Close range accurate TOF sensor

Hello,

I am looking for a TOF ranging sensor with sub millimetre accuracy… ie returns 3.74 mm instead of rounding to 4 mm … as accurate as possible for distances below 12 cm, for a project I am working on.

Tom

Hi @Tom252259, welcome to the forums :smiley:
Sub-millimeter precision is a pretty demanding application - what’s the project? Is it necessary for the sensor to be optical?

Intuition tells me you might be at the edge of precision for common ToF sensors.
Do you need to resolve absolute measurements or just detect relative change?

Vishay make ToF sensors available in small ranges, some are good for use as a force sensor by measuring the (presumably very small) deflection of a plate. That sounds promising.

Otherwise, we have load-cells that can correlate deflections. You can see the performance of our Load-cell amplifier in this episode of The Factory

Hello,

Answers inline:

Hi @Tom252259, welcome to the forums :smiley:
Sub-millimeter precision is a pretty demanding application - what’s the project? Is it necessary for the sensor to be optical?

Yes, need to sense distance between two pieces of flat steel that are usually moving very small amounts, between range of 20 cm and 30 cm away at .01 mm increments, but sometimes sudden jarring motion, so I want no mechanical connection between them.

Intuition tells me you might be at the edge of precision for common ToF sensors.
Do you need to resolve absolute measurements or just detect relative change?

Absolute measurements

Thank you for your help.

Hi Tom
Like Jeff said, a pretty demanding task.
To measure a resolution of .01mm using TOF techniques the distance there and back is .02mm
This is 0.00002metres. Now light travels at approx 300000000 metres/sec (actually I think from memory it is 299820000 but 300 is new enough).
Now my maths might be out by a decimal place or 2 but to travel 0.02mm would take 6.66 * 10^-14 or 0.00000000000066 seconds.
Good luck with measuring that.
Cheers Bob

Hi Tom

Don’t quite know how you would go about this. The only thing that comes to mind is the measuring apparatus that measures the head position of a milling machine but I am not sure of the resolution. May be only 0.1mm and requires a mechanical connection to one object. The measurements here are relative, you start from a known reference and call it zero and go from there.

Outside a sophisticated Lab you could have a problem with that sort of non contact resolution. There is probably a way to do it using optics to magnify the reflection of one plate (providing it is shiny) and measuring beam deflection. I have seen this done with a reflecting plate on a sensitive galvanometer deflecting a spot behind a frosted screen to measure tiny electric currents. A similar principle could be employed here. Have a laser or some spot light source reflecting off he moving plate and measure the spot movement some distance away to get your amplification of movement.

But I don’t think you are going to get anywhere with TOF measurement.
Cheers Bob

1 Like

Thanks for the physics references to clarify.

What if we change the requirements to be 1/2 or 1/4 of a mm accuracy, is tof going to be able to do that level?

Hi Tom
It all comes down to the smallest time you can measure.
To put it in perspective if it is 1µSec that is a 300m round trip (there and back) which is 150m resolution
Thus 1 nSec would be 0.15 m or 150mm.
We are now getting down to pico seconds which is getting pretty impossible in the home environment.
Cheers Bob

1 Like

Hi @Tom252259 ,

When I was involved in one of the theme park attractions at Fox Studios way back, we were asked to provide a distance measurement system (to play music notes depending on the height of various people). We wound up using an fairly hi-spec Allen Bradley optical sensor that could provide (via measuring its’ current output and converting to 12 bit digital (4096 values in theory). But the best sensor (== $$)would only work over a 0.5m to 3m range (i.e. a close in ‘dead zone’) and the resolution was plus/minus 1 mm.
And there was some jitter in that as well, which didn’t worry us as we eventually only wanted 4 ‘octaves’ of sounds so we threw away most of the bits… And in the automation world you are usually operating at 24VDC or so.

This is a recent A-B sensor, and the spec is still similar …

Murray

1 Like

Hi Murray
Yes you would need a fair sort of ADC to be able to count enough numbers for that resolution. This will have a greater bearing on range of the measurement, in Tom’s case 30cm. As the measuring medium (laser??) travels at the speed of light I was thinking in terms of the clock speed to measure the required small time intervals for 0.01mm resolution. I think somewhere in the region of 15000Ghz give or take a decimal place or two. I make it 1.5 * 10^13. Math might be a bit out but it will be still too fast for “normal” current technology I think.
Cheers Bob

1 Like

Hi Robert,
For the A-B analog input it was 3.90 μA/bit with a 4-20mA sensor attached. And the maximum network scan rate was 500Kbps, but the analog module had its’ own ‘ready to read’ flag so it was converting much slower - about 1 mS at best, and usually up to 32mS for ‘solid’ values.
And that was mixed in with heaps of other network bidirectional data…
Cheers
Murray

1 Like

Measuring small distances is common in machining, it is common to be working with distances in the order of <0.1mm.

Many machines are outfitted with a Digital ReadOut (DRO) which replaces traditional analog dial readouts.
Perhaps @Tom252259 you could use a DRO for such a task? Better still, they commonly have a serial interface that you can read with eg. a microcontroller.

This is well outside CE’s swim-lane. After some light research I’ve found a DRO product sheet from Mitutoyo, which are a very reputable measuring-instrument manufacturer (Mitutoyo fanboy here).

Best of luck with your project :smiley:

2 Likes

Hi Michael
Already suggested something similar

Have not had any feedback from Tom as to whether this system is feasible or not.
Cheers Bob