Continuous Elevator Gearing

Hello,
I am designing an elevator for CAD practice. It is a 2 stage continuous elevator and is on the heavier side because I made the carriage out of 8020. The total cable length should be around 83" . All I have left to do is designing the winch gearbox. However, since this is my first time designing an elevator I am not sure how to proceed.
My team’s 2019 elevator had 4 775 pros in a 8:54 gear ratio. So far, every elevator CAD I’ve looked at has used 2-4 775 pros. Would it be better to use Neo motors instead of 775’s? If so, how many Neos would be appropriate for a non-endgame elevator? This thread is slightly related but it mostly talks about high torque applications.
For calculating the optimal gear ratio, I plan on using the JVN Calculator’s Linear Mechanism tab. Please let me know if this is meant for elevators or not.
The following is a screenshot from 254’s 2018 build blog. Should I aim for similar speeds, or is this showing climbing speed instead of their normal speed?


Finally, almost every elevator design I’ve seen at has used a 2 speed gearbox. I understand this is necessary if the elevator is being used for climbing, but many robots that did not use their elevator to climb still used 2 speed gearboxes. What benefits does a 2 speed gearbox provide in an elevator other than endgame strength?
Thank you in advance. :slight_smile:

4 Likes

The speeds in the screenshot make me think that that is the speed during normal operations (not climbing). The reason I believe this is that ~1 second travel time from one end to the other is something to aim for. It is quick, but elevator travel speed is very important for quick cycles in 2018 and 2019. As for your other questions, I’ll defer to someone more knowledgeable. However, I do think that 2 NEOs would be a strong replacement for 4 775-pros.

1 Like

I think you’re seriously overpowering this elevator. We ran our 2019 elevator (also constructed of 80/20 extrusion) that was lifting a similar load and used only one 775pro in a 25:1 gearbox. Our time from rest to tier three on the rocket was about 1.1 seconds using Andymark wheel hubs for the cable spools. Four 775pros is manically overpowered, as the stall load in your calculations will tell you. 478 lbs is almost 16x the load you’re actually lifting (you put the load in at 30 lbs). One 775pro (or one Neo) is more than enough to do the job. The only reason to put four 775pros on a winch is if you’re lifting your entire robot with it.

I’ll disagree with this slightly. We would have gladly switched to 2 775’s last year if we had the space. The reason being, current draw. We frequently burned out motors on our elevator that year with similar gearing (don’t remember exactly off hand). Had we had a second motor to spread the load between, I think we would have seen very different results.

So while one motor on a ~25:1 gearbox can do the job, there can definitely be benefits to using 2 or more motors in this scenario.

1 Like

image
Thank you for your advice. Does this look alright? I might change the gears around depending on the gearbox geometry. It says 775 pro but the specs are for two Neos. The gear ratio is low because of the large pulley diameter (I designed a helical pulley). I don’t care about climbing and I want it to be fast.
I would appreciate it if someone would explain the advantages of having a 2 speed gearbox on an elevator other than the ability to climb.

Disclaimer: I have not been on a team that has built an elevator with a 2 speed gearbox while I was a member of the team, this is just my speculation

That said, having a 2 speed gearbox would make it a lot easier to justify not having a brake on the elevator. If the elevator needs to be held in the air for a decent amount of time, having a high gear to get it there and a low gear to minimize current draw to hold it there could be beneficial.

I don’t think a 2 speed gearbox is necessary for an elevator unless you intend to climb with it, and even when you do intend to climb with it sometimes you don’t need a 2 speed gearbox.

Any system, elevator or not, should be designed based on its load and performance requirements. The gearing and number of motors is based off of how much load the elevator experiences under use, how quickly you want to reach your target distance, and how much current you are drawing. System design is a balancing act of requirements, and understanding how each variable affects the others is important when designing a system.

At a base level:
-Higher gearing decreases speed and increases the torque (how much load can be moved, which decreases the current draw per motor)
-Increasing motors adds power (how easily load can be moved), decreases the current draw per motor, and increases the number of motors (obvious, but worth considering).

