I have a test rig whereby a shaft (an M4 threaded bar) is being pulled ~11.5mm by a mechanism (rather not detail that) in about 20ms. I want to plot the movement of this shaft vs the mechanism pulling it.

I made up my own sensor using an IR switch with a portion of a hair comb passing through it. Of course a custom 3D printed solution would have been ideal but I don't have that.

I got the finest comb I could find and measured the distance between first and last tooth and divided the length by the number of teeth (minus 1) which gives me a 1.54mm pitch, or 0.77mm between tooth and gap.

I made up an L shaped bracket to hold the comb and passed through the IR switch and put a scope on that.

So, I was expecting to see 11.5/1.54 = 7.5 full on/offs on the scope or 15 high/low transitions.

In practice I got much more than I expected. I figured my shaft is bouncing when it hits the stops so tried to damp it by creating a deformable structure with masking tape around the lock nuts that limit travel. This improved things greatly as my last pic shows, but this was the best of several tests. I need data I can trust.

I'm pretty sure the teeth are not hitting the switch, and I shortened the teeth to make them less flexible.

Unfortunately both the comb and switch are black, but hopefully you can see what's going on. For scale, the screws holding the comb are M3.

Any thoughts/suggestions please?


EDIT 10th Aug

As per the suggestions in the comments, I tried it by hand and was consistently getting the expected results, 7 full waves which is 10.8mm, within 1 step of what I should get and I'm happy that it's repeatable.

EDIT 19 Aug

Sorry for the delay in this update, I've been working on and off at it. I think I'm getting pretty good results now. I re-made the bracket that holds the comb (could not figure a way to attach directly to the bar) and used glue instead of screws.

For damping, I used those felt like things you put on furniture etc to prevent scratches. And, I soaked them in oil to improve damping. Finally, I added a small spring that comes into play near the end of the stoke. See last pic.

enter image description here enter image description here enter image description here enter image description here enter image description here enter image description here

enter image description here

  • 1
    $\begingroup$ In order to stop stored kinetic energy without a bounce back, you need more of the same damping. It is a damping resistance to match the impedance of the crashing object at some g value. What are you expecting? a constant force to prevent backlash or springback? $\endgroup$ Aug 10, 2022 at 14:26
  • 1
    $\begingroup$ the use of the comb is very clever $\endgroup$
    – jsotola
    Aug 10, 2022 at 15:20
  • 1
    $\begingroup$ steel is elastic and there will always be bounce back for an end-stop until you have achieved perfect damping in the shortest desirable distance. Your purpose needs to be defined. $\endgroup$ Aug 11, 2022 at 17:42
  • 1
    $\begingroup$ Can you describe what exact velocity at end stop you are trying to achieve with this ? A dead end will reflect energy and push back. $\endgroup$ Aug 11, 2022 at 20:21
  • 1
    $\begingroup$ @TonyStewartEE75 thanks for your help. A new damper made the final difference, see my updated OP. $\endgroup$
    – KevInSol
    Aug 19, 2022 at 16:31

1 Answer 1


The placement of your sensor on that long arm is very bad. It's going to vibrate like a tuning fork.

The arm is long which gives it a lot of flex. And those overly long bolts make it heavy which increases the inertia at the comb. These combine to give a large vibrational amplitude. There is also very little damping, as others have mentioned.

The best solution would be to attach the comb rigidly to the main shaft.

You may see some improvement by lightening the end of the sensor arm as much as possible (perhaps switching to plastic screws or at least shorter ones). Shortening and stiffening the arm, and adding damping (maybe a thin plastic truss).

  • $\begingroup$ I replaced the screws with glue, no difference really. Re-working the comb bracket, I won't get it finished today, $\endgroup$
    – KevInSol
    Aug 11, 2022 at 15:49
  • $\begingroup$ Just to re-iterate. If you can attach the comb to the threaded shaft directly that would be by far the best option. $\endgroup$
    – Drew
    Aug 12, 2022 at 1:29
  • $\begingroup$ I updated my question to show the results. I think my new bracket made the most difference, thanks $\endgroup$
    – KevInSol
    Aug 19, 2022 at 16:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.