# Electronically measure distance to the nearest 0.05 mm

#### GregJ7

Joined Jun 7, 2014
42
What is a low cost, reliable method for providing a distance measurement to a CNC controller PCB that is accurate to the nearest 0.05 millimeters? The possible distances would be 0.0 to 1.0 meters. It would be for a CNC intended for wood carving. (If it were intended for metal cutting, the accuracy would need to be even better.)

The first thing I thought of was measuring the strength of a stationary magnetic field, but I doubt I could create a clean enough environment for that. Then I imagined the round trip time of a momentary light or laser blast, but I am pretty sure I can't afford anything that can be accurate to the nearest 15 picoseconds. Now I am wondering if measuring the round trip time of a momentary audio burst is feasible. It would probably need to be self-calibrating to account for changes in temperature and humidity. I don't want to freak out the dogs (or humans) in the neighborhood, so I'm thinking it would need to be a 150k+ Hertz signal. (Some animals can hear very high frequencies—although I doubt anyone nearby has a pet porpoise or beluga whale.) However, are there even low cost emitters and detectors that work for those audio frequencies? (I couldn't find any at any cost.) It's easy to build a crude emitter, but I have no idea what its capabilities or limitations would be.

Any other methods, perhaps more realistic, of measuring come to mind?

#### drc_567

Joined Dec 29, 2008
881
If a metal bar with a mark or registration at 50 μm spacings could be fabricated, then it may be possible to devise a slider with a laser diode and phototransistor. The slider would provide a single count as each mark passed under the diode-transistor assembly. The count accumulation would serve as an input to the CNC mechanism.
... One or two problems to sort out, like how to arrange for two axis operation.

#### Sensacell

Joined Jun 19, 2012
2,595
Look to the real world of existing CNC technology.

It's a very mature science, it's hard to imagine that you could come up with something better than what already exists.

#### WBahn

Joined Mar 31, 2012
26,141
What is a low cost, reliable method for providing a distance measurement to a CNC controller PCB that is accurate to the nearest 0.05 millimeters? The possible distances would be 0.0 to 1.0 meters. It would be for a CNC intended for wood carving. (If it were intended for metal cutting, the accuracy would need to be even better.)

The first thing I thought of was measuring the strength of a stationary magnetic field, but I doubt I could create a clean enough environment for that. Then I imagined the round trip time of a momentary light or laser blast, but I am pretty sure I can't afford anything that can be accurate to the nearest 15 picoseconds. Now I am wondering if measuring the round trip time of a momentary audio burst is feasible. It would probably need to be self-calibrating to account for changes in temperature and humidity. I don't want to freak out the dogs (or humans) in the neighborhood, so I'm thinking it would need to be a 150k+ Hertz signal. (Some animals can hear very high frequencies—although I doubt anyone nearby has a pet porpoise or beluga whale.) However, are there even low cost emitters and detectors that work for those audio frequencies? (I couldn't find any at any cost.) It's easy to build a crude emitter, but I have no idea what its capabilities or limitations would be.

Any other methods, perhaps more realistic, of measuring come to mind?
If it's for a CNC (computer numerical control), then you need something that is compatible with whatever controllers you are using. Check with the manufacturer.

If it's just for DRO (digital read-out), then a much more reliable way of measuring is to simply use one of the many off-the-shelf systems designed for that purpose.

#### GregJ7

Joined Jun 7, 2014
42
I'd build the controller board. It's a project to enjoy designing and building, not just to have a CNC—besides which I can't afford a CNC with the features I want, which are not all that special, but don't appear in the low-cost CNCs. One of them, for example, is 0.05 mm accuracy. (Look at "intricate" carvings done with CNCs under $4k.) #### jpanhalt Joined Jan 18, 2008 10,210 For that level of accuracy, I would use a linear encoder such as a DRO (Digital Read Out) as used on mills, lathes, and so forth. Mine are Mitutoyo (https://shop.mitutoyo.eu/web/mitutoyo/en/mitutoyo/Linear Scale/DRO linear scale & counter/index.xhtml), but they do get quite expensive the longer they are. The Mitutoyo controller has SPI output as I recall. Some devices use a standard SPI, some are Mitutoyo modified, but it is still an easy interface to use. The encoder by itself is considerably cheaper. Zoro (Zoro.com) advertises a 700 mm one for$287. Price takes a good jump at 1000 mm. My experience is with Mitutoyo, but those comments probably apply to other brands as well.