So where to start? I would start by considering how my system will handle one of the most common high-load scenarios, which for most, is stall. Stall is basically when you hold your system (an elevator, in this case) still at a single point. There are two primary ways of stalling a system: a) a separate mechanical locking brake, or b) powering the motors just enough to hold the system in place without moving it. Each comes with their own advantages and disadvantages.

A mechanical brake allows you to avoid stalling the motors (this is good because stalled motors draw current continuously, heat up, and potentially die/burst into smoke), but comes at the price of an additional mechanical brake system. If you choose this route, you don’t need to worry too much about your motor overheating in stall because you’re not really stalling your motor for any significant amount of time, especially since you’re only pushing current through your motor during the short durations that you are moving the system.

Stalling the motors to brake the system simplifies your mechanical system, but provides another constraint for you to design around: holding voltage. Holding voltage is the voltage (between 0 and 12 volts) that you must run the motor at to provide enough power to counteract the load of your system (the weight, in this case), to keep it in place. With resistance in the system being mostly the same, if we look at the classic V=IR, the higher our holding voltage is, the more current the motor draws. The more current the motor draws, the hotter it gets, and the hotter it gets, the more likely the motor will burn out. This means that when designing a system that stalls motors to brake, it is important to know what holding voltage is “safe”.

Luckily for us, VEX has done some independent testing to provide us with these values. If we visit motors.vex.com, we can see a list of motors that have been tested (as well as the Falcon 500 and the NEO, both of which are missing tests). If you click on any of the non-brushless motors, you will see a series of graphs describing the outcomes of various tests on the motor (here is the page for the 775pro: https://motors.vex.com/vexpro-motors/775pro). What we care about is the Locked Rotor Stall Test. On the X-axis is time in seconds, the Y-axis is output torque from the motor, and each line is the motor at different voltages. Some of the lines drop off immediately, some stay for a little and then drop off, and others move far along the X-axis. These represent how long the motors can survive at stall at the given voltage. As you can notice, the trend is such that the motors last longer the lower voltages they stall at, which supports our V=IR math we brought up earlier. If you notice, only three lines - 6V, 4V, and 2V - last beyond the length of a match. Of those three, we really only have one - the lowest, 2V, that stays mostly consistent throughout a match’s worth of constant running (unlikely scenario, but remember the motor keeps getting hotter throughout the length of a match, and in elims there can be some back to back matches pretty quickly). As a result, some people (myself included) tend to try and keep our system holding voltage at 2 volts or less when using 775pros.

The next question, then, is how we determine holding voltage. There’s a couple ways to do it, but the way I’ve learned how is the following:

Holding Voltage = (Applied Load (lbs) * Pulley Radius (in) * 12 volts) / (Total Gear Ratio * Motor Stall Torque (in*lbs) * # of motors)

Some things to note:
-In the JVN Calc, the linear system tab uses pulley diameter. To turn into radius, multiply the diameter cell by 0.5.
-The math requires that the motor stall torque must use the same force units as the applied load and the same length units as the pulley radius. Since the stall torque in JVN Calc is provided in NewtonMeters, I multiply the Motor Stall Torque by 8.850745792 to convert the NewtonMeters to Inch*Pounds. I suggest making an additional cell in JVN Calc that shows you the holding voltage of your system.

Once you’ve figured out how you are handling stall (I usually use motor stalling, but it is up to you), you can play around with the variables of motor quantity, motor type, and gear ratio to figure out how to best accomplish your system performance goals (go x inches in t amount of time) within your system requirements (Pull no more than n amps of current per motor at stall with a holding voltage under n volts)

4 Likes

The original gearing looks fine to me, based on the duty cycle. We ran similar numbers in 2019 to account for a healthy factor of safety and climbing. Dual falcon or neo is likely an improvement that saves motor slots without sacrificing performance.

We ran dual 775 redlines at 5:1 ratio and a 1.25" diameter spool (Pulley Diameter) initially on our 2019 elevator and the motors got dangerously hot after only a few minutes of practice. We changed the ratio to 25:1 (still with 2 motors) and it ran fine for the entire season (except for one time in practice when the lower limit in the code got re-set and the motors were stalled for 30 seconds or so against the lower stop of the elevator). It still had plenty of speed.

