Society of Robots - Robot Forum
Mechanics and Construction => Mechanics and Construction => Topic started by: calvino on May 31, 2011, 01:41:31 PM
-
Hi everyone, I'm trying to build an ultra fast following line robot, it will have differential steer and I think I will build it with wooden plates (like 200g), I will put this wheels:
http://www.dynamoelectronics.com/dynamo-tienda-virtual.html?page=shop.product_details&flypage=dynamo.tpl&product_id=232&category_id=106 (http://www.dynamoelectronics.com/dynamo-tienda-virtual.html?page=shop.product_details&flypage=dynamo.tpl&product_id=232&category_id=106) (100g max)
And this battery: http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=6518 (http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=6518) (170g)
And I will control it with this beautiful fpga: http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=139&No=593&PartNo=1 (http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=139&No=593&PartNo=1) (uhmm I'm not sure of its weight but it's really small, so max 100g)
And finally, the real problem is that I found this brushless motors, hobbyking says that they are for airplanes, but what it is the difference? there are others for cars, the real problem is that I don't know if my motor will fit my requirements, the total weight of the robot is 670g without motors, I can't use the formula that societyofrobots give because I dont know the torque, so what can I do?
I think I can use this http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=7338 (http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=7338) but I'm not sure if the motor can move the car, and that is all the problem, I want to know what brushless motor can I use to achieve a high velocity in a short time. Something like 10m/s in 1 second or something like that.
to control the motor I think I will use something like these one: http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=6457 (http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=6457)
Thanks very much, I hope I get understood, and sorry for my english. Any other specification you need, please let me know.
-
Hi,
And finally, the real problem is that I found this brushless motors, hobbyking says that they are for airplanes, but what it is the difference? there are others for cars, the real problem is that I don't know if my motor will fit my requirements,
Brushless motors are nice and powerful compared (by size) to brushed motors, but some of them are way too fast to be efficient. You need a very slow motor - a Kv of less than 1000 at least (I'd go for around 600 tops). The motor you found is very fast - up to ~25,000 RPM and will be inefficient with a gearbox.
Often you'll find the same motor size can be had in different speeds and the higher the speed, the lower the torque.
I want to know what brushless motor can I use to achieve a high velocity in a short time. Something like 10m/s in 1 second or something like that.
Brushed motors works very well, is easier to control and is cheaper, all things concerned (but heavier too).
10m/s^2 is a bit unrealistic for a robot (it's a higher acceleration than if you drop it from a tall building).
Provided your 'bot is supposed to do anything besides drag racing, you might need to adjust your expectation potentiometer a bit.
Robots aren't supposed to go 0 to 36km/h in 1s, but to do stuff along the way.
Assuming the 37mm wheels, they'd need to spin at 5162 RPM, to go 10m/s (and the motors needs to provide the necessary torque for moving the total weight of the fully loaded 'bot at that speed).
Have you given any thought as to how you will control/steer a differential drive at such speeds?
-
Thanks Soeren for the fast answer, and ok, you are right in than it is a bit unrealistic thinking in control 10m/s, I was thinking doing it with 6 optic sensors, if that doesn't work, I was thinking in use the camera of an optical mouse (that camera is like 18X18 pixels, so, you can think them as 18 small sensors), finally I was thinking in use neural nets to predict the next point based in a number of foregoing points (i'm not sure if this point is clear, but you can ask me) , so I can slow down before a curve, but all that are only ideas, the point is that I want to achieve a high velocity so I can prove them, even if doesn't work. If not, I can slow down my robot, like you said, adjust my expectation potentiometer, and prove other ideas :D
So, what motors I have to look for than has the necessary torque, at that speed?
Thank you very much :D
-
I think your limitation at high speed is less your ability to track the line, and more your ability to maintain traction between the wheels and the ground. 'High velocity' is entirely dependent on scale - if you use a track with a turn radius of 10cm, travelling at 10cm/s is essentially equivalent to 10m/s on a 10m turn radius for line-tracking purposes. Just use a normal-speed robot (up to 3m/s or so), and a track with smaller turns/runs. Scaling can be accomplished later - at the very least, this will avoid traction problems, and the (slight) danger of a 22mph block of metal flying into people when it loses the line.
-
One observation is that that battery might be overkill.... e.g. I've seen vids of those fast line following contests. You need very short battery life... thus weight that you can save. But I could be wrong, too (I only skimmed your post...)
In general, it's definitely a 'start slow' type of project. I'd be expecting a good amount of troubleshooting in order to perfect the line following. You could thus start with an imperfect motor.
-
I think your limitation at high speed is less your ability to track the line, and more your ability to maintain traction between the wheels and the ground. 'High velocity' is entirely dependent on scale - if you use a track with a turn radius of 10cm, travelling at 10cm/s is essentially equivalent to 10m/s on a 10m turn radius for line-tracking purposes. Just use a normal-speed robot (up to 3m/s or so), and a track with smaller turns/runs. Scaling can be accomplished later - at the very least, this will avoid traction problems, and the (slight) danger of a 22mph block of metal flying into people when it loses the line.
Yes, I guess I will have problems with traction, in fact, I had them a time ago in a competition, my robot followed the line perfectly outside of the race, but in the competition was a disaster because my algorithm try to adapt to the terrain after lots of trys, so, after all was perfect, the terrain was changed, so, you can imagine what happened. With the experience I also learned that the width of the robot should be of the minimum radius of the curve the robot has to take (I'm not sure if I'm being clear) and that will reduce the problems with the line following.
Later, in another competition, my robot was really fast, but when it reached a high speed to take a curve, the wheels slipped (I think that this is about what you are talking), so, I was thinking to solve this with an accelerometer and an encoder for the motors. The fpga I showed you has an accelerometer (that board is beautiful ::)) so, I can know the position and the orientation with both, I mean, I will have two sources that will show me if the car is slipping (the encoders will show me the speed of the motors so, I can integrate it to know displacement, even if the robot is in the air, I mean, is not moving, the encoders will show me the speed. With the accelerometer, I can know the same integrating two times and this will show me the real movement of the car, so, if they are different I guess there will be slipping, I'm not sure, I have to test out).
And due to your advice, I was thinking to paste two wheels, so, the traction area will be bigger
Thanks for the answer
Edit: I forgot to finish the idea about the accelerometer and encoder issue, so, all that is to detect if the car is slipping, if the robot detect is slipping I was thinking to do something like turn on the two motors so it will go forward in the correct orientation (of course, this is a vague idea, but it is an idea)
-
One observation is that that battery might be overkill.... e.g. I've seen vids of those fast line following contests. You need very short battery life... thus weight that you can save. But I could be wrong, too (I only skimmed your post...)
In general, it's definitely a 'start slow' type of project. I'd be expecting a good amount of troubleshooting in order to perfect the line following. You could thus start with an imperfect motor.
What do you mean with imperfect motor?
Sincerely I never have utilized those batteries but the specifications say that it can give 33Amps of constant current, if it explode and I am alive to tell it, I assure you that you will know it ;)
You are right, there will be a lot of troubleshooting, not only with the line following but with all the mechanics issues.
I really want to build this project, in the worst of the cases, it won't work, uhmm, and maybe I lose a hand or a finger, but I don't think so :)