On it!...
due to time constraints, I did not make all the LED's flash...But the "new" encoders work great!
[video=youtube_share;W_5Rrwx9qmE]
Whatcha think?...needs work eh?...:blank:
After much pestering, I'm ready to add my first video to this thread.
I haven't added any distance computation algorithms to my encoder program yet so this attempt just uses speed and time in an attempt to get the robot to travel in a figure 8.
I have the program monitor the landing gear channel for any significant change (indicating the switch as been thrown). Once the switch is thrown the robot will change mode from manual control to autonomous.
I edited this video without listening to the sound so sorry if I chopped off any words at the beginning or end.
Even though the figure 8 was kind of butchered, I still think it's fun to see those mecanum wheels move the robot. I'll likely give this another attempt tomorrow.
Awesome Duane, You have got the Mecum machine rolling sweet now..
Figure 8's are not as easy as they look, alot harder to do than a square.
I am working on a new Magnamatic Encounterator System for this challenge,
Trying to get it done fast, as I am commiting to going to the UPEW this year,(gonna get me one of those mini S2's)
and I want to bring this figure 8 bot, and hopefully a line following robot.. oh, and the "LightShow" robot too.(some assembly required...)
It's first infant steps without a tether.... In honor of Erco and his challange. Two problems, one I don't understand yet and one that is obvious. The obvious one is the fact it computes speed in steps from 0 to 127 which comes out to 4 RPM at the motor. That's not good enough to get the speeds to match up correctly so one motor stops before the other. Two is why turning to the left causes it to miss the mark by over 40 degrees. The controller is spitting back the same numbers just reversed but somehow that is not getting either wheel to the correct placement.
The guts is a simple calculation of circumfrence. You tell it how many degrees of arc and what radius you want the inside wheel to follow and it calculates the rest. You tell it what speed the fastest wheel should move and it calculates the other based on the ratio of the two distances.
Even if you ran over the plate and a few kittens and small children your's is much cooler to watch!
You shouldn't 'need' encoders with steppers true but that does leave you with an open loop. I said to move the wheel this fast and for this long but I have no idea if it moved at all or stopped early or what-have you. The belt drive introduces some small errors that are difficult to correct for so the difference between what I asked for and what I get increases over more distance. If I just ran the wheels direct off the motor shaft I could probably do better but then a grain of rice or a stiff breeze might stall it out.
Yeah, I hate when there's something weird that I can't account for. The need to add 40 degrees to a left turn to get a full circle has got me stumped. The same basic concept is used to do a spin turn and it executes those nearly perfectly. (shrug)
Yup, that's a QuickStart board alright though I'm not yet using all those nifty buttons and LEDs. The other board is a ChipKit UNO32. It does all the heavy lifting as far as step generation and figuring out how far to move ect.
Not in y'alls class here, but I had to try this with Sid, my blue Scribbler.
First off, getting Sid to run in a straight line is a challenge. the motor speeds seem to vary from one run to the next. Still... I sat down with a cold adult beverage and the Big Ten playoffs for background noise. Took a couple of hours, but I got a decent, kinda repeatable circle... using the GUI interface. Copied that for the second circle, did some tweaking to compensate for the difference in the motors and tried the figure-eight.
Not =too bad=... I managed to end up about four inches from the starting point. But I suspect that will change on the next run. If I can get some measure of repeatability, I'll post a video.
Back to the drawing board!
Amanda
Edit: Sure enough... after a (visually) perfect first circle on the first try, the second yielded a first circle of about 370 degrees skewing the "8" a bit. Still, the starting point fell under the 'bot. Now I'll be up the rest of the night twiddling with it. :-)
Come on Amanda, enough talk. Post that video! First person to Youtube the orig blue Scibbler doing a Fig-8 wins. Starting point falling under the 'bot's final position is plenty good.
Well, here it is. The video is pretty dark... took it when the sun was way over in the west. If you look just where Sid starts moving, there is a quarter on the carpet... hard to see tho'.
Man! Was I late to this party or what?!?! Great job every one. I will challenge ScribblerKart and the Gatorbots next Thursday and we will try to post a video...
Can't believe nobody put a Sharpie in one of those S2s or old Blues...
Well, here it is. The video is pretty dark... took it when the sun was way over in the west. If you look just where Sid starts moving, there is a quarter on the carpet... hard to see tho'.
Go! Amanda Go!,
Hey, maybe you could add the dinner plates for Sid's next video, They help add to the excitement of a figure 8...
