Go to Post I love how these contests always seem to find their way to go back to Dave Lavery and (pick one) Krispy Kreme/Heidi/Photoshop - DCA Fan [more]
Home
Go Back   Chief Delphi > CD-Media > White Papers
CD-Media   CD-Spy  
portal register members calendar search Today's Posts Mark Forums Read FAQ rules

photos

papers

everything



Cyber Blue FRC 234 Motor Test Report 12-8-2017

By: Chris Fultz
New: 12-08-2017 12:11 PM
Updated: 12-08-2017 12:11 PM
Total downloads: 874 times


During the fall of 2017, Cyber Blue completed performance testing of CIM, MiniCIM and VEXPRO 775 based drives. Multiple numbers of motors and combinations were tested. This report summarizes the testing.

During the fall of 2017, Cyber Blue completed performance testing of CIM, MiniCIM and VEXPRO 775 based drives. Multiple numbers of motors and combinations were tested. This report summarizes the testing.

Tests included acceleration, top speed, turns and pushing, using low and high gear options for some of the tests.

Attached Files

  • pdf Cyber Blue FRC 234 Motor Test Report 12-8-2017

    Cyber Blue FRC 234 - Motor Test Report 12-8-2017.pdf

    downloaddownload file

    uploaded: 12-08-2017 12:11 PM
    filetype: pdf
    filesize: 1.58MB
    downloads: 872



Recent Downloaders

Discussion

view entire thread

Reply

12-08-2017 12:15 PM

Chris Fultz


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

During the fall of 2017, Cyber Blue FRC 234 completed a series of tests on FRC motors that are commonly used for drive systems.

Motors tested included CIM, MiniCIM and VEXPRO 775's.

Multiple combinations of motor counts and combinations were used. Tests included acceleration, top speed, time to 50 feet, a turns course and a wall push.

This report summarizes the findings from the report in both table and graph forms.


The base data can be found here: https://tinyurl.com/FRC234MotorTestData



12-08-2017 12:29 PM

Mark McLeod


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Ahh, the sweet smell of data...

Thanks for the work Chris & company!



12-08-2017 12:33 PM

noah.gleason


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Sweet data! From a quick glance, looks like 6 mini-CIM is the best?



12-08-2017 12:33 PM

nuclearnerd


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Super useful, thanks!

One thing: On page 6 you list the power from the 775 pro drives at 80% of their max power ratings. However you limited the motors to 80% of their input voltage, not power. According to this recent discussion, the power available at each voltage-limited motor is actually 0.8^2 or 64%.



12-08-2017 12:46 PM

JesseK


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

4 MiniCIMs do not beat 4 CIMs in a drag race! Your experiment results are backwards. Please fix it.

This is great baseline data. Great job with the experiment setup, data collection, and presentation! It is a little difficult to discern between some of the blues in the charts of the PDF, but the provided data is helpful in sorting it out. Thanks for all of the hard work!



12-08-2017 12:48 PM

nuclearnerd


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by noah.gleason View Post
Sweet data! From a quick glance, looks like 6 mini-CIM is the best?
I'm not sure that conclusion can be drawn. Since the gear ratio was (roughly) the same for all tests, the drive trains weren't optimised for different amounts of power available. That said, this is *really* useful data for validating acceleration simulation models.



12-08-2017 12:51 PM

aciarniello


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Welp. I was having a very productive afternoon. Can't wait to dig into your results.

Thanks for doing this test and sharing. This will be hugely helpful in our decision making process next year!



12-08-2017 12:53 PM

Mike Schreiber


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Awesome data. It appears the available current from the battery posted in the battery spec sheet is about right, don't plan to get much more than 270 amps out of it. The behavior you've shown has confirmed a lot of the modelling I've done on this. Thanks for sharing!

Quote:
Originally Posted by noah.gleason View Post
Sweet data! From a quick glance, looks like 6 mini-CIM is the best?
Depends what you mean by best. The 6 mini CIM drive had a slightly higher free speed than the other CIM configurations and the voltage limited 775 pro configurations. In a 50ft sprint I would expect the slightly higher gearing to be an advantage.