Edit: Depending on the accuracy of your drive systems, you can also use a rotary encoder on those drives. With ground ball screws, you can be well within that accuracy. Even modern rolled versions can get you there. You can also create an error file and get rid of some of the mechanical error.

Last edited:

#### MeSat

Joined May 16, 2019
16
This is related to a project that I am working on for my retirement project. I am assuming you want to measure outside your stepper motor and system electronics as a reference.

I designed but didn't build an optical encoder for a project years ago that would provide the accuracy you want but it was going to be expensive at the time.

For your CNC, how is it driven, belt, rack and pinion, screw drive? All have different accuracy levels to them and that is important if you want to be that accurate with your measurements. I have a laser system and all measurements are calculated from the reference point I start with and there is no need to measure anything from an external source.

I have seen micrometer resolution magnetic pickups that as similar to what is in a caliper and from what I see on Ebay, not that expensive now. Prices depend on the length.

#### Wolframore

Joined Jan 21, 2019
1,783
It really depends on your size limits and accuracy requirements

glass scales are considered the most accurate with micron level accuracy... they are also very expensive/

If your accuracy requirement is less (since wood is not that stable) there are a number of options

If you're using ball screw drives you can put a rotary encoder on it and get pretty good accuracy. If your travel is relatively small people use the cheap electronic verniers (6") that has digital outputs built into them these are not bad for the X on a lathe.. Use servo motors and position is built in...

I've seen systems made by Wixey and iGaging...

I have some MAXIN product that is a magnetic linear encoder.. haven't played with it yet.

many options... using something like a laser might work as long as you keep the beam path clear of debris. But unless it's completely sealed stuff gets everywhere.

With wood 0.040" (1mm) will more than suffice.

For metal you will need 0.025mm (0.001") or better tolerance and a machine strong enough to contain the pressures required safely and without transmitting vibrations to your work piece. My mill is a lightweight at 600lbs... I really need the 4,000lb beast to get better results especially with 1 meter travel... but for softer metals like brass or aluminum you can do quite a bit with a lighter machine.

#### Papabravo

Joined Feb 24, 2006
14,382
Use an LVDT. We could use them to measure the diameter of a workpiece in a grinding machine to nanometers.

Last edited:

#### MeSat

Joined May 16, 2019
16
Thank you for the suggestion for LVDT (Linear Variable Differential Transformer). I have never heard of it before and it may be something I can use at my work as we do look for very small displacements. I will have to learn more about them.

They are expensive though for a small range. 2.5 to 10mm at $140. #### Back to school Joined May 22, 2019 106 Have you tried doing a web search for used CNC parts? Here in the rust belt there are a lot of shops that deal with the reconditioning of machines and parts sales. You can get a worn, 3 axis table with ball screws and synchro's from a bench top machine for a reasonable price around here. A "worn table" from a metal cutting machine could be fine for what you're after. I'm not sure I would worry about vibration unless you're in a hurry with you're work. If you keep the diameter and depth of cut appropriate to the machine vibration shouldn't be a problem. The same goes for controllers from the same places. There's also a lot of free shareware available for running the machines. At least there was at one time. Edit: Buying over the internet can be expensive because of shipping and handling. Most of this stuff is heavy. Local pickup is usually a better choice. #### Papabravo Joined Feb 24, 2006 14,382 Thank you for the suggestion for LVDT (Linear Variable Differential Transformer). I have never heard of it before and it may be something I can use at my work as we do look for very small displacements. I will have to learn more about them. They are expensive though for a small range. 2.5 to 10mm at$140.
Talk to these guys. They are experts, and I worked for them once upon a time.

http://www.controlgaging.com/