ILITE Drivetrain Simulator, v2020

introducing this year’s ILITE Drivetrain Simulator. Not only have we incorporated more elements about how ILITE designs our drive trains, I’ve included several rounds of feedback I have received from many teams across FRC.
ILITE-DriveTrainSimulator.xlsx (3.0 MB)

Input a target motor, number of motors, and over 25 other elements about your drive train design in order to see more characteristics, charts, and design tradeoffs than ever before! Furthermore, there is more emphasis on the KOP drivetrain, in order to help teams understand what impacts “steroids” can have.

What's New?
  • Added a changelog tab to track everything that has changed
  • Added the Falcon and NEO550 motors
  • More accurate modeling of deceleration, wheel slip, current draw during wheel slip
  • A static KOP drivetrain image so you can compare your choices against the original KOP
    or strategize about what to do if you have the KOP but face a more optimized drive train
  • Now you can show or hide all of the individual lines on the tradeoff graphs, in order to clean them up a bit
  • Added a few enhancements related to shifting transmissions
  • More useful elements like turning deadband, estimated throttle response times, and motion profiling graphs

This year, the spreadsheet has many more elements!

To help guide you through this simulator, I created a tutorial! I will be available on ChiefDelphi and FRC Discord to answer questions (CD is typically faster, since Discord is a bit noisy).

Want to get notified about spreadhseet updates with fixes or further enhancements? Then “Watch” or “Star” the github repository here:

35 Likes

I was looking for one earlier this week and now my prayers have been answered. Thank You!

4 Likes

I really like this simulator. I am math/engineering/physics challenged and find it relatively easy to understand and manipulate. The tutorial is excellent.
There are a couple of things I could use some help with.
Is there a way to put in wheels that aren’t matched, like a combination of omnis and traction wheels? If this is a stupid idea, you can just tell me.
You mentioned published COF data but i have had a difficult time finding it on vendors websites, and can’t find anything that separates lateral COF. Would it always be the same?

1 Like

Generally speaking, lateral CoF is the same as normal CoF except in specific conditions, such as 25’s treaded wheels, omni wheels, or (depending on how the dynamics are adjusted for) Mecanum wheels.

To put in wheels which aren’t matched, I’d like to know more about the setup itself. Is it a 4+2 (4 traction, 2 omni, all touch the floor) setup? If so, and if the c.g. is close enough to the center of the robot, then you can average the two lateral CoF’s for the purposes of the simulation. If it’s something different, let me know.

Thank you for this awesome calculator and the new tutorial!

1 Like

2 designs. Corner omni’s with center traction and 4:2 as you described. Both setups without center drop.

For the corner omni setup, I would use 0.5 lateral traction. It may come out slightly less than that though.

In both cases, I would average the forward/reverse traction. Say you use blue nitrile (~1.25 CoF conservatively) on the middle wheels and omnis on the outer wheels. Then the approximate static CoF for the robot would be (1.0 + 1.25 + 1.0)/3 = ~1.08 with a lateral CoF of 0.5. This should get you close enough to actual.

VEX’s data suggests their omni wheels are 1.1 CoF. But in our testing the VEX colsons (1.0 CoF) get better traction than their omni wheels on carpet. The carpet could be the difference between our tests and theirs. I want to say conventional wisdom (from ~2012 or so, JVN, 234, and several other teams did traction testing) put omnis at 0.9 CoF, but the blogs/papers seems to have evaporated.

2 Likes

@JesseK from watching your video (thanks, there are some cool features I never really knew about or understood!), it seems like you attempt to set your current limits, and more importantly low gear ratios, to avoid being traction limited. Why? I’ve generally viewed by traction limited as a positive thing, as it prevents stalling out motors when things go astray (your auton drives into a wall, for instance).

The reality is, I’ve never designed to push, and the video probably reflects that. Yet this was a request, so I added it. I still need to understand duty cycle with brushless motors and their current limits. It seems like a drive train that is current limited to 240A output at a single stator does not actually pull 240 from the battery. For now, that’s an improvement for next year.

1 Like

Just wanted to drop in and say that I’ve recently learned about the VEX 13T pinions which have a 14T center-to-center distance.

In the video I mentioned that the Falcon cannot have a different gear ratio because the KOP gearbox isn’t designed to be configurable as the KOP kit. Yet with the 13T VEX pinion, not only can either the NEO (217-3416) or Falcon (217-6921) be adjusted slightly, the Falcon itself can be brought down to the same floor speed as a NEO. This should give the KOP even more throttle response (~33% better) for teams who do not want an excessive floor speed for a 2020 short-field strategy.

1 Like

Thanks for creating and posting this. From my initial playing around with gear ratios, it looks like we might be able to optimize how our KOP chassis performs on the field.

I am using your simulator with an old version of EXCEL (2007) and none of the drop down menus seem to work. For instance, I have to go over to the Motors tab, copy a motor name and paste it into the Motor field to change motor types. Similarly, I have to type SHOW or HIDE in the fields at the very right for controlling the three charts. Is this due to a compatibility issue with my old version of EXCEL? I can try this at work tomorrow where we have a much newer version of EXCEL.

Yes, this is an issue with older versions of Excel. There is also an issue with Excel 2010 since the ISNA() function doesn’t exist until v2013. I have successfully tested this spreadsheet with 2013, 2016, and 2019 version of excel.

Thanks for the quick reply. Would certain functions not really work properly such that I would get incorrect or misleading results if I continue working on it with the 2007 version? If so, I will stop and do it some other time.

Yes - specifically you’ll notice that when the motor changes none of the calculations will update. This is because there are 3 critical calculations which need the IFNA() functions.

Thanks again for your quick replies.

I’ve been thinking a bit about the time-to-stopped where the robot ends at the sprint distance instead of beginning deceleration there. For the case of reversing to stop, the wheels will be slipping the whole time so the robot will be decelerating at a constant a = \frac{F_f}{m} = \frac{mg\mu_k}{m} = g\mu_k. So we can calculate the stopping distance as:

\begin{cases} \Delta x = v_i t - \frac{g \mu_k}{2} t^2 \\ v_i - g \mu_k t = 0 \end{cases} \implies \Delta x = v_i \left( \frac{v_i}{g \mu_k} \right) - \frac{g \mu_k}{2} \left( \frac{v_i}{g \mu_k} \right)^2 = \frac{v_i^2}{2g \mu_k}

By modifying column OY (“Hit Target?”) to =IF(PU8+LO8^2/(2*32.2*'Drive Train'!$F$17)>='Drive Train'!$F$29,1,0)* the robot begins decelerating at \Delta x before the sprint distance, so it ends right at the final distance. For the example I tested, the robot stopped at 20.2 ft for a 20 ft sprint distance instead of 22.2 ft normally. Something similar can be done for braking and coasting to stop, taking the average deceleration since it isn’t constant for the other methods.

 
* This formula for cell OY8, then drag down to fill the column