12-08-2017 12:54 PM

Chris Fultz


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by JesseK View Post
4 MiniCIMs do not beat 4 CIMs in a drag race! Your experiment results are backwards. Please fix it.
The results are recorded correctly. This is a summary of the raw data, with no factoring or adjustment for input speeds. That is why we also included the raw data files.



12-08-2017 12:54 PM

Chris Hapstack


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

This is spectacular, and ever so timely! Thanks, Cyber Blue!


Quote:
Originally Posted by noah.gleason View Post
From a quick glance, looks like 6 mini-CIM is the best?
I'm not convinced of that result - what makes you think so?



12-08-2017 01:04 PM

noah.gleason


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
I'm not sure that conclusion can be drawn. Since the gear ratio was (roughly) the same for all tests, the drive trains weren't optimised for different amounts of power available.
Ah yeah, good point.

Quote:
Originally Posted by Chris Hapstack View Post
I'm not convinced of that result - what makes you think so?
It was at the top of most of the charts, especially the on-ground speed tests. It's definitely not conclusive, just a first-glance takeaway.



12-08-2017 01:10 PM

JesseK


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by Chris Fultz View Post
The results are recorded correctly. This is a summary of the raw data, with no factoring or adjustment for input speeds. That is why we also included the raw data files.
I hope you took the emojis to mean that my mind is blown by the finding, rather than a challenge to the results .

Page 22 shows that nearly all of the setups took a full 1 second or more to stop from full speed. It is significant that such a "small" action in a full cycle of 2017 could eat away at 2 or more seconds of cycle time, and isn't an insight I would have even thought of.

In test #4, did the "stop" happen simply from brake mode on the talons, or was a reverse thrust applied before the robot motion was fully-arrested?



12-08-2017 01:20 PM

nuclearnerd


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by JesseK View Post
Page 22 shows that a nearly all of the setups took a full 1 second or more to stop from full speed. It is significant that such a "small" action in a full cycle of 2017 could eat away at 2 or more seconds of cycle time, and isn't an insight I wouldn't have even thought of.
How soon until the elite teams start fitting disc brakes to their drivetrains for extra-high deceleration



12-08-2017 01:22 PM

Mark McLeod


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
How soon until the elite teams start fitting disc brakes to their drivetrains for extra-high deceleration
I call that a "wall"



12-08-2017 01:25 PM

Oblarg


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
How soon until the elite teams start fitting disc brakes to their drivetrains for extra-high deceleration
If you have a shifter, immediately shifting to low-gear to slow down basically does this. We found it quite helpful last year.

We might run our own versions of some of these tests to verify the data here. I find the shifting results a bit surprising, but I think "shift after 2 seconds" is the likely culprit; it'd be nice to test with actual autoshifting code.



12-08-2017 01:29 PM

Chris Fultz


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by JesseK View Post
I hope you took the emojis to mean that my mind is blown by the finding, rather than a challenge to the results .

Page 22 shows that a nearly all of the setups took a full 1 second or more to stop from full speed. It is significant that such a "small" action in a full cycle of 2017 could eat away at 2 or more seconds of cycle time, and isn't an insight I wouldn't have even thought of.

In test #4, did the "stop" happen simply from brake mode on the talons, or was a reverse thrust applied before the robot motion was fully-arrested?
I wasn't sure on the CIM comment, but it was a good point to make that we are reporting raw data, so the different input speeds are an influence. We chose to not do a lot of data manipulation.



12-08-2017 01:34 PM

Collin Fultz


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by JesseK View Post
In test #4, did the "stop" happen simply from brake mode on the talons, or was a reverse thrust applied before the robot motion was fully-arrested?
You can see this in the data for each test. Starting with Config 2 (6 CIMs), we go from 1 to -1 over approx 300 ms



12-08-2017 04:03 PM

Tom Line


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Chris, this is an incredible service to the community. Thank you and your team for the dedication.



12-09-2017 06:49 AM

Basel A


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

A CIM drivetrain (4 or 6) geared to the same free speed as a miniCIM drivetrain would almost certainly be faster (if it weren't, I'd be questioning the test methodology). In this test data, 6 miniCIM is faster than 6 CIM almost across the board. It's interesting that, in most cases, the marginal difference in free speed makes a bigger sprint speed difference than the 50% torque difference.



