View Full Version : Team 610 pre-season drivetrain
Rob Stehlik
13-12-2013, 13:44
Two of our grade 11 students took it upon themselves to design and build a drivetrain during the pre-season. Matt and Jonathan did a spectacular job, putting in a lot of time working out all the details in CAD, and then machining all of the parts themselves.
http://i.imgur.com/OmtX6Lt.jpg
I encouraged them to try some new things, since that's what the pre-season is for. Here are the specs:
-Fully geared
-8wd, front wheel raised 3/16"
-6" aluminum performance wheels
-Single speed: 10fps (it's all about acceleration)
-Riveted frame construction
-Fiberglass belly pan
-Beefy, heavy construction (didn't care about weight on this one)
http://i.imgur.com/EAyf9Fu.jpg
It should be driving soon!
Akash Rastogi
13-12-2013, 13:53
Pretty sweet, I like how much space you save with the thin tubing.
Did you have issues with assembly? Any access holes at the bottom to help assemble the geared drive?
What gauge is the wire going from the DSC to the controllers?
Very nice! It's laid out much better than most of the stuff I have to work on. It looks better thought out too.
How did they install the gears in the tubes down each side?
If I only had hiring and firing authority...
Rob Stehlik
13-12-2013, 15:35
Pretty sweet, I like how much space you save with the thin tubing.
Did you have issues with assembly? Any access holes at the bottom to help assemble the geared drive?
The tubing is 1 x 4, the gears just roll in from the end. Assembly was surprisingly easy.
Jay O'Donnell
13-12-2013, 16:14
Looks very well designed and made! Mind posting a video when it's driving around?
Very nice!
One speed 3 cim definitely has it's advantages, less parts weight and complexity. easier to predict when driving(less thinking about shifting/waiting for auto shift). At 10fps I can't see you guys popping a breaker unless you were REALLY trying, maybe a push fight against a 2cim shifter going at 6fps. I remember you guys browning out a few times last year at 9fps, is there any way you are looking to avoid these brownouts in the future?
This project probably started before vex did their unveiling. Now that the 3 cim ball shifter is available would you consider it as a replacement for your beloved single speed?
6 Cim's and a battery at the back. May want to distribute weight better. Always beware of Center of Mass.
We typically run only 12 Gauge wire to our motors. We try and standardize on two gauges: 12 and 18. Very generally (there are always special cases) motors will get 12 gauge, and everything else that requires a circuit run from the PDB will get 18 gauge. It's pretty simple for our students to follow, and for them to explain during inspections.
This is the first time we've used Talons, and are very curious about their linearity, and how they'll behave with our software PID control.
Once it's driving we'll likely post a video, but I doubt it'll be much more interesting than any other FRC drivetrain doing laps (http://youtu.be/ybyGcg7u7Cw).
The brownouts from our 2013 robot were traced back to a faulty 12v->5v DC-DC converter for the radios. We didn't catch this until we were on Einstein when the field support crew informed us that our radio (not cRIO as we originally thought) was losing power. We replaced it at some point on Einstein as we were placing the robot on the field for a match - I don't remember during which match. After that, the brownouts ceased.
Our drivetrain choices are driven by game strategy requirements. I wouldn't rule out a multi-speed gearbox in the future, but there would need to be very good reasons. We optimized last year for multiple 36ft start-stop sprint between the feeder station and pyramid, and for getting around defenders. Optimizing for, say, Overdrive in 2008 was a different matter completely. I don't think our team has ruled out anything, but we always try and find the simplest and most reliable solution to achieve the objectives that our Strategy team sets. Being able to build something that's maintenance free is also a big deal.
If the drivetrain strategic objectives were the same as last year, I think it's safe to say we'd stick with the single speed. In fact, we might've gone down to 4 CIMs, as the 6 CIM decision was somewhat a product of our power-takeoff climbing time requirements.
The weight distribution is definitely biased for just a drivetrain, however I don't lose any sleep over it. We developed this drivetrain with the intent that we hope to actually use a similar design on a full FRC robot someday. There's room if we need an open front for an intake, and the top-half of a robot is usually weight biased as well. We actually didn't know which end of our drivetrain would be the front of our robot until close to week 3 last year - and this gave our programmers fits! We intend on testing this drivetrain at full competition weight at some point, so ballast will be going on, and the weight distribution will be adjusted at that point.
Once it's driving we'll likely post a video, but I doubt it'll be much more interesting than any other FRC drivetrain doing laps (http://youtu.be/ybyGcg7u7Cw).
Some dubstep (http://www.youtube.com/watch?v=0EC0T7UF_OE&feature=youtu.be) will make the video more awesome.
What are you expecting that is different with Talons regarding PID control? I believe that you guys mentioned that you were using CAN last year. I'm not too familiar with it, but if I recall correctly it does some of the framing for you regarding PID control. I guess that part will be different.
However, we used PID control last year for our drivetrain running 4 Talons. Proved to be accurate within 3 inches over a 100 inch drive if the count-to-inch conversion was accurate, and that was with little tuning on the drivetrain ("Test mode" only works for one motor at a time, and we relied upon it for rapid tuning). For FRC purposes, I can't imagine that non-linearity is that significant between, for example, a Talon or a Jaguar running CAN.
