I am designing a sensor for a machine that manufactures small springs, and I would like to know if anyone has knowledge of a capacitive sensor capable of measuring precise displacements in the range of 0.2 mm, with the result of this displacement provided as an analog output or another type of communication proportional to the displacement.
There are sensors from Micro-Epsilon, but they are too expensive for the application. I would like to know if there is another sensor I could use
Ok. How many are you needing over what time frame? Since the Micro-Epsilon unit has all the parameters needed for the project, please tell us those the parameters.
What kind of displacement in a 5 DOF room?
Think of capacitive sensors—they’re elegant, but in a spring-forming machine with oil mist, metal dust, and vibration, they can drift or pick up unwanted signals, especially over a small 0.2 mm range. A better fit would be a short-range analog inductive (eddy-current) sensor: non-contact, rugged with metal, repeatable to a few microns, and with a smooth analog output like 0–10 V or 4–20 mA. If you can afford slight contact wear, a miniature spring-loaded LVDT is even more precise and stable. Less expensive but trickier to engineer is a magnet + linear Hall effect setup: cheap and compact, but it requires careful calibration and mechanical design. You might also explore laser triangulation if non-contact measurement on non-metal targets is required, but those tend to cost more and are sensitive to sprays or surface finish.
For a DIY angle—and especially if you’d like to experiment with capacitive principles in a simpler, low-risk context—there’s a neat project here: https://www.pcbway.com/project/shareproject/DIY_simple_Capacitive_Proximity_sensor_392be74b.html . It’s a basic capacitive proximity sensor that’s great for learning and proof-of-concept, but it won’t match the precision or environmental robustness you’d get from a properly engineered inductive or LVDT solution.
For example: let’s say the sensor starts its reading at 5 mm of distance when the spring is approaching horizontally towards the sensor. As the spring gets closer to the sensor at 4 mm, 3 mm, 2 mm, this sensor should be able to provide a voltage proportional to that approach with precision. In this way, I can measure the spring length of, for instance, 10 mm. If the spring is below 9.8 mm, it is defective; if it is above 10.2 mm, it is also defective. This defines the acceptance tolerance of the spring as ±0.2 mm. And one very important detail: the sensor must not touch the spring.
Thank you very much for your suggestion and analysis. I considered using an inductive sensor, but what prevents me is the different types of materials used, such as stainless steel, copper, and brass. As for LVDT, the only difficulty I see is that when the spring is cut, it might get caught on the contact rod. I thought about using the AD7746, but while researching it seems quite problematic to program the registers. I will take a look at the project you recommended, thank you very much.
Instead I'd try two light barriers for the min and max spring size.