12-09-2017 07:51 AM

Chris Fultz


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by Basel A View Post
A CIM drivetrain (4 or 6) geared to the same free speed as a miniCIM drivetrain would almost certainly be faster
I am not sure you can say that. The data shows that after a certain power point, there is not a noticeable difference in acceleration for the gearing and weight of these robots. The "2 CIMs" is noticeably slower in acceleration and top speed, but after the next step up in motor power combinations everything seems to converge and the differences are very small.

Quote:
Originally Posted by Basel A View Post
(if it weren't, I'd be questioning the test methodology).
Honestly, that is why we published the methodology and all of the data. It would be very beneficial if someone duplicated our tests and we could compare results - Even if a team just did one or two configurations. There were some things in the data that surprised us and confirmation or contradiction of these findings would benefit everyone.



12-09-2017 11:13 AM

wilsonmw04


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

thank you to everyone who worked on this. This is a great tool for the community. It seems that some of my assumptions about motors and robot performance is wrong.



12-09-2017 11:20 AM

SPang


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

This is excellent work. I was wondering how the 4 CIM 2 Mini combo compared to the all 775 and the all MiniCIM. The results were interesting to say the least. Thanks for producing this resource for the community!



12-26-2017 06:37 PM

nuclearnerd


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
I'm not sure that conclusion can be drawn. Since the gear ratio was (roughly) the same for all tests, the drive trains weren't optimised for different amounts of power available. That said, this is *really* useful data for validating acceleration simulation models.
So I sat down to try to simulate your testing runs with Ether's acceleration model. I eventually got pretty good agreement, but only after putting in some pretty weird parameters:
  • Internal battery resistance is 0.021 ohms, which is pretty high
  • The mechanical efficiency is 0.55, which is very low
  • The gear ratio is 5.3 which is 20% higher than your actual high gear ratio (4.4)

These parameters are far enough from intuitive values that I've either got a wrong combination that happens to work well (over fit), or the simulation math isn't quite right.

I'm going to take these parameters and see if they predict your other runs in your data set. We took our own measurements this fall, which I'll check too. My hope is to come up with parameters that gives reasonable simulations for future drivetrains. In the mean time, take this data with a grain of salt.

Graph: https://drive.google.com/open?id=1nZ...G86IUoQIcNdtqs
Data: https://drive.google.com/file/d/1KSz...ew?usp=sharing
Ether's Sim Batch File Params: https://drive.google.com/file/d/17cF...ew?usp=sharing



12-26-2017 08:29 PM

Oblarg


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
The mechanical efficiency is 0.55, which is very low.
This is similar to direct measurements we have made of this in the past; see our drive characterization whitepaper for details.



12-27-2017 07:50 AM

JesseK


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
So I sat down to try to simulate your testing runs with Ether's acceleration model. I eventually got pretty good agreement, but only after putting in some pretty weird parameters:
Try this one. It's pretty close to 234's data, without a lot of weirdness. By default it's setup for ILITE's 2017 drive train.



12-27-2017 11:46 AM

messer5740


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Correct me if I'm wrong, but it seems to me that there is a very small difference in performance from a 4 cim (or mini cim) version compared to the others in terms of speed, but not in terms of draw on the wall test. Would this mean that if teams were limited on budget but want high performance they should chose the 4 cim option, or do other motor options such as the 775 pros blow it out of the water?

Basically what I can see from the 4 cim data compared to the others is avoid running it into walls for 10 seconds and you're good



12-27-2017 12:18 PM

JesseK


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by messer5740 View Post
Correct me if I'm wrong, but it seems to me that there is a very small difference in performance from a 4 cim (or mini cim) version compared to the others in terms of speed, but not in terms of draw on the wall test. Would this mean that if teams were limited on budget but want high performance they should chose the 4 cim option, or do other motor options such as the 775 pros blow it out of the water?

Basically what I can see from the 4 cim data compared to the others is avoid running it into walls for 10 seconds and you're good
4 CIMs have been good enough to run fully-weighted robots for a decade now, and they're still good enough moving forward.

Teams who use different types or quantities of motors are optimizing for sprint time (8x775, 6xMiniCIM, or 6xCIM are better), weight (6x775 or 8x775 are better), or thermal buildup during finals (6 MiniCIMs are better).



