One speed vs Two speed gearboxes

[RoboChair]

Originally Posted by Abhishek R

So we've been thinking about auto-shifting some more, and how to increase our efficiency. I've also heard that 1678 achieved speeds of 22 ft/s on the field and wondered how they did that, especially if they started in first gear then shifted or if they were in 2nd the entire time (assuming they used shifters). We were happy with our auto-shifting system last year and would like to improve it further. This led to a few questions:

How do we figure out the correct gearing for the low gear supposing we know we want to have a high gear of 17 ft/s? This needs to maximize acceleration as well as battery life.

What exactly does it mean to be traction-limited and what are the physics equations and calculations surrounding that?

We are planning on using either 4 CIMs, 4 CIMs and 2 MiniCIMs, or straight up 6 CIMs, so any advice on those configurations would help as well. I didn't want to start a new thread since there has been so much discussion on the topic, but since using auto-shifting as a method of avoiding blowing the main breaker was discussed here, I figured it was apt. We didn't have to worry about that last year since we only had 4 CIMs on the main drivetrain.

We were GEARED for 22fps(no loss), we measured somewhere between 17 and 18 actual. About 12 miles per hour or if you want to give a good value to some pit scouts, that's 5.3% the speed of sound!

[asid61]

Originally Posted by RoboChair

We were GEARED for 22fps(no loss), we measured somewhere between 17 and 18 actual. About 12 miles per hour or if you want to give a good value to some pit scouts, that's 5.3% the speed of sound!

As a general rule, most speeds posted for gearboxes and drivetrains are free speeds.
Because everybody posts free speeds, 22fps is still extremely fast, even if the actual is 17 or 18fps. A robot geared for 10fps may only reach 8.

[Chris is me]

Originally Posted by asid61

As a general rule, most speeds posted for gearboxes and drivetrains are free speeds.
Because everybody posts free speeds

Honestly, most people just post what the JVN calculator spits out with the default constant values. Unless someone says "free speed", I don't generally assume they are posting free speed. Otherwise a "12FPS" drivetrain would be a lot less impressive. Certainly not "everybody".

[Abhishek R]

Originally Posted by RoboChair

We were GEARED for 22fps(no loss), we measured somewhere between 17 and 18 actual. About 12 miles per hour or if you want to give a good value to some pit scouts, that's 5.3% the speed of sound!

Oh ok, I see. What was your lower gearing set to, and did you have to upshift/downshift in order to get to that 17 fps, or were you able to start in 2nd gear just fine?

[cgmv123]

Originally Posted by RoboChair

that's 5.3% the speed of sound!

If you want to make it sound less impressive, call it Mach 0.053

[Michael Corsetto]

Originally Posted by Abhishek R

Oh ok, I see. What was your lower gearing set to, and did you have to upshift/downshift in order to get to that 17 fps, or were you able to start in 2nd gear just fine?

I don't know what our high/low free speeds were.

However, I do know we did not auto shift. We just stuck it in high gear until we ran into so traffic, or the drive train was stalling (to prevent tripping the breaker).

Remember kids, it's not about the speed of your drive train, it's how you use it

Merry Christmas!

-Mike

[Abhishek R]

Originally Posted by Michael Corsetto

I don't know what our high/low free speeds were.

However, I do know we did not auto shift. We just stuck it in high gear until we ran into so traffic, or the drive train was stalling (to prevent tripping the breaker).

Remember kids, it's not about the speed of your drive train, it's how you use it

Merry Christmas!

-Mike

Thanks for the advice, I appreciate it. Merry Christmas to you too!

[thefro526]

Originally Posted by Abhishek R

So we've been thinking about auto-shifting some more, and how to increase our efficiency. I've also heard that 1678 achieved speeds of 22 ft/s on the field and wondered how they did that, especially if they started in first gear then shifted or if they were in 2nd the entire time (assuming they used shifters). We were happy with our auto-shifting system last year and would like to improve it further. This led to a few questions:

How do we figure out the correct gearing for the low gear supposing we know we want to have a high gear of 17 ft/s? This needs to maximize acceleration as well as battery life.

What exactly does it mean to be traction-limited and what are the physics equations and calculations surrounding that?

