Society of Robots - Robot Forum
Electronics => Electronics => Topic started by: Tsukubadaisei on September 07, 2007, 05:15:57 AM
-
It is the first time I am using continuous rotation servos in wheels(I have already used them for other things though), and something is bugging me a lot. I have built a robot this July using double stearing. You can see pictures of it in http://www.societyofrobots.com/robotforum/index.php?topic=1551.msg10458#msg10458 (http://www.societyofrobots.com/robotforum/index.php?topic=1551.msg10458#msg10458).
I have been looking at ADMIN's Taurus2 and the structure is similar. And I think I noticed that Taurus 2 doesn't use rotary encoders. Following my original design I installed simple homemade encoders in my Hanako but they are driving me crasy (every time I used wheels I used stepper motors so I am a n00b when using encoders for wheels).
So I would like to know what you guys think. Should I use encoders or not? What kind of precision do the modified servos have? I know that when you mod a servo you lose angle control but you get speed control, but is the angle control important? Should I use encoders in order to recover it? I will be doing a lot of line tracing and object detection by the way.
-
If you want precise speed/acceleration/position control or odometry based map position you will need encoders.
-
Just as I thought. Thanks.
-
My Taurus2 worked perfectly fine without encoders . . . if you dont need precision dont use them . . .
-
Yes. The linetracing routine was easy to program and it is working great without encoders. Now I am trying to use only sonars and object detection in order to follow the second half of the contest (without lines but with objects). Since I have 4 sonars and 2 of them are attached to servos (wide area) I think I can do it. So far I have got better results than when using encoders. Anyway, thanks admin for confirming my expectations.
-
I have another question: ADMIN, what did you do to solve the energy problem for Taurus2? Today I tested my robot for the first time using internal batteries, it worked great... for 4 min :'(. I need about 15 min (actually only 10, but just in case...). By the way, I am using 11 servos(4 for the wheels) and used 5 x 1.2V 2000mAh Ni-MH batteries connected in series(6V). I should have guessed that it is not enough but I have no idea what to do.
So far I used only 12V steppers and I used to connect 12V batteries in each one. Or when I built a humanoid I used 12V servos so I had no problems.
If anyone else can help, please be my guest.
-
I cant remember the mAh on the batteries I used, but Im willing to bet it was 2000mAh. I used two of these 7.2V NiCad batteries attached in parallel.
Other than adding extra batteries, you can also try being more mindful of power draw in your actuators:
- dont send PWM commands to servos that arent needed
- measure power draw on each of your servos, determine which are draining too much, and try to redesign to fix it
I dont remember ever having power problems on it. It would last 30+ minutes on one charge . . . The servos I used were the cheap HS-311's . . .
-
You might consider switching to Li-poly batteries...
More expensive but more mAh per pound (kg?)
Here is a good site to buy them: http://www.all-battery.com/index.asp?PageAction=VIEWCATS&Category=258
-
Thanks ADMIN, I talked with some teachers and colleages and it seems that is the solution after all. I am using some cheap FUTABA servos as well, but I was worried about increasing the Voltage to 7.2 and the current 2 times. I appreciate the quick answer.
And thanks Jesse, that was what I was considering in the beggining but the set up I am using now has more charge (2000mAhx5=10000mAh) then the batteries in you link. Also I dont know about the maximum current and I have heard that those batteries are explode easier. And even if they are better it will not worth replacing my collection of 20 1.2V NiMH batteries, I am saving money. But thanks anyway.