Pros and Cons of Dual Motor Elevators

Hey CD!

Recently, I have been doing a lot of research in elevators as the variety of them and the aspects that separate the good from the best fascinates me. Feel free to look at this other thread I made while you are at it.

Anyway, I am wondering what people think of dual (or tri) motor elevators. Whether this is in gear boxes or some of the form that creates the motion in an elevator system, cascade or continuous, I’d love to hear your thoughts on why these multiple-motor elevators are outstanding, normal/decent, bad, or somewhere in the middle.

Thanks in advance!

1 Like

Depends on the application.

You can get any amount of torque out of a single motor, but the variable will be final output speed.

You can get any output speed you want, but at the cost of torque and current draw.

How much do you want to lift? A game piece? A robot? Three robots?
How fast do you want to lift it?

These are the questions to consider for your system before deciding on # of motors. The more you want to lift and the faster you want to lift it, the more motors you’re going to need.

JVN’s design calculator can help you figure these things out so that you’re not scrambling to debug current draw and torque issues last second.

5 Likes

Thanks for the utility! Super useful at first glance!

One of the biggest issues I’ll see teams run into when designing an elevator with multiple motors, is not mechanically linking each motor together. By this I mean, don’t have a left side motor and a right side motor that aren’t connected together. A good example of this would be a rack and pinion drive on each side of the elevator, each individually powered by its own seperate motor. While in theory it sounds easy to just tell the motors to “go the same speed,” in reality it is much much more challenging. More often than not, you will run into the issue of your elevator racking, or rather, one side moving at a different speed than the other. Even a small difference in speeds can cause huge issues.

The solution to this can be one or several things. You could put all the motors into one big gearbox, you can have two seperate motors/gearboxes but with a shaft connecting the two (watch out for how much your shaft can twist), you can go to only one motor (at the cost of either speed or torque).

5 Likes

For the past two seasons, we have ran dual motor elevators. In 2018, we started with each one in it’s own VP, and just connected at each end of the shaft using chain, partially for redundancy, and partially for the fear of the torque of two 775s accelerating the elevator rapidly under load would destroy the small gears. To save weight, we ended up putting them on the same side, and used a dual motor input into a VP, and had no issues.

Since we had good luck with 2 775s in one VP, we did the same thing again (but used Vex’s new dual motor input they came out with), and even used the output shaft of the VP to mount the sprocket that would run the main chain up and down, and never had any issues.

Could we have used 1 775? Yes. But crazy fast speeds are definetly a lot harder. Having nearly a perfectly balanced elevator, and 2- 775s, the elevator was faster than we could control almost. In the end, I think we got pretty close to .5 seconds from the bottom to top.

This wasn’t full speed I dont believe, but only video I have that shows it moving pretty fast. Teams have done it, but I know we personally would have struggled to make it that fast with 1 motor.

1 Like

^^ This. While it was a double-reverse four bar rather than an elevator, 3468 had exactly this issue this year. They managed to improve things from competition to post-season by slaving one to the position of the other, but it still wasn’t quite right. If you aren’t going to mechanically link the two, motion profiling or some sort of software control which adjusts at least one of the motors to account for the differences is a must.

@gixxy, want to chime in here with details?

@GeeTwo, We came up with the solution during Build Season, and were using it at competition, not in the Post-Season.

As for the details, 3468 did a double-reverse four-bar lift, with a 12" DART Linear Actuator for each side. While we did have plans for trying to link them together mechanically, we ran out of time first, so we had to fix it in Software.

The solution we went with was setting up each side as a P(ID) Loop, using the included potentiometers as absolute position feedback devices. The First PID Loop’s setpoint would be set to our actual target, and the other PID Loop would just constantly update its own setpoint to the first PID Loop’s current position. You can see our code here: https://github.com/frc3468/DeepSpace/blob/master/src/main/java/frc/robot/subsystems/Lift.java and you can see a (not realtime…) video of the lift in action here: https://www.instagram.com/p/BuF48udnSDZ/

This wasn’t the most optimal solution, and we did have to slow down the speed of our motors a bit (mostly to give us some more reaction time, and so it couldn’t tear itself apart, if something went wrong, which it never did during a match thankfully), but it was simple enough for me to get our programmers to understand the concept and write the logic themselves, so I called it a success, lol.