We are planning on using either 4 CIMs, 4 CIMs and 2 MiniCIMs, or straight up 6 CIMs, so any advice on those configurations would help as well. I didn't want to start a new thread since there has been so much discussion on the topic, but since using auto-shifting as a method of avoiding blowing the main breaker was discussed here, I figured it was apt. We didn't have to worry about that last year since we only had 4 CIMs on the main drivetrain.

Figuring out the ideal High / Low gear split requires a bit of thinking, a bit of math, and honestly becomes a bit of a preference issue in the end. If you're planning on auto-shifting, especially if you're trying to optimize acceleration, you want to look at the acceleration of both options, and pair them together in a way that is going to give you the most smooth accel. If your split between high and low gear is too extreme, you'll notice a "dip" in acceleration, between the top end of low gear, and it's shift into high, where the robot appears to struggle a bit. Anecdotally, anything beyond a 2.5 split will stair having a significant "dip" between high and low gears, but it's been a while since I've looked at the issue in detail.

With that being said, and as Michael mentioned above, you can get away without shifting from low gear to high gear even when gearing aggressively - if your driver / team knows what they're doing. If you get used to driving at speed, and or optimizing paths so that you're not changing directions a lot, you're minimizing the amount of time spent getting up to speed.

As far as being "Traction Limited" is concerned - it's a function of you're drivetrains traction, relative to the "thrust" of the drive. Essentially if your drivetrain can produce enough torque (thrust) to spin the wheels before your breaker would trip, that'd mean you're traction limited. To calculate this, you need to first find the force of friction (Ff = Fn x CoF - is a good start) and then find the effective "Trust" (wheel torque) of the Drivetrain. (Input Torque x Gear Ratio / Wheel Radius - as a starting point) From there, you look at the current draw for that torque, and then compare that to the performance of your breakers.

One of the biggest things to be aware of is the current per motor, and the number of motors in the system. In a 4 CIM drivetrain, designing to be traction limited usually means designing around the 40A breakers in the PDB, so that you're not drawing enough current to trip them - But in a 6 CIM drivetrain, it's better to design around the 120A main breaker, specifically not exceeding a total current draw of about 150% (180A) which works out to about 30A or so a motor - doing this ensures that you shouldn't ever (or at least rarely) trip the main breaker...

[Ekcrbe]

Originally Posted by thefro526

Anecdotally, anything beyond a 2.5 split will stair having a significant "dip" between high and low gears, but it's been a while since I've looked at the issue in detail

I know that for the last few years, we've had high/low splits of right around 3:1 and free speeds of ~5/15 FPS, and I can't imagine wanting them to be any closer together than that. Even with only 4 CIMs, starting in high gear is fine and low can be almost exclusively reserved for pushing and fine control. A simple software adjustment like squaring the drive joystick values makes control in high gear noticeably more manageable, and if we were to change anything, I would say we would try to push it faster before deciding to slow it down.

[Abhishek R]

Originally Posted by Ekcrbe

I know that for the last few years, we've had high/low splits of right around 3:1 and free speeds of ~5/15 FPS, and I can't imagine wanting them to be any closer together than that. Even with only 4 CIMs, starting in high gear is fine and low can be almost exclusively reserved for pushing and fine control. A simple software adjustment like squaring the drive joystick values makes control in high gear noticeably more manageable, and if we were to change anything, I would say we would try to push it faster before deciding to slow it down.

On our 2013 robot, which did not auto shift, we often had trouble starting in 2nd gear, and it seemed to take a toll on the batteries whenever we tried. The robot simply did not like to get going very quickly in 2nd gear, so we started in 1st and upshifted to 2nd once we had some motion.

Do you have any guesses as to why this may be?

[Ekcrbe]

Originally Posted by Abhishek R

On our 2013 robot, which did not auto shift, we often had trouble starting in 2nd gear, and it seemed to take a toll on the batteries whenever we tried. The robot simply did not like to get going very quickly in 2nd gear, so we started in 1st and upshifted to 2nd once we had some motion.

Do you have any guesses as to why this may be?

Well, I took a glance at your season highlight video, and the first thing that struck me was your wheels. How much did they weigh? They look relatively heavy. If you were rolling on heavier wheels (especially if a lot of that weight is concentrated ~4" from the center in the thick rubber tread), the wheels' moments of inertia can come into play. It probably took a lot more torque to get those wheels rolling than lighter ones like the VexPRO traction wheels (6" in 2013, 4" in 2014) that we've used.