@Whit,
I bet we will be stunned by the Whit/Scribblerkart team effort, it's a shame you had to give up the weekend though.
Oh, and don't forget the dinner plates, Not sure it's official without the plates...
Here is one side of the Magnamatic Encounterator Sub System, or MESS for short,
I have plenty to clean up before it's good enough though, Just thought I would share this little bit of progress...
[video=youtube_share;74AzrJ5e1F0]
Magnets and Reed switches sure can be fussy...
"Magnets and Reed switches sure can be fussy..." Ttailspin
I wasn't gonna say anything, but they are not real precise. I used reed switches reed switches on Wastebot and it was quite difficult to get any consistency with them. http://www.youtube.com/watch?v=Ie1AWV3-yS4
BoeBot Figure 8 couldn't wait! This is Ken's BoeBot I just assembled. I have to add the flamethrower next, but I figured I should test the chassis first. This is actually the first BoeBot I have built or programmed...
My Fig 8 second loop could have been a tad smaller for symmetry, but I did all this in about 5 minutes and figured it was good enough for a first pass. I'll leave room for improvement so somebody else can have some fun!
It hooked a bit when it stopped at the end. Shoulda added some pulsout 750's for dynamic braking at the end!
' {$STAMP BS2}
' {$PBASIC 2.5}
PAUSE 1000
FOR W0=1 TO 435
PULSOUT 13, 790 ' left
PULSOUT 12,727 'right
PAUSE 20
NEXT
FOR W0=1 TO 360
PULSOUT 13, 777 ' left
PULSOUT 12,680 'right
PAUSE 20
NEXT
After a day of playing with Sid, Bob (my Boe Bot) was looking a bit ignored, so I decided to give him a shot at the figure eight. After some discussion, the crew thought the dinner plates were "so yesterday" and they wanted a couple of observers on the course to make sure everything was on the up and up. ;-) Under the pivot of the Ping servo is a black quarter for my reference point.
I found out the smallest change in the starting position had a big change in the finish.
Repeatability is the biggest problem I've had... I don't think carpet is the best place to try this!
Comments
due to time constraints, I did not make all the LED's flash...But the "new" encoders work great!
[video=youtube_share;W_5Rrwx9qmE]
Whatcha think?...needs work eh?...:blank:
-Tommy
I haven't added any distance computation algorithms to my encoder program yet so this attempt just uses speed and time in an attempt to get the robot to travel in a figure 8.
I have the program monitor the landing gear channel for any significant change (indicating the switch as been thrown). Once the switch is thrown the robot will change mode from manual control to autonomous.
I edited this video without listening to the sound so sorry if I chopped off any words at the beginning or end.
Even though the figure 8 was kind of butchered, I still think it's fun to see those mecanum wheels move the robot. I'll likely give this another attempt tomorrow.
Tommy, I think I've seen those new quad encoders in the Parallax shop.
DY-NO-MITE!!! Beautiful job, stop apologizing already! Everything is a work in progress.
http://atom.smasher.org/construction/?l1=Robot&l2=Figure+8+&l3=in+progress&l4=Expect+Delay
Figure 8's are not as easy as they look, alot harder to do than a square.
I am working on a new Magnamatic Encounterator System for this challenge,
Trying to get it done fast, as I am commiting to going to the UPEW this year,(gonna get me one of those mini S2's)
and I want to bring this figure 8 bot, and hopefully a line following robot.. oh, and the "LightShow" robot too.(some assembly required...)
-Tommy
It's first infant steps without a tether.... In honor of Erco and his challange. Two problems, one I don't understand yet and one that is obvious. The obvious one is the fact it computes speed in steps from 0 to 127 which comes out to 4 RPM at the motor. That's not good enough to get the speeds to match up correctly so one motor stops before the other. Two is why turning to the left causes it to miss the mark by over 40 degrees. The controller is spitting back the same numbers just reversed but somehow that is not getting either wheel to the correct placement.
The guts is a simple calculation of circumfrence. You tell it how many degrees of arc and what radius you want the inside wheel to follow and it calculates the rest. You tell it what speed the fastest wheel should move and it calculates the other based on the ratio of the two distances.
Everything is a work in progress photomankc.
But I know how you feel. You feel it should work better.
With steppers you shouldn't need encoders, right? That's a QuickStart board right?
Great video. Thanks for posting.
Even if you ran over the plate and a few kittens and small children your's is much cooler to watch!
