. That's not realistic, especially for a low-cost commodity TOF system.

I'd suggest looking at non-TOF approaches, if possible. Off the top of my head:

  • Free-space ultrasound can meet the sort of resolution requirements you require, although clutter rejection may be a pain if the operating environment changes. You also may need to calibrate for air pressure, temperature, and humidity.
  • Computer vision could work very well, especially if the system geometry allows setting up a camera looking at the scene in profile.
  • Thinking really outside the box, what about interferometric radar? Even mmWave radar will not get the range resolution you want (COTS solutions do ~4 GHz bandwidth which corresponds to a range resolution of a few centimeters), but if you looked at the phase of the return pulse, you might be able to subdivide the resolution cell.
\$\endgroup\$
3
  • \$\begingroup\$ The question asked for accuracy, but you're explaining resolution. They're related but different things, and getting 1 mm accuracy is possible without having 1 mm resolution. Just requires high SNR and good calibration and you can be accurate to much less than your resolution. \$\endgroup\$
    – user1850479
    Commented May 25, 2021 at 19:15
  • \$\begingroup\$ Hmm...ultra sonic radar? \$\endgroup\$
    – euraad
    Commented May 25, 2021 at 19:29
  • \$\begingroup\$ By the way! I like your answer! :) \$\endgroup\$
    – euraad
    Commented May 25, 2021 at 19:36
2
\$\begingroup\$

This is not a Time-of-flight sensor, rather a fancy reflectivity sensor and you cannot avoid doing the optical design of your requirements to meet a low error rate without doing the homework I requested on your previous similar question.

\$\endgroup\$
4
  • \$\begingroup\$ Do you have any better idea of a sensor? \$\endgroup\$
    – euraad
    Commented May 25, 2021 at 18:21
  • \$\begingroup\$ Same answer as last time. Sense what reflectance ratio? Reject what and where? \$\endgroup\$
    – D.A.S.
    Commented May 25, 2021 at 18:36
  • \$\begingroup\$ I don't know the reflecance ratio, as I said before. \$\endgroup\$
    – euraad
    Commented May 25, 2021 at 18:38
  • \$\begingroup\$ Then you have to test it and find out. \$\endgroup\$
    – D.A.S.
    Commented May 25, 2021 at 18:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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