One of the first things that comes to mind when you wonder about the physics of accelerating a robot is inertia, but rotational inertia can become a big factor too in some designs. That's my best guess as to what it was, especially if you didn't have as much of a problem with it in other years.

[Abhishek R]

Originally Posted by Ekcrbe

Well, I took a glance at your season highlight video, and the first thing that struck me was your wheels. How much did they weigh? They look relatively heavy. If you were rolling on heavier wheels (especially if a lot of that weight is concentrated ~4" from the center in the thick rubber tread), the wheels' moments of inertia can come into play. It probably took a lot more torque to get those wheels rolling than lighter ones like the VexPRO traction wheels (6" in 2013, 4" in 2014) that we've used.

One of the first things that comes to mind when you wonder about the physics of accelerating a robot is inertia, but rotational inertia can become a big factor too in some designs. That's my best guess as to what it was, especially if you didn't have as much of a problem with it in other years.

Hmm, yes, they were 8 inches Colsons, so pretty heavy. We were planning to use 6 inch wheels but then some complications arose that caused us to use bigger wheels, We didn't think it would have that adverse of an effect on the drive performance.

Could electronics be an issue? Or is that a minor concern, efficient wiring?

Thanks for the help!

[Jared]

Originally Posted by Abhishek R

Hmm, yes, they were 8 inches Colsons, so pretty heavy. We were planning to use 6 inch wheels but then some complications arose that caused us to use bigger wheels, We didn't think it would have that adverse of an effect on the drive performance.

The rotational inertia of the wheels does little to hinder robot acceleration.

For anybody interested, I compared the kinetic energy of the robot moving at 10 feet per second to the rotational energy of the wheels on a robot moving at 10 feet per second. If we compare these energies, we'll see how much of our power goes to spinning the wheels, and how much goes toward moving the robot.

A fully loaded robot weighs 150 lbs, which is 68 kg. 10 feet per second is 3.05 meters per second.

KE = 1/2*mv^2 = 0.5(68)(3.05)^2 = 316.285 Joules
That's the amount of energy it takes to bring your robot up to speed.

For the rotational energy, we've got to find the moment of inertia for the wheels. The radius of the wheel is 4", which is equal to 0.1016 meters, and I'll guess that the mass of the wheel is 2 pounds (probably heavier than the actual wheel), which is equal to 0.91 kg. The wheels are disc shaped, so we can use I = 1/2 * m *r^2 = 0.5(0.91)(.1016)(.1016) = 0.0046968 kg * m^2

One rotation of the wheel causes the robot to travel 8*pi inches = 25.1327 inches = 0.63872 meters/revolution.

3.05 meters/second (divided by) 0.63872 meters/revolution = 4.77517 rev/second = 30.00 radians/second

For the rotational energy, E = 1/2 I*omega^2 = 0.5(0.0046968)(900) = 2.114 Joules per wheel.

For six wheels, that's 12.684 Joules for a robot.

tl;dr, it requires 328.969 Joules to bring your robot to 10 feet per second, and 12.684 of these Joules (3.8%) are used to get your wheels up to speed. This assumes that you have 6 colsons with a diameter of 8" and a mass of 2 pounds.

The fact that the wheels spin when the robot moves makes the robot feel 5.7 pounds heavier to the drive system.

[Ekcrbe]

Originally Posted by Jared

The rotational inertia of the wheels does little to hinder robot acceleration.

Good to know. I was actually in the process of working out this math myself, but you beat me to it. I guess I'm not all that surprised that it works out to be so small, but it was the only reason I could come up with other than "it just is." I did expect it to be more like 8-9% of the robot's weight rather than <4%, but kudos for proving it.

And those wheels are just under 2 pounds each.

[Jared]

Originally Posted by Ekcrbe

Good to know. I was actually in the process of working out this math myself, but you beat me to it. I guess I'm not all that surprised that it works out to be so small, but it was the only reason I could come up with other than "it just is." I did expect it to be more like 8-9% of the robot's weight rather than <4%, but kudos for proving it.

And those wheels are just under 2 pounds each.

I expected it to be more, which is why I did the math. I wonder how much the gears/chain/shafts contribute to the rotational inertia.