Has anyone had issues with limit switches tripping due to excessive vibraton?
No, but a way to fix that would to be to put a filter on your limit switch in code. That way it has to stay down for a certain amount of time before the code reads it as tripped.
What limit switches are you using? Also, that much vibration may point to a mechanical issue (for example, non symmetric shooter wheels)
We have noticed back EMF noise generated from the motors to cause false tripping “limit switch closures” on our robots. Many of the false triggers can last for a second or two. Filtering in code, suppressing via capacitor, using 5 volts as input, using ferrite doughnuts, and assuring good wiring practice(keeping wires as short as possible for motor and signals, twisting pairs, not running signal wires near or in parallel with motor power wires). generally eliminates these issues. Be sure to ward off this in all areas to limit the issue. Have you tried to put a meter on the limit switch? Start by using a AA battery in-line and set meeter to hold peak or min/max hold (assure the robot wires are not connect) and then drive the robot causing vibration to see if it actually mechanically opens or bounces? Try to confirm that the limit is mechanically vibrating open and closed. We have often thought that we must have a bad limit or excessive vibration only to find noise.
One other simple thing to try is just making a mechanical “filter”. This year, my team had issues where a limit switch with a longer paddle was bouncing up and down, preventing us from using our lifter. The solution: We wrapped a piece of surgical tubing around the paddle, and attached it above the limit switch, so the tubing was slightly stretched. It worked great, and we didn’t have any more false triggers.
1 Like
Do you have a picture of this?
What is a good robust type of limit switch?
We had an issue where the switch where the micro switch that we used to position frisbees before the shooter would bounce mechanically, ie occasionally it would detect two frisbees passing through instead of the one. The switch was set up so that it had a flexible surgical tubing extension that contacted the frisbee. We found the extra mass of the tubing would give it a higher inertia, and give it a higher tendancey to bounce.
We solved this by removing the tubing and by repositioning the switch so that it contacted a bent metal section of the lever, in tandem software also implemented debouncing code. (Which I think we disabled after fixing the problem mechanically)
As for what limit switch, in Australia we use: http://australia.rs-online.com/web/p/microswitches/515735/ It is pretty much identical to the one that came in the KoP (except grey, and slightly higher current rating) and has been very versatile.
Tim
Lonnie,
I am a little surprised at the noise problems you are having. If you use the normally closed contacts, then the line impedance is so low (effectively zero ohms) that noise would only affect the input circuitry if a motor was directly connected to the switch. Even when using the normally open contacts there is still a low enough impedance by the DSC input to be relatively immune to noise pickup. What you may have is an intermittent switch contact that is affected with vibration. The result would be a rapidly changing trigger to the DSC. This should be easily diagnosed by rapping the switch with the insulated end (handle) of a screwdriver. Microswitches can be damaged by rough handling of the exterior contacts. In rare cases, the internal spring can be damaged or broken resulting in the effect you describe.
Is there any way that i can see the code that you wrote?
612 uses similar logic with our joysticks:
(see GetRawButton())
Hope this helps!
To update this problem has not yet been solved.
It has been determined that vibration inside the limit switch is causing the problem.
Hardware solutions would be prefered.
Perhaps vibration isolating the switch by mounting it on shock absorbing foam rather than directly to a solid piece of vibrating frame.
Maybe a firmer style switch like the big ones we’ve received in older kits.
Recommend going optical.
(Get the breakout board with the photo interrupter.)
Or maybe try this:
If nothing else, extend a small, opaque tab in a position such that the above products give you a return at an appropriate location.
We tested the photo interrupter out for a home made linear encoder this year. While we never used it, the concept worked quite well, with the interrupter/optical limit switch hooked to the digital input on the sidecar. The Optical Detector / Phototransistor could pointed across the path of your mechanism, but close enough to the path so that when the mechanism passed across the interrupter, you got a response. Should be very accurate. While we have not tried this, I plan to test out the concept this summer, to have another item in our bag of tricks.
Can you post a photo of the mounting of the switch. These are used in industrial applications all the time and vibration is rarely a problem. Often the actuator is actually the problem and that is easily solved with external damping. As I stated earlier it is always possible you have a damaged switch. The construction of the microswitch uses a spring to hold the contacts in place.
Are you experiencing “contact bounce” instead? There are certainly electronic and programming techniques that can eliminate this condition.
We had some issues with limit switches and noise this year. I think we worked it out. We have some good troubleshooting tools and methods that we can help you out with.
Would you like to meet up (our team and your team) at some point this summer? We have some summer R&D projects we’d like to do as well… 
Non contact object detection is an alternative to a limit switch. We used these switches in 2012.
http://www.pololu.com/catalog/product/1132
We did have problems with them in direct sun light and certain florescent lighting. They are very good if they are shielded from direct lighting. We also had to add a 22 uf cap on the power supply to eliminate noise. Later it was found that the DS was the problem. If you go to the website you will see some other IR distance sensors. Optical break beam is also a possibility.