Selecting Scales for a DRO - Six Parameters that Matter

Thursday, December 5, 2013

Things were easy when iGaging "Remote DRO" were the only scales Android DRO supported. Now that it can work with most of the common scales things are becoming a bit more confusing, judging from the number of emails I've been getting lately. With other things being equal, the choice of scales will have the biggest impact on you DRO's performance. There is a number of options, ranging from cheap Chinese calipers to glass scales and other quadrature encoders. Although I can't definitively answer the question "which scales should I use for my DRO setup", in this post I'll try to explain which parameters matter most in a DRO application.

Most of the scales used in hobby DROs fall under one of the two major categories: capacitive scales and quadrature encoders. The former category encompasses things like AccuRemote and iGaging DigiMag* scales, digital calipers and the common "Chinese" linear scales that look like large calipers with jaws cut off. The latter consists of glass linear scales and rotary encoders as well as magnetic scales. Generally, capacitive scales are much less expensive but have some shortcomings, thus most of the commercial DRO units use either glass or magnetic scales. If money is not an issue you can stop reading and get a set of brand-name glass scales

*Many retailers have the word "magnetic" in the product name or descriptions. This is a bit misleading, since the scales are not magnetic; the DRO display can be magnetically attached to the machine. These are capacitive scales and should not be confused with magnetic quadrature scales.

Deciding Factors

There are a few factors that will drive your decision. In addition to the technical parameters, such as accuracy, resolution and reliability, there is an issue of cost and availability. I think a good place to start is to clarify what each of those parameters means and how it might influence the end result.


Resolution determines how fine the measurements will be. For example, the micromenter in the picture above has much better resolutions at 0.0001" than either of the rulers. For the purposes of the DRO the value that matters is counts per inch [for linear scales] and counts per revolution for rotary encoders.


Accuracy is often confused with resolution but it's a very different parameter. Accuracy tells you how close the reading from the scale will match the actual dimension. For example, when you measure a 1" standard, an accurate scale will read exactly one inch, while a bad one will show a different value. It's possible to have very high resolution but very poor accuracy. I.e. if I grind off some metal off the micrometer's anvil, it will still have the same resolution but won't be accurate at all.

Side note: Please keep in mind that some manufacturers/retailers often use the term "accuracy" to encompass such important parameters as repeatability, hysteresis and runout.


Repeatability defines how close together a set of successive readings will be to each other. For example, I can affect the readings from a micrometer by applying different amounts of torque, so the readings won't be repeatable (unless I use the clutch properly). In other words, when you move a scale by exactly one inch several times (from the same starting position), a repeatable scale will provide readings that are identical or very close together.


Hysteresis is the amount of travel the reading head can move before the scale starts registering the change in position. The best way to see the hysteresis is to first move the scale one way by exactly one inch in one direction, then by the same amount in the opposite direction, and compare the readout. In essence this can be thought of as scale's backlash.


Runout is another aspect of the overall accuracy that is as important as the repeatability. In a nutshell it describes how uniform the resolution of the scales is. For instance, when you move the scale from one end to the other in exactly one inch increments a good scale will give you a series of identical readings. You can think of runout as deviation in resolution over the scales length. If there is any difference it will lead to an error of measurement that will accumulate over longer distances.


Reliability is a somewhat subjective value and depends to a large degree of the shop environment, usage etc. Every scale will fail under some conditions, so it's up to you to determine what your particular usage envelope will be and if the scales are fit for it. Never the less some scales will be more reliable and last longer as a result of better protection, higher quality materials etc.

Practical Implications

All of the above parameters will influence the overall performance of the DRO. I've noticed that people tend to be most concerned with the resolution. Although its true that [to a point] a higher resolution will let you hold tighter tolerances, it's only one of the key factors you should look at. In practice repeatability is even more important specification, as it determines the overall performance of a DRO. For instance, 0.0001" resolution is pretty much worthless when the scale is only repeatable to 0.005". Moreover, as long as the scale is repeatable, the software can compensate for the offset, as long as it's constant.
Consider the illustration below; the first two cases will work equally well in a DRO, since the application will compensate for the offset. Third case is less precise but will get you in the "ball park". The last case, though, is pretty much worthless for any precision work as the readings will be all over the place.

Accuracy vs. Repeatability

Second most important property is the runout. Similar to the repeatability, runout determines the DRO's accuracy. Imagine that you need to drill a row of holes exactly 1.000" apart. If a scale has high runout, the holes will not be spaced equally.

Runout can be thought of as repeatability over the scales length

Finally, high hysteresis will add some "backlash" to the DRO. Most glass and magnetic scales have very small hysteresis; calipers and other capacitive scales list this parameter, unfortunately.

Effect of scale hysteresis is similar to backlash
Most of the scales you will come across will have resolution of at least 0.0005", which is more than adequate for most metal work.

One value that I haven't mentioned above is the scales refresh rate. I.e. the frequency at which the scale will report the position to the interface. If the refresh rate is too low there will be perceivable lag between the machine's movement and the DRO display. In my experience a refresh rate of 20Hz or above feels pretty much "real-time", while 5Hz and less is noticeably laggy.


  1. Nice and well written analysis. Everything seems much clearer now.

  2. Yuri, I would like to add a controller for (older) Heidenhain glass scales (example: LS803). These have an analog output, two phases. Each phase provides about 25 microamps, peak-to-peak. To help me in repairing a broken scale I build a transimpedance amplifier that converts this to a 5 volt peak-to-peak signal. The typical period of the scales is 40 microns or 20 microns. So to count, I need to track the "phase circle": to get one count every five microns I need to identify either 8 or 4 counts per circle. This is easy enough to code up. Would you suggest I try producing new controller code or modifying the existing code to do this? Bruce

    1. Bruce,
      If I were you, I would use a circuit similar to the "voltage converter" I built for the mixed scale controller to convert your since/cosine waves to a 50% duty cycle square wave. You don't want to be reading analogue values via the micro-controller, unless you're willing to move up to an ARM M3 or M4-based DSP. MSP430 (or Arduino for that matter) don't have enough horsepower to do that.
      Hope this makes sense.