EDIT: Code LInk and Demo Video

1 Like

Do you have a link to Vex to where I can find something like this?

Oh, I believe I found it. Was it this? Whatsoever, there are plenty of great gearboxes we could look into. Thanks!

More motors, more power. More power, score faster. Score faster, win more.

Cons: two motors and controllers cost more and weigh more than one.

Pros: faster cycle times, less load per motor per unit performance, and redundancy.

5 Likes

In offseason 2018 we built a 4 motor elevator (so we could climb), and it was incredibly fast, we just used a gearbox and had chain running back (from the center of the elevator and the carriage, I know that the space savings of having the motors close to the elevator are great, but that is an easy way to make a super fast elevator.

If you ask 148 in 2018, their Uppercut elevator was running on a 3 or 4 to 1 versaplanetary with 1 775 and was like way fast. Although they did balance it with constant force spring pretty well and the whole elevator assembly was like <15(maybe 10) lbs. So really it’s just designing to what your trying to accomplish.

1 Like

Like in other applications, it’s just a question of the requirements. First, figure out how fast you want it to go. Increase by ~50% and pick your desired motor, and you can figure out the total gear ratio you need. Next, figure out how much torque you need. Increase by ~50%. Compare that to the output torque you get from your motor and gear ratio combination - that ratio will tell you how many motors you need.

As others said, keeping everything mechanically linked is pretty important. It’s easiest to do that with a multi-motor gearbox, in most cases.

2 Likes

More motors is not the only way to get more power. We used Neo motors on our climb elevators and hatch elevator and ended up being very happy with both the speed and torque. This is a lighter solution that 2 775’s, 1 extra speed controller, and a dual motor input would be. I believe final reduction was 20:1 on a 8lb cascading elevator. Constant force springs and carbon fiber are two good options to drastically reduce elevator load. I will also add a +1 for mechanically linked elevators.

This is somewhat a copy of above, but my take. I have built 2 elevators in the past year, each with two motors. The one in 2018 needed to lift the 140lb robot, and get to the scale, so it had two motors to speed it up. It also helped with reliability as the lift had other issues and continually burned up a motor, thus, it could run on one motor and keep our robot useful. In 2019 it had to lift an intake with counterbalancing springs. However, we wanted to have reliability (be able to smoke a motor) and have the lift power through sticky sections. Thus, even though we could have gone with one motor, it was only 1.5 lb more and $40 to add some reliability and extra power. We have looked at four motors, but it significantly increases complexity (custom gearbox) and seems to be too much (unless you are lifting multiple robots).

Some extra information regarding the 2019 elevator:

Tuned for top speed of 16fps and acceleration of 16fps/s

(#TeamRealUnits until I get my math wrong and it isn’t real anymore)

The below video is from a PID tuning exercise with no manipulator on the carriage, to which we attempted to run our profiles at 45fps & 45fps/s; no idea how close it was to hitting it’s target velocity.

I find it more useful to figure out the max power (torque times angular speed, or force times speed) needed, add a buffer (+100% typically) to figure out what motor(s) will work, then work out the gear ratio based on the maximum speed and torque individually. Probably gives the same answer, but with less likelihood of rework being needed.

Either more [bigger] motor or more motors are required to increase power, unless you only need it for a short duty cycle and are pre-storing mechanical energy (e.g. a flywheel or spring).

3 Likes

No, but it’s the easiest.

195 ran 4x 775pro on our elevator this year. It was geared to go to full extension in under 1s (~12:1), so we needed the power from the 4 motors to get it there with the heavy turret carriage.

1 Like

This. By a lot.

Also, @RobotMark, it’s the whole topic of the thread: 1 motor vs multiple motors. Not motor selection or design. I do agree with your points in general principle though.

In the real world you probably choose a single motor that’s strong enough to do the job. With the motor rules imposed in FRC, you often need multiple motors to satisfy the requirments of your mechanisms. You can use my design spreadsheet to help you figure out how many (and what kind of) motors you need.

2 Likes