You shouldn't 'need' encoders with steppers true but that does leave you with an open loop. I said to move the wheel this fast and for this long but I have no idea if it moved at all or stopped early or what-have you. The belt drive introduces some small errors that are difficult to correct for so the difference between what I asked for and what I get increases over more distance. If I just ran the wheels direct off the motor shaft I could probably do better but then a grain of rice or a stiff breeze might stall it out.
Yeah, I hate when there's something weird that I can't account for. The need to add 40 degrees to a left turn to get a full circle has got me stumped. The same basic concept is used to do a spin turn and it executes those nearly perfectly. (shrug)
Yup, that's a QuickStart board alright though I'm not yet using all those nifty buttons and LEDs. The other board is a ChipKit UNO32. It does all the heavy lifting as far as step generation and figuring out how far to move ect.
I have most of the mechanicals in place on this "SpotBot"...
It will be doing figure 8's in no time,
I just have to perfect the Magnamatic Encounerator System before I turn it loose around the plates..
-Tommy
No last minute 'hook' to the slow side now. They stop pretty close to the same moment.
First off, getting Sid to run in a straight line is a challenge. the motor speeds seem to vary from one run to the next. Still... I sat down with a cold adult beverage and the Big Ten playoffs for background noise. Took a couple of hours, but I got a decent, kinda repeatable circle... using the GUI interface. Copied that for the second circle, did some tweaking to compensate for the difference in the motors and tried the figure-eight.
Not =too bad=... I managed to end up about four inches from the starting point. But I suspect that will change on the next run. If I can get some measure of repeatability, I'll post a video.
Back to the drawing board!
Amanda
Edit: Sure enough... after a (visually) perfect first circle on the first try, the second yielded a first circle of about 370 degrees skewing the "8" a bit. Still, the starting point fell under the 'bot. Now I'll be up the rest of the night twiddling with it. :-)
Come on Amanda, enough talk. Post that video! First person to Youtube the orig blue Scibbler doing a Fig-8 wins. Starting point falling under the 'bot's final position is plenty good.
I'll get to it today if you don't! Now git!
http://youtu.be/PV_D-bie3ZE
Pretty rough and I'm working on it!
Try again tomorrow when the light is better. :-|
Now who will get a BoeBot around first? Will it be me?
MATT G? WHIT? There's only one "first". This is low hangin' fruit!
Doggiedoc: Toddler!
Humanoido: Penguin!
idbruce: Whaddayagot?
Everyone: Run whatcha brung!
Bonus points for first wooden robot!
Time's a-wasting!
Can't believe nobody put a Sharpie in one of those S2s or old Blues...
The train is leaving. Are you on it or under it?
I'm sure you & Scribblerkart will do the right thing this weekend!
Very cool. Love the fluid move from one circle to the other.
Hey, maybe you could add the dinner plates for Sid's next video, They help add to the excitement of a figure 8...
@Whit,
I bet we will be stunned by the Whit/Scribblerkart team effort, it's a shame you had to give up the weekend though.
Oh, and don't forget the dinner plates, Not sure it's official without the plates...
Here is one side of the Magnamatic Encounterator Sub System, or MESS for short,
I have plenty to clean up before it's good enough though, Just thought I would share this little bit of progress...
[video=youtube_share;74AzrJ5e1F0]
Magnets and Reed switches sure can be fussy...
-Tommy
I wasn't gonna say anything, but they are not real precise. I used reed switches reed switches on Wastebot and it was quite difficult to get any consistency with them. http://www.youtube.com/watch?v=Ie1AWV3-yS4
My Fig 8 second loop could have been a tad smaller for symmetry, but I did all this in about 5 minutes and figured it was good enough for a first pass. I'll leave room for improvement so somebody else can have some fun!
It hooked a bit when it stopped at the end. Shoulda added some pulsout 750's for dynamic braking at the end!
' {$STAMP BS2} ' {$PBASIC 2.5} PAUSE 1000 FOR W0=1 TO 435 PULSOUT 13, 790 ' left PULSOUT 12,727 'right PAUSE 20 NEXT FOR W0=1 TO 360 PULSOUT 13, 777 ' left PULSOUT 12,680 'right PAUSE 20 NEXThttp://youtu.be/YXit5lZTQZ8
I found out the smallest change in the starting position had a big change in the finish.
Repeatability is the biggest problem I've had... I don't think carpet is the best place to try this!
Amanda
Not a single Eddie robot entry yet. Are Eddie, Kinect and MSRS up to the Figure 8 challenge?