I would be a little dubious about 10:1 with 775s, especially with a 3" pulley diameter. That seems like a lot of load, especially if you are going to be using the motors at stall to hold the elevator. But, I think you are saying that the calculations are actually for NEOs and not 775s? In that case, you should be fine with that ratio.

1 Like

Gear ratio and pulley diameter are important, but ultimately the loaded current draw is the most important. The loaded current draw is only 7.4A in the first image, which is a fairly low heat output (maybe 5W per motor). Even if the current draw ends up being double what you expect, that’s still only about 20W of heat per motor, which can easily be handled with some 40mm case fans if needed.

Our 2019 elevator had 4 775s geared for a loaded current draw of about 6A, and due to large cantilevered loads and an aggressive duty cycle we saw a lot of heat, killing a few motors by champs. A better designed system or added cooling fans would have been a big improvement.

Going to NEOs or Falcons will cut the generated heat due to lower coil resistance, for the same output speed. That lets you use fewer motors. Falcons have about half the equivalent winding resistance as 775pros, and are more efficient with their torque to boot. Falcons also have built-in overheat protection which is awesome.

1 Like

If I am not wrong 254 made a double climb for 2018. So 154(max weight per robot with bumpers and battery)x2(two robots )=308 Ibs. You need at least 308 Ibs to climb with two robots. We have 170.37 Ibs extra for safety. FOS(factor of safety) is 478,37(Stall Load)/ 308(Weight of two robots )=1,5

Note that I am not pro in these calculations so watch Spartan Series/ Mechanical Design for Controllability - Travis Schuh.

Looking through their technical binder they used 4 775’s with a gearbox, geared 12.06:1 for normal operation and 42.44:1 for their climb. Their elevator is on pages 12 and 13

1 Like

Sorry I though they used 20:58 ratio by Jvn calc screen shots

I’ll repeat what I always say about gear ratios: free speeds, ‘adjusted’ speeds, whatever… are meaningless.

Back up, don’t worry about the prior art and JVN sheets and all that good stuff.

Q: What are you trying to do with an elevator?
A: Move an object from point A to point B in as little time as possible.

As such, the best metric to gague an elevator by is… how much time does it take to move from point A to point B- the time to target (or sprint distance).

Using ‘adjusted speed’ to measure any mechanism is like trying to figure out how fast your car is going only by the engine sound: it takes a well-tuned ear, and even then, you’re going to be wrong, and even then, you have to re-tune your ear when you try to figure out the speed for a motorbike, or a semi truck.

You can simulate that here, or with the ILITE simulator, or many other tools. The process is almost the same as a drivetrain with the only exception that you’ll have an external load equal to the force due to gravity acting on the elevator when it rises (which you could reduce/eliminate by means of a counterweight).

Now, you’ve got a way of calculating a metric you actually care about… the question becomes “how low does the time to target need to be”? Obviously, throwing more motors at the problem is ideal (lower time to target = better, right?). But there’s obvious issues, namely:

  1. Limited PDP slots
  2. Motors are heavy
  3. Drawing more current will only serve to brown out the battery (electrical limitations)
  4. Too many motors will break your gearbox
  5. At some point, the elevator will no longer be the limiting factor on how quickly you can perform useful actions.

That last one is (usually) the most important- if your elevator goes up in 0.01 seconds, that’s great- but you probably can’t drive fast enough that the elevator is the limiting factor. If you can lift your elevator while you’re moving from point A to point B (which takes 2 seconds)… then your elevator only needs to have a time to target of 2 seconds to not be the limiting factor. Anything else is wasted power.

5 Likes

Like I said, we built this and used it in competition for our hatch panel elevator and it worked perfectly with no overheating or burnouts. This is with just stalling the motor to hold the elevator in the extended positions, often for 5-10 seconds depending on how much opposition we had in setting the hatch panel. The same motor has been performing like a champ all through the off-season and for demo work (which was a lot of up and down for extended periods) right up until the beginning of this build season. It’s still working now (or would be, if we could get to the robot in storage.) Like I said, we built an elevator very similar to the one the OP is describing, right down to the use of 80/20 for the structure, so what we found in practice applies pretty well.