12-27-2017 11:47 PM

nuclearnerd


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by JesseK View Post
Try this one. It's pretty close to 234's data, without a lot of weirdness. By default it's setup for ILITE's 2017 drive train.
I don't know where to get the raw data from the ilite spreadsheet, so I'm not sure what to plot to compare with the 234 data, but just entering the test parameters naively, I think it predicts a top speed of ~21 fps for the 6 cim configuration, whereas the 234 data gives a top speed of only 17 fps.

I'm still tweaking parameters on Ether's sim. The impetus for all this work is that we chose a ratio this year (G=4.0 with 4" wheels) that we thought would get us close to 20fps. In reality we maxed out around 16 fps, similar to 234s data. I'm anxious to get a simulation that predicts why, so that we can make more informed choices next year. (or else figure out whether our robot was unusual)



12-28-2017 12:20 AM

nuclearnerd


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Similarly, the WCP calculator is a couple of ft/s behind the 234 data, using default parameters. For 6 cims in high gear, it predicts a top speed of 19fps (actual is more like 17), a speed at 20ft of 17.3 fps (actual is 14 ft/s) and a time to 20ft of 1.75s (actual is 2s): https://drive.google.com/open?id=1Hm...EZgBoo1TRyFeke

(btw, is there a way to attach files in reply-to-threads? Uploading to google drive is getting tedious, and I'd like to show a picture directly)



12-28-2017 01:31 AM

Oblarg


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
Similarly, the WCP calculator is a couple of ft/s behind the 234 data, using default parameters. For 6 cims in high gear, it predicts a top speed of 19fps (actual is more like 17), a speed at 20ft of 17.3 fps (actual is 14 ft/s) and a time to 20ft of 1.75s (actual is 2s): https://drive.google.com/open?id=1Hm...EZgBoo1TRyFeke

(btw, is there a way to attach files in reply-to-threads? Uploading to google drive is getting tedious, and I'd like to show a picture directly)
The WCP calculator does not account for constant frictional losses, which are substantial, and appears to scale top speed proportionally by "overall efficiency," which is not correct. It is also possible to make the "speed at distance" value greater than the "top speed" by making the "overall efficiency" sufficiently low, which makes me skeptical of their math in general.



12-28-2017 07:23 AM

JesseK


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
I'm still tweaking parameters on Ether's sim. The impetus for all this work is that we chose a ratio this year (G=4.0 with 4" wheels) that we thought would get us close to 20fps. In reality we maxed out around 16 fps, similar to 234s data. I'm anxious to get a simulation that predicts why, so that we can make more informed choices next year. (or else figure out whether our robot was unusual)
The point of my spreadsheet is that top speed is irrelevant, particularly for shorter distances. 234's time to target was pretty close to what my sheet predicted.



12-28-2017 09:35 PM

mylodon


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Not familiar with the gearbox used, how do you derive the effective gearing ratio from "2.92 spread and a 14-tooth input gear"?



12-29-2017 12:13 AM

nuclearnerd


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by mylodon View Post
Not familiar with the gearbox used, how do you derive the effective gearing ratio from "2.92 spread and a 14-tooth input gear"?
Yeah, I wish they posted the actual value. But they do list the gearbox part number, a WCP 3 cim DS, which according to the wcp calculator linked above has a high gear ratio of 4.4.



12-29-2017 07:03 AM

Collin Fultz


Unread Re: paper: Cyber Blue FRC 234 Motor Test Report 12-8-2017

Quote:
Originally Posted by nuclearnerd View Post
Yeah, I wish they posted the actual value. But they do list the gearbox part number, a WCP 3 cim DS, which according to the wcp calculator linked above has a high gear ratio of 4.4.
Good feedback, sorry about that. Here's what we used.

Initial Gear Reduction: 14:40
Low Gear Reduction: 14:60
High Gear Reduction: 30:44

Overall High Gear Ratio: 4.2:1
Overall Low Gear Ratio: 12.2:1

https://www.dropbox.com/s/bwuoe7enpc...pture.PNG?dl=0



view entire thread

Reply

Tags

loading ...



All times are GMT -5. The time now is 06:08 PM.

The Chief Delphi Forums are sponsored by Innovation First International, Inc.


Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi