Detect "clicking" of spring plungers. Need some feedback on current solution

[Vince_vd]

Hi

I have a motorized assembly that reaches a stop at some point, where spring plungers act as a clutch to spare the motor. I need to be able to detect the "clicking" of the spring plungers to make the motor change direction when it happens. 

I have a working program at the moment, but I feel like it could be done easier/more reliable so I was hoping for some feedback. What I do now is the following:

- Continuously get current using DAQ NI USB 6009

- Apply a bandpass filter to the signal which makes the clicking clearly visible

- Apply FFT to this signal

- Trigger when FFT rises above certain amplitude

- Start a timer for 1 second and count how often it is triggered (I collect 250 samples 4 times a second). I count the clicks because changing the motors direction wil also make the FFT spike.

- Change direction if 2 or more clicks are detected in 1 second.

I have attached a screenshot of the signal after the bandpass and from my VI. The attached VI can read data from the attached lvm file with real readings to simulate the signal. 

[JonasM]

With the low cost equipment and the very clear amplitude change you have between normal mode, clicking and change direction I would say you method works if you feel it is fast enough.

Other options would be linear encoders that would give you the position from counters, and you wouldn't have to spend time on FFT calucations etc.

[JoeMoore1331]

Why not just use limit switches?

[Vince_vd]

---

@JonasM wrote:

With the low cost equipment and the very clear amplitude change you have between normal mode, clicking and change direction I would say you method works if you feel it is fast enough.

Other options would be linear encoders that would give you the position from counters, and you wouldn't have to spend time on FFT calucations etc.

---

Well I'm not sure yet if it is fast enough. In the situation I described here it seems to be ok since it'll only be run sporadically. There will be another assembly though without the physical stop, here the detection vi would have to run continiously to detect the clicking once something obstructs the motor (which will definitely happen in time). This is the situation where I'm a bit afraid keeping this running all the time might take up a bit much of the resources.

---

@JoeMoore1331 wrote:

Why not just use limit switches?

---

Limit switches aren't reliable enough sadly. Once it is up and running it'll be running for a long time withou anyone being able to get to it, and in the second situation I described above they wouln't work anyway. Thanks for your responses though, I had already given up hope of anyone finding this question 🙂 

[JoeMoore1331]

I don't understand how limit switches aren't reliable enough? I've used them with every linear slide that I've used and some have a repeatability of around 0.01mm!

Mechanically something sounds wrong with your setup if you have no idea where the motors are and are relying on a 'click' from your clutch - what if your clutch fails and your motor keeps driving your shaft?

Every setup I've used has switches that operate before the clutch, the clutch should be a backup incase the limit switches fail and not your primary indication of when you have hit your limits.