“wasted power” is usually small compared to drivetrain power for things like elevators. Defining how fast it needs to be is important, but once you do so, throwing motors at it and increasing FOS can be done by managing loaded current draw and speed. I find that PDP slots are usually a much more valuable asset than wasted watts.
Never hurts to go a little faster than you need to. If you plan on rasing the elevator in 2 seconds and it only takes 1, that’s 1 more second your driver has to make a mistake and fix it. For something like an elevator as my primary scoring mechanism, if I know I can dedicate 4 motors to it without affecting other subsystems, can’t hurt to add a little redundancy and speed.

2 Likes

Did you use constant force springs in you elevator design? From what I’ve read here, it doesn’t sound like it. A really good goal would be the elevator, with a game piece, and the motor mechanically disconnected, should be able to be moved by hand, with 2 fingers, and when you let it go it should stay in that position. It should NOT come crashing down into its lower stops, it should stay right where you put it. This is the same way all garage doors operate. This is how a 1/2hp motor can open a garage door that weighs close to 1000lbs…

If your elevator is balanced then the only force the motor needs to overcome is the acceleration of the mechanism. Our 2018 elevator used 2 775Pros, the current draw was the same going up as it was coming down, and it got to its endpoints in about 1 second, very little holding current. A balanced elevator also is easier to tune from a PID standpoint also. The PID values can be the same for both directions and set way hotter.

Balanced elevator good…

8 Likes

Just an anecdote on balancing of a mechanical system:

From the Wiki on the Panama canal lock gates:

These gates are of enormous size, ranging from 47 to 82 ft (14.33 to 24.99 m) high, depending on position, and are 7 ft (2.13 m) thick. The tallest gates are at Miraflores, due to the large tidal range there. The heaviest leaves weigh 662 t (730 short tons; 652 long tons); the hinges themselves each weigh 16.7 t (36,817 lb). Each gate has two leaves, 65 ft (19.81 m) wide, which close to a “V” shape with the point upstream. …

The original gate machinery consisted of a huge drive wheel, powered by an electric motor, to which was attached a connecting rod, which in turn attached to the middle of the gate. These mechanisms were replaced with hydraulic struts beginning in January 1998, after 84 years of service. The gates are hollow and buoyant, much like the hull of a ship, and are so well balanced that two 19 kW (25 hp) motors are enough to move each gate leaf. If one motor fails, the other can still operate the gate at reduced speed.

1 Like

Disclaimer: I haven’t read through all of this so I might be repeating some points.

Typically speaking, the reason that teams throw a bunch of 775’s on elevators isn’t because they need the power, it’s to spread the heat out over more motors. To hold an elevator in place with just the motors means stalling them, and 775’s cool themselves with fans that are connected to the rotor. So, when the motor stops, the cooling stops but the heating doesn’t. By adding more motors each one sees less torque, which means less heat per motor. (Oversimplifying :slight_smile: )

If you’re just lifting a “typical” elevator (IE: you aren’t buddy climbing with it like 254-2018), and you either have a brake or a counterbalance (or both), then you can get away with relatively little power. Sourcing from my experience:

Our 2018 elevator had a single 775 and a bike brake. It lifted fast enough that it wasn’t our scoring bottleneck, and the motor never needed replacement. (picture below, left)
Our 2019 elevator had a single NEO and a “971 brake.” It was lifting a pretty heavy elevator with no counterbalance. That motor was also able to lift the elevator in about a second, and never needed replacement. (picture below, right)

That said, if you have the extra weight and PDB slots, there’s no real downside to just throwing an extra motor on. It’s nice to have some headroom, and it helps counter any unexpected friction in the system.

1 Like

Are you suggesting a water game? Dude, I’m not going to make our elevator buoyant. I’ll stick to springs :diving_mask:

3 Likes