Macroware Technology

Welcome Portlet


Welcome to Macroware Technology Blog.

The purposes of this engineering blog:

- Share knowledge
- Learn
- Have fun
- Document what I was doing / thinking

Search Box

 

Private Area

Mailing List

Engineering Education

Engineering Organizations

Professional Career Networking Groups

Electronics Hobby

Trade Publications

Dirt Dog Needs a Bath

posted Saturday, 23 December 2006
My dirt dog got to the point where it wouldn't stop acting erratically and no amount of blowing on the sensors or vacuuming it would help.  So I got my air compressor out as the manual stated and thoroughly cleaned the whole thing off with that and a vacuum.   After that it worked fine again.  It went through the main room of the addition again and picked up a full canister of dust and dirt.  I emptied out the canister and sent it back to work again, but shortly after restarting it, it was acting erratically again.  So I went and cleaned it again, and then it worked fine.  I went through about three cycles of this, with slightly less in the canister each time, but then the battery was drained, so now it is charging.  It seems like from my experience, it can go through a very messy environment once and then it needs a thorough cleaning.  This time the room still had a layer of drywall dust (I don't think we're ever going to get all of that cleaned out) but also quite a bit of sawdust from the finish carpenter.  Again, the dirt dog did a pretty good job of cleaning up the room, although I think it might need a couple of more passes.  It's just that the sensors are not robust enough to deal with a lot of dust.  One other thing I did try this time was putting black electrical tape over the cliff sensor openings, and that didn't seem to help much with the dust problem, dust was still able to get to a place where it could affect the sensors.

I think the cliff sensors use light to determine if there is no floor under the front of the robot, although I'm not 100% sure of that.  Maybe there are some other sensors in the robot that are getting clogged up too that my thorough cleaning is taking care of but I'm not aware of exactly where they are?  Anyhow, my general engineering experience is that optical light based sensors usually aren't all that robust, and are prone to being affected by non ideal conditions.  In many applications this is okay, however I think it was a design mistake to use optical sensors for the cliff sensor in a robot that is going to be subjected to a very dusty and dirty environment.

So far the amount of time the dirt dog is saving me by cleaning up the room is being offset by the amount of time I'm spending cleaning the dirt dog.  Now in the future when I'm using it mostly in my workshop, it probably won't ever be quite as dirty as a construction site, and especially won't have all the drywall dust, so maybe it will work out okay, but so far I'm a little disappointed.