Motor for arm: what should be used when?

[Ether]

Originally Posted by clayman

Ballpark, I'd say 40 ft-ibs.

I assume that's 40 ft-lbs before you balance it with springs :-)

Just for fun, let's do a little engineering and see what happens at 40 ft-lbs
and why the gear ratio is so important.

Let's say you want to move the arm at 90 degrees per second. That's 15 rpm.

You said your total gear ratio is 12.75*2:1 = 25.5:1.

So your speed at the motor is 15*25.5 = 383 rpm.

Your torque at the motor is (40/25.5)*12*16 = 301 ozin.

Using the motor curves for the CIM and a bit of algebra, 11.4 volts will give you 301 rpm at 383 ozin. The motor efficiency at this operating point will be 6.4% and you'll be pulling 117 amps (until your circuit breaker trips).


Now let's see what happens if you change the gear ratio to 250:1

The rpm at the motor is now 3750 and the torque at the motor is 31 ozin.

Using the CIM motor curves again, we see that 9.6 volts will produce 31 ozin of torque at 3750 rpm. The motor efficiency will now be 65% and you'll be drawing only 14 amps.


Of course, the problems with high gear ratios are

- some loss of efficiency (which I haven't accounted for here)

- weight, size, expense

- problems with overloading the gears if they are too small


But there are problems with using balancing springs instead of gear-down:

- you can't balance over the entire operating range

- if you don't gear down enough, you lose fine control of the arm


So, no surprise, as with everything in engineering it's a matter of finding the right compromise to meet the task at hand.

[HarveyAce]

We have two window motors powering the upper half or our arm. We had the problem of the motors not being able to drive the arm past parallel with the ground due to max torque on the arm due to gravity, which was stalling or window motors. We have two strands of surgical tubing as a counter-balance. An estimation of the torque load would be about 31.5 ft-lbs. (3.5ft*approx 9 lbs weight). Our windows get warm but don't overheat (they did only once due to excessive testing in a short time span). we are using 2 vics to control the motors, using an analog control (a second Attack 3 joystick). this way we can have a little more control of the motor going down (lower voltage), but max voltage going up (it won't move past 90 degrees even with the tubing without the voltage ). besides adding more tubing, what could we do to help get more control and use less voltage? Should we change to an FP with a speed controller? if so, how do we get it to not back-drive?

[JamesCH95]

We've got two RS775s running through a a ~380:1 reduction on our arm (64:1 gearbox, 6:1 sprocket). The arm has two window motors, two pistons, some 25 series chain, and a pretty serious manipulator. We simply stall the motors to hold position controlled through a PID loop, and they don't heat up much at all. They've never been more than "warm to the touch" even after minutes of holding position at horizontal. They lift the arm nice and fast, the rate is controlled in software, and using the jags' dynamic braking it coasts down nicely as well.



Two RS775s might be overkill, but they handle the stalling a lot better than any other BaneBots motors and have proven to be wonderfully robust so far.

[Ether]

Originally Posted by JamesCH95

We've got two RS775s running through a a ~380:1 reduction on our arm (64:1 gearbox, 6:1 sprocket).

Can you give a ballpark estimate of the torque load on the arm?

(I use an inexpensive fish scale - available in the sporting goods section of any discount store - and a tape measure to get an estimate)

[JamesCH95]

The robot has been shipped to another regional, but I can make a reasonable guess of around 25lb-ft

[Ether]

Originally Posted by JamesCH95

The robot has been shipped to another regional, but I can make a reasonable guess of around 25lb-ft

25 ft-lb = 4800 ozin at arm => (4800/380)/2 = 6.32 ozin at each motor

From the motor curves for RS775:

RS775 @ 12.0 volts:
   oz-in      Nm     rpm    rpm%    amps   watts out    watts in    eff%
     6.3   0.045   12272    94.4     6.6        57.3        78.6    72.9

12,272 rpm at motor = 194 degrees/sec at the arm.

A bit fast for the top end, but motors running nice and cool.



Throttle it down to 90 degrees/sec by reducing the motor voltage:

90 degrees/sec = 15 rpm at arm => 15*380 = 5700 rpm at motor

RS775 @  5.9 volts:
   oz-in      Nm     rpm    rpm%    amps   watts out    watts in    eff%
     6.3   0.045    5700    88.7     5.6        26.6        33.5    79.5

Still running cool, and the voltage is in a reasonable range for control.


How about stall?

RS775 @  0.7 volts:
   oz-in      Nm     rpm    rpm%    amps   watts out    watts in    eff%
     6.3   0.045       0     0.0     4.9         0.0         3.3     0.0

Very reasonable 3.3 watts dissipation.

[JamesCH95]

Very close to my "napkin" design

I aimed for about 180deg/sec at full speed, but couldn't quite get there with the finite available sprocket ratios.

We limit the 775s to 60% of full command speed, and the PID loop still does a great job of controlling them.

I hadn't done a stall calculation, we had been planning on using a DeWalt transmission with anti-back-drive pins, but found that the pins didn't like it when the arm was powered down, so we removed them. This also has the side benefit of being able to move the arm without powering up the robot or disassembling anything.

[Ether]

Originally Posted by JamesCH95

Very close to my "napkin" design

I aimed for about 180deg/sec at full speed, but couldn't quite get there with the finite available sprocket ratios.

I didn't account for gearbox inefficiency and manufacturing tolerances, so the numbers aren't anywhere as exact as the displayed values might suggest. So it might actually be closer to your "napkin" 180. I left the displayed values as-is in case anyone wanted to play around with the calculations and compare them.

[JamesCH95]

Originally Posted by HarveyAce

We have two window motors powering the upper half or our arm. We had the problem of the motors not being able to drive the arm past parallel with the ground due to max torque on the arm due to gravity, which was stalling or window motors. We have two strands of surgical tubing as a counter-balance. An estimation of the torque load would be about 31.5 ft-lbs. (3.5ft*approx 9 lbs weight). Our windows get warm but don't overheat (they did only once due to excessive testing in a short time span). we are using 2 vics to control the motors, using an analog control (a second Attack 3 joystick). this way we can have a little more control of the motor going down (lower voltage), but max voltage going up (it won't move past 90 degrees even with the tubing without the voltage ). besides adding more tubing, what could we do to help get more control and use less voltage? Should we change to an FP with a speed controller? if so, how do we get it to not back-drive?

I would get a stronger motor (RS775 is awesome, as I've mentioned) and gear it way down. You might considering gearing the window motors down even further.

[clayman]

Originally Posted by Ether

Just for fun, let's do a little engineering and see what happens at 40 ft-lbs
and why the gear ratio is so important...

Thank you so much! It looks like we were having a problem with the arm hitting the stall amperage. Tomorrow I will change the ratio to 51:1, much more reasonable (and ideally below the max power rating). It should supplement our spring counterbalance nicely.