Sensor Recommendations

[DavidB]

Currently I'm trying to get as much information on the variety of sensors you can use in FRC, such as potentiometers etc, as possible. So basically I have four questions

1. What kind of sensors have you used the most in the past

2. What kind of situations have you used them for and how were the results.

3. What brand of sensors have you found to be the most durable and reliable (and cheap) or does FRC have its own specific type it recommends.

4. Where do you purchase your sensors.

Thank you for any help you can provide, because this is my rookie year I'm trying to learn as much as I can before season kick off.

DavidB

[samir13k]

1501 uses gyros from sparkfun.com whenever we do arcade drive. It helps keep our robot running on a straight path. We also utilize ultrasonic sensors to give us feedback as to our position on the feild by using surrounding lanmarks in the field such as the walls in overdrive. We also used them to find the overpass this year. We did stop using them half way through the game.

But as i mentioned earlier GYROS!!!

[EricH]

For sensors, in the past few years, anything goes provided that a) it doesn't exceed the cost limits and b) it doesn't violate any other rules.

Subject to change for 2009.

Some sensors have been included in the KOP (camera, gear tooth, accelerometer, gyro, IR, light sensor, depending on the year).

[EHaskins]

The WPILib has driver code for the Hitechnic Compass sensor I've tested it, and it works well.

The KOP gyros are often useful, and I've used Maxbotix ultrasonic senors autonomous mode.

[Phalanx]

In the past we've used:

Quadrature Encoders with great success.
Maxbotix Sonic Range finders with minor success.
Gyro(300) from Sparkfun without success.
Compass with limited success (I can't remember which one).
Custom made IR Board with excellent success.
CMUCam with limited success.

We've also written all of our own code from scratch for these which is mostly responsible for the limited successes.

[DavidB]

Cool, thank you this is the kind of stuff I was looking for.

[DonRotolo]

Potentiometers we use a lot, often the mid-price ones ($4 each or so) from Mouser.
Same with limit switches, we pick ones with long levers on them we can bend to fit whatever our application, usually from DigiKey or Mouser.
The maxbotix ultrasonic range sensors are good, but we has poorer results from the very narrow beam type. We only use #1 now.
The gyro and accelerometers from the KoP are pretty good for many applications. Certainly the price is right. The gear tooth sensor, we have not had good success, but I personally do not blame the sensors, they work fine.
Haven't had a need to sense pressure, temperature, and many other physical parameters, but with exceptions most any sensor was legal last year.

I would approach this differently: What do i need to sense, and why, then find something in a catalog to specify how.

[keen101]

Sparkfun has a lot of cool sensors. So does parallax.

we've used:

CMU cam with moderate success
Parallax PING>>> ultrasonic sensor with success.
Limit switches
IR board
Potentiometers
1 wire temperature sensor (obtained from allelectronics)

all electronics
electronic goldmine
bg micro

are all places that sell surplus stuff, and usually have some cool sensors on the cheap.

oh, almost forgot..... last year we used line sensors as encoders for our wheels. We had to print out a circle that was split up into 8ths half dark half white, but it did work as wheel encoders to tell us relatively how far our robot had traveled. Feel free to use sensors in unexpected way's.

[Doug Leppard]

For 1902 we use.

In order of priority:

Optical Encoder - used in robot base to tell speed, direction and control robot mostly in autonomous mode. We have used it also for arm position.

Gyro – used in base in conjunction with encoder. Tell the direction.

Ultrasonic Sonic sensor – Matbotix is the best by far. Used to sense walls objects. Easy to use.

Thumbwheel switch – to set the autonomous mode. You can then dial what mode you want fro various versions of the autonomous mode.

100k pot - Used to position arm.

Limit switches – Various switches to test if at the limit of travel.

Camera – 2006 and 2007 used effectively. Of course 2009 new setup.

I would not use the compass. Mainly it is highly sensitive to metal around it. If another robot or metal in the arena comes close it can throw off the readings as much as 90 degrees. I walked by with a metal trash can and totally changed the readings. Nice in the laboratory but bad in reality.

Autonomous mode is ALL about sensors and using them effectively.

[dpeterson3]

I didn't get to use any sensors with the FRC bot last year (granted I tried, but my ideas were shot down). With other bots however,
Lego's light sensors (old RCX style) are very good for sensing different colors
Light encoders are very cheap and easy to build. (1 cds cell, 1 LED, and one old cereal box )
a compas would be very useful to tell the bot which alliance it is on
switches alone are very powerful for sensing a barrier or limit
pots are good for telling angles

I've never tried this one, but it makes sense. I have seen designes for using cheap mic cartagres as touch sensors. When the robot 'hears' the bumper move, it senses it hit something/

[DavidB]

So it seems that a large majority of the basic sensors that you will need come with the KoP. Thanks for the posts this will be very helpful.

[Ian Curtis]

We've used limit switches, and encoders to great success. We've had limited results with the 300 deg/s version of the '08 (and '07, I think) kit of parts gyro. We've also had some limited success with Sharp IR distance sensors and the MaxSonar distance sensor. Both the IR and MaxSonar sensors work great, it's just our implementations were not well tested enough. The Gyro on the other hand was, but for some reason we never figured it, when the OI was plugged into the Competition Port (i.e. the Robot was on the playing field) the Gyro never initialized and gave a completely bogus reading, causing our robot to drive across that nice white Lane Divider, conveniently scoring us a 10 point penalty.

But yeah. Limit Switches. They're awesome! Cheap, easy, and can save you from accidentally having major rebuilds. As I've posted in another thread, the Sharp IR sensors "miss" clear material (polycarbonate, glass, acrylic) unless the radiation beam (that just sounds nifty!) is perpendicular to the see through surface. Sharp IR sensors also can have problems with shiny multifaceted surfaces (aka, the aluminum diamond plate FIRST puts all over their fields). The MaxSonar sensors have none of these issues. The Gyro also works great, as long as it initializes (and our problem was pretty one-in-a-million). Just be sure the gyro can keep track of a robot that is spinning pretty rapidly (150 deg/s does not cut it).

[Gdeaver]

If you look at Mouser's web site for Cherry electrical products, you will find a non-contact absolute angular position sensor. They output an analog voltage from .5 to 4.5 volts thru 360 degrees. They make a great replacement for continuous rotation pots and could be used as a quadrature encoder replacement. Automotive grade and tough unit. Includes the magnet. Might be hard to find the corresponding connector. 26$ plus shipping.

[nathanww]

Just a hint about sensors from our last year:

Figure out what kind of sensors you want(and a general idea of where to put them) at an early stage in the process. This is important because often the sensors themselves can have quirks, or there can just be cases of PEBKAC leading to some truly bizzarre results. Ideally you should know the specs of every sensor you are using before you start writing code for any of them.

[Sparks333]

This is an odd one...

For the Aim High challenge, we built a rotating turret that could shoot balls in a wide range of directions. It required motors and stops for both the X and Y direction for the turret, and the initial design depended on mechanical limit switches. When we tried it in competition, the switch arm often bent out of shape, allowing the motors driving the turret to stall and pop breakers. The solution to this problem (which worked wonderfully, I might add) was to install reed switches and magnets at the extremes. Vibration-induced false positives really never became a problem, which surprised us all. Future iterations of the robot are supposed to use hall-effect sensors (a reed switches' modernized cousin), but they require power as well, so for a quick 'n' dirty fix, use a reed.

Sparks