View Full Version : Designing with sensors
Munchskull
06-12-2016, 00:37
This last year our robot had some issues with rotation sensors that I would would like to avoid this coming year. Last year we used some magnetic continuous potentiometers to determine our arm posistion. The issue became these pots had a dead zone and then would restart on the other side of the voltage. Looking back on it they were probably not the use of the sensor. My question is what sensors OTHER THAN Versa planetary encoders do people use for position track on arms and other moving parts and how do you integrate them in to your design?
This last year our robot had some issues with rotation sensors that I would would like to avoid this coming year. Last year we used some magnetic continuous potentiometers to determine our arm posistion. The issue became these pots had a dead zone and then would restart on the other side of the voltage. Looking back on it they were probably not the use of the sensor. My question is what sensors OTHER THAN Versa planetary encoders do people use for position track on arms and other moving parts and how do you integrate them in to your design?
US Digital MA3 for absolute encoders. You can also do a normal quadrature encoder and a secondary sensor for zeroing, we did this with WCP Hall Effect sensors on our 2015 elevator.
Munchskull
06-12-2016, 01:00
Are HallEffect sensors used like limit switches?
cadandcookies
06-12-2016, 01:13
Are HallEffect sensors used like limit switches?
Hall Effect sensors can be used for a variety of purposes, including limit switches and something similar to an encoder. Here's a basic overview from wikipedia. (https://en.wikipedia.org/wiki/Hall_effect_sensor)
how do you integrate them in to your design?
3D printed mounts can go almost anywhere, even after the design was built without sensors.
Others have used bent polycarb as encoder mounts to absorb shock or something.
Are HallEffect sensors used like limit switches?
Exactly like limit switches, but the non-contact nature makes them (IMO) much more reliable.
Munchskull
06-12-2016, 01:55
Exactly like limit switches, but the non-contact nature makes them (IMO) much more reliable.
2015 we used a reed switch for the same thing.
Hitchhiker 42
06-12-2016, 10:47
3D printed mounts can go almost anywhere, even after the design was built without sensors.
Others have used bent polycarb as encoder mounts to absorb shock or something.
Additionally, to connect the shaft of a pot to the rotating shaft, we've used a couple of tricks: either get flexible shaft couplings (one of our mentors brought some in from work) or, if you don't have that, some pneumatic tubing with zipties at both ends works well too.
My team used linear actuators to make an arm move up and down, pivoting on an axis. The linear actuator's had potentiometers but they quickly broke, so we added in physical limit switches to stop the arm from going too high (jamming the actuators) or too low (destroying bearings, sometimes jamming, messing up lead screw). We also added in a string potentiometer which actually worked very well for arm position. It isn't always good though because the housing is fragile and the string can sometimes get in the way. AndyMark sells the string potentiometer and I would highly recommend it.
This last year our robot had some issues with rotation sensors that I would would like to avoid this coming year. Last year we used some magnetic continuous potentiometers to determine our arm posistion. The issue became these pots had a dead zone and then would restart on the other side of the voltage. Looking back on it they were probably not the use of the sensor. My question is what sensors OTHER THAN Versa planetary encoders do people use for position track on arms and other moving parts and how do you integrate them in to your design?
We used the MA3-A10-250-N (AndyMark's Absolute magnetic pot) on our shooting arm and had very good luck with it, no dead zone at all. Our high goal shots, among other things, were dependent on it and we were over 80% consistent in Auto. One thing to point out is that if you use v1 of the Universal Breakouts on the Talon SRX's they only supported 3.3v sensors and would therefore give you a dead zone for that 5v pot. In version 2 they've fixed that issue.
Munchskull
06-12-2016, 14:08
We used the MA3-A10-250-N (AndyMark's Absolute magnetic pot) on our shooting arm and had very good luck with it, no dead zone at all. Our high goal shots, among other things, were dependent on it and we were over 80% consistent in Auto. One thing to point out is that if you use v1 of the Universal Breakouts on the Talon SRX's they only supported 3.3v sensors and would therefore give you a dead zone for that 5v pot. In version 2 they've fixed that issue.
Pictures of how you mounted these would be great. Last year we had issues of the entire POT rotating at times.
Luckily (in this case) we use the Victor SP and not Talon SRXs.
Additionally, to connect the shaft of a pot to the rotating shaft, we've used a couple of tricks: either get flexible shaft couplings (one of our mentors brought some in from work) or, if you don't have that, some pneumatic tubing with zipties at both ends works well too.
How well does this actually work? The last trick like this I tried was using hot glue and it was terrible. Right now I am leaning towards drilling and tapping the shaft for using set screws.
..get flexible shaft couplings (one of our mentors brought some in from work)..
There are several varieties on FIRST Choice this year - oh, it just closed a couple of hours ago. For 1/4" to 1/4" solutions, see FC17-122, or a combination of two FC17-110s with an FC17-124 between them.
Note that to couple two 1/4" shafts with tubing, you would need tubing with a 1/4" inside diameter, not outside. I have never done this, but it seems that you would want to keep the length short if your application is sensitive to sensor backlash.
I heard secondhand (no details) about a team who used a 1/4" drive socket universal joint in a coupler to isolate a potentiometer from shock loads. Yep, sounded kludgy to me too.
Hitchhiker 42
06-12-2016, 17:41
Note that to couple two 1/4" shafts with tubing, you would need tubing with a 1/4" inside diameter, not outside. I have never done this, but it seems that you would want to keep the length short if your application is sensitive to sensor backlash.
Absolutely. The tubing was probably about a 3/4" plus what went on each end of the shaft. The only goal is to not have to align the pot and the shaft perfectly.
Pictures of how you mounted these would be great. Last year we had issues of the entire POT rotating at times.
Luckily (in this case) we use the Victor SP and not Talon SRXs.
How well does this actually work? The last trick like this I tried was using hot glue and it was terrible. Right now I am leaning towards drilling and tapping the shaft for using set screws.
We used a lathe to bore a 1/4" hole into the end of the 1/2" hex shaft our arm was rotating on, we then slotted it on the 6 sides of the 1/2" hex shaft by making 3 cuts straight across, and finally added a collar clamp to tighten the end of the hex shaft after the encoder shaft was inserted. We simply used 2x2 angle to mount the POT in place.
If you're interested I could take pictures next time I'm up at the shop.
We used a lathe to bore a 1/4" hole into the end of the 1/2" hex shaft our arm was rotating on, we then slotted it on the 6 sides of the 1/2" hex shaft by making 3 cuts straight across, and finally added a collar clamp to tighten the end of the hex shaft after the encoder shaft was inserted. We simply used 2x2 angle to mount the POT in place.
If you're interested I could take pictures next time I'm up at the shop.
I've done similar as well. works with even just a single slit across. Since it's more or less a rigid coupling, good practice would be to make a somewhat flexible mount so minimal side load is put on the encoder shaft. The encoder body mount simply needs to stop the body of the encoder from rotating, not locate it. Bent 1/16" polycarbonate works well.
chapman1
07-12-2016, 15:46
We used PVC tubing for a U-shaped pickup mechanism in 2014, chosen for its flexibility, so that the arms could spread around the 2' diameter game pieces. The problem was that because it flexed, the encoder magnet at one end of the shaft did not stay in perfect alignment with the encoder.
We solved it by using a gates belt off the pickup shaft to drive a steel shaft for the encoder.
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.