Sensor Recommendations

[Workaphobia]

I'm a bit embarrassed to post this without extensively searching and researching the matter, but I'm low on time and the next couple of days are critical in my (rookie) team's design stage.

I'm the lead programmer. I essentially gave up on trying to comprehend all the drafts and discussion regarding the arm mechanism for our robot, because it kept changing so many times and we didn't seem to be getting anywhere. I'm now fully willing to butt out and let my teammates commit to whichever idea they will, but I'm wondering how critical it is that I tell them early on exactly what sensors I'll need and where. Is a potentiometer something that can't easily be added to an arm after it is already built? Does a gyro need a special mounting place reserved in the chassis? What if in two weeks I suddenly realize that my job would be vastly simplified if I only had an extra input device somewhere - am I doomed if I don't have the foresight to bring it up now?

I was thinking of at least a few touch sensors for collision and limit switches wherever there are moving parts. Is it worth it to add a potentiometer so I can also get the angle of a variable arm, or is that not such a great idea for a rookie team?

I'll admit that I haven't been as active in reading all of the manuals and white pages as I would have liked, so if you just want to throw me a link or two, I'll understand.


Thanks in advance.

[Kevin Watson]

Originally Posted by Workaphobia

I'm a bit embarrassed to post this without extensively searching and researching the matter, but I'm low on time and the next couple of days are critical in my (rookie) team's design stage.

I'm the lead programmer. I essentially gave up on trying to comprehend all the drafts and discussion regarding the arm mechanism for our robot, because it kept changing so many times and we didn't seem to be getting anywhere. I'm now fully willing to butt out and let my teammates commit to whichever idea they will, but I'm wondering how critical it is that I tell them early on exactly what sensors I'll need and where. Is a potentiometer something that can't easily be added to an arm after it is already built? Does a gyro need a special mounting place reserved in the chassis? What if in two weeks I suddenly realize that my job would be vastly simplified if I only had an extra input device somewhere - am I doomed if I don't have the foresight to bring it up now?

I was thinking of at least a few touch sensors for collision and limit switches wherever there are moving parts. Is it worth it to add a potentiometer so I can also get the angle of a variable arm, or is that not such a great idea for a rookie team?

I'll admit that I haven't been as active in reading all of the manuals and white pages as I would have liked, so if you just want to throw me a link or two, I'll understand.


Thanks in advance.

It really sucks when this happens. You have at least two options that come to mind. Most importantly, you need a systems person(s) who can see the "big picture". This person(s) is responsible for making sure things like sensor mounts are in the design before the mechanical design team makes some assumption that will prevent you from mounting your sensor on the finished 'bot. If you can't make this happen, you gotta get in there and make people listen to you, so you'll get what you need. Sometimes engineering is a contact sport <grin>.

-Kevin

[russell]

Potentiometers are good, and (at least if your robot turns out to be kind of a roughly built machine) you can generally find a way to attach them after the build process, though if you can incorporate them into the design now then that is even better. Just remember, when you have to sort of calibrate them every so often, especiallyif you are working on the robot allot. I would say more but my spacebar is starting to act up so I will just leave it at that.

[jgannon]

During the off-season, I added a potentiometer to last year's robot as a coding exercise in preparation for this year, so it can be done after the robot is built. At the very least, request that your team leave at least one end of every shaft on your robot exposed. It's not too difficult to build a little mount, and then you can couple an encoder or potentiometer to whichever shafts you decide you'd like to measure. We used plexiglass for the mount, and fuel line to couple the shafts. It's not the neatest way of doing things, but it will work, and it allows you to focus on more important tasks in the mean time.

[phrontist]

Adding a pot is, IMHO, worth the minimal extra effort.

[jgannon]

Originally Posted by phrontist

Adding a pot is, IMHO, worth the minimal extra effort.

Oh, sure it is. The motto around here is, "If it moves, we measure it." The issue was whether or not one can be added after the robot is built, which it most certainly can, if the team takes some minimal precautions. Best of luck, Workaphobia.

[Squall]

I may be wrong here, but if you're thinking about using a Gyro you should consider somewhere safe and far from vibrations, and still near your robot spin center.

The Gyro is very sensitive, so you must protect it as much as you can so you won't get lot's of wrong values from it.

[Mark McLeod]

The mechanical design team needs to make allowances now for the mounting of sensors. Better to have the sensors designed in and not used then not having it when you need it or attached as an after thought and hanging out for a falling tetra or another robot to violently remove.

[Bentow]

I too find myself as a lead programmer on a rookie team. Are there any other things I need to look out for besides ensure spots to mount sensors. Are there are absolutely critical readings I need, that I would otherwise miss until the last minute?

Also OT: who makes the best drivers, especially on a small team, SW guys, Mech guys? We haven't gotten anywhere with picking them yet.

[jzampier]

Having been a programmer/electronics person for the last few seasons...
here's my notes:

1. Don't 'butt out'. It's bad then the mech folks ignore your requests and then complain when the bot doesn't do anything useful in auton mode and various other places. Get the sensors you want worked into the design early. Shaft encoders, pots and other stuff. I recommend buying good pots that are nicely sealed.

2. Get a breakdown of what motors and assemblies are being used and lay it out in a spreadsheet to assign inputs and outputs on the RC. This makes wiring easier, and neater. I can send you some old breakdowns if you want examples.

As for the 'best driver'... well, I've always been a fan of holding driver trials with old robots. If you don't have an old robot, try with a new one. Let anyone who wants try and rotate.