Quote:
Originally Posted by AdamHeard
Exactly... definately a lesson learned there....
Another interesting tidbit I forgot to mention. Our original system w/ FPs wasn't geared low enough and was a pain to tune in PID because it was sooooo fast. The new system with the CIMs (not that the motors make a difference, there were other reasons for the change in motor type) was much slower and torqu-ier so we were able to tune a two jointed arm w/ straight P control in 15 minutes at a San Diego with even better accuracy the the original.
|
Power = force x distance / time
Think about the weight you are lifting. That's the force, The height you have to lift it to is the distance. Those are pretty straight forward.
What many teams forget is that it will take time to move from where you pick up an object to where you have to place it. Typically that time is 3-4 seconds. Give your arm operator something to do while the driver is moving from place to place. The joint of a single joint arm will typically move 90-120 degrees. Even if you want to pick up really fast, you can still design a joint that will take 9 seconds for a full rotation or 6.66 rpm. It will then take about 3 seconds to raise the arm to the correct height. The 05 and 07 BeachBot arms are running at more like 4 or 5 rpm and they are plenty fast enough. The arm almost always gets to "place" position before the rest of the robot.
Use a gear reduction such that your motor is running at 75% of free speed and your arm is moving at 4-6 rpm and you will get the most out of whichever motor you use.
The BeachBots have developed a very slick way of using the FP motors and a stock FP gearbox to power our arms. (actually I think we modified an idea we stole from Team 60) In 07 we started out using two FPs to power our arm as we did in 05. This set-up can and has lifted other robots silly enough to get entangled with our arm. But after a competition or two this year we decided we didn't need all that power in the arm as much as we needed the extra pound or two so we could have both ramps. So we removed one of the motors and its gearbox.
In 07, because the objects were so lightweight, virtually any motor in the kit could have driven an arm if it was properly geared.
As an example let's run the numbers:
We want to move a 0.5N object approximately 3 meters in 4 seconds. That will take 0.375 watts. The least powerful kit motor is the Mabuchi at 16W. The Mabuchi has a free speed of 4700 rpm. So you want it turning at 3500 rpm or 59 rps.
At this speed the output torque should be 25% of 61mNm or 0.015Nm. Assuming a 120 degree rotation for the final output, it should take 12 seconds for one revolution or 0.083 rps. Dividing the motor speed by the output speed, 59rps/0.083rps gives a reduction of 710:1. But that provides plenty of torque: 0.015Nm*710= 10.65Nm. If we assume a 2m arm, we only actually need 0.5N *2m = 1Nm.
That looks like a big reduction, but 20:1 is easily done using chain and sprockets and Banebots puts out a 56:1 gearbox that would work nicely. Using these together gets 1120:1 which should make the motor even happier.
There will be some losses in the drive train, but even with 50% loss there will be way more torque than needed. I puposely left out the power required to move the arm itself as that can be adjusted by using counter balances or springs.
So it really does not matter too much which motor you use as long as you use it correctly and keep all hardware within its operating limits.