This is the first time we've used Talons, and are very curious about their linearity
I can't imagine that non-linearity is that significant between, for example, a Talon or a Jaguar...
If it helps, there is actual test data available here:
http://www.chiefdelphi.com/media/papers/2720
Joe Ross
16-12-2013, 12:40
http://i.imgur.com/EAyf9Fu.jpg
It should be driving soon!
You've probably fixed this by now, but your cRIO modules are not in the correct slots.
I assume you purposely wired the Digital Sidecar and Radio the way you did? The 5v supply on the PDB has a worse dropout then the 12v supply.
Rob Stehlik
16-12-2013, 15:11
This project probably started before vex did their unveiling. Now that the 3 cim ball shifter is available would you consider it as a replacement for your beloved single speed?
I doubt it. In my opinion, two speed transmissions are silly and largely unnecessary. In most robot games, the vast majority of your movements will be short sprints from game piece to goal, and avoiding other robots in between. This calls for acceleration, not a high top speed. Slower is faster.
Kernaghan
16-12-2013, 19:24
This is really cool. I've been CADing some very similar designs (putting gears inside the rectangle tubing), so I have a few questions. What type of gears are you using, aluminum, steel, what manufacturer? Also gears do not absorb force like belts and chains do, how are you planning on mounting the bumpers so the forces from collisions are dispersed correctly? The drivebase looks awesome, I can't wait to see what 610 comes out with this season.
JohnFogarty
16-12-2013, 19:56
Where do you buy your 1x4 tubing?
Kernaghan
16-12-2013, 20:09
Where do you buy your 1x4 tubing?
You can get it here (https://www.metalsdepot.com/products/alum2.phtml?page=aluminum%20rectangle%20tube&LimAcc=%20&aident=), stock number is T34118
waialua359
17-12-2013, 02:18
I doubt it. In my opinion, two speed transmissions are silly and largely unnecessary. In most robot games, the vast majority of your movements will be short sprints from game piece to goal, and avoiding other robots in between. This calls for acceleration, not a high top speed. Slower is faster.
Over the years based on our experiences, I find this very true.
However, when we had our software controlled auto transmission in a previous year, we found that to be highly effective......except we scrapped it when our roll pin from the dog gear went sour in matches. We didnt last an entire tournament without having issues with it.
The first time I saw shifting was 2001 (I think) at the SVR regional when 254 used low gear shifting to balance on the bridge much easier. It was very quick that year. EJ drove......maybe he can comment if it was worth it.
DampRobot
17-12-2013, 02:39
Over the years based on our experiences, I find this very true.
However, when we had our software controlled auto transmission in a previous year, we found that to be highly effective......except we scrapped it when our roll pin from the dog gear went sour in matches. We didnt last an entire tournament without having issues with it.
The first time I saw shifting was 2001 (I think) at the SVR regional when 254 used low gear shifting to balance on the bridge much easier. It was very quick that year. EJ drove......maybe he can comment if it was worth it.
I'm kinda lolling at the thought of EJ driving a robot in a FRC competition.
I used to share this thought, but have found through experience that the idea that slower top speed=faster isn't necessarily true. My current belief is that you should optimize your acceleration by reducing the mass of the robot as much as possible, and as a last resort, adding more power. Limiting your top speed to improve your acceleration is a risky proposition at best. We geared our robot too low in 2012 (where conventional wisdom would say that you needed acceleration a lot more than speed), and we really found it limited our performance. Sure, we accelerated quick, but it took forever to go anywhere.
This offseason, we played with an approximately 90lb bot (not including bat and bumpers), and geared to go around 18fps with a WCP DS. Simply put, the acceleration and speed were fantastic, even compared to a bot only 30lbs heavier. 18fps adjusted was a good speed for short runs around defense too. We got heavy d played on us all through quals, and we only really started shifting in elims. That robot flew, and it was because it was light and we geared really fast.
My opinion tends to be that for robots with a decent center of gravity, speeds of 16+fps are reasonable for most games (with a shifter). Of course, others have different well supported ideas about the "best" top speed.
MichaelBick
17-12-2013, 07:19
I'm kinda lolling at the thought of EJ driving a robot in a FRC competition.
I used to share this thought, but have found through experience that the idea that slower top speed=faster isn't necessarily true. My current belief is that you should optimize your acceleration by reducing the mass of the robot as much as possible, and as a last resort, adding more power. Limiting your top speed to improve your acceleration is a risky proposition at best. We geared our robot too low in 2012 (where conventional wisdom would say that you needed acceleration a lot more than speed), and we really found it limited our performance. Sure, we accelerated quick, but it took forever to go anywhere.
This offseason, we played with an approximately 90lb bot (not including bat and bumpers), and geared to go around 18fps with a WCP DS. Simply put, the acceleration and speed were fantastic, even compared to a bot only 30lbs heavier. 18fps adjusted was a good speed for short runs around defense too. We got heavy d played on us all through quals, and we only really started shifting in elims. That robot flew, and it was because it was light and we geared really fast.
My opinion tends to be that for robots with a decent center of gravity, speeds of 16+fps are reasonable for most games (with a shifter). Of course, others have different well supported ideas about the "best" top speed.
I agree with this thought. 6 motors is just so easy now, especially if they keep even half of the motors available. If you have the motors I see little reason in not going 6 motor drive.
DampRobot
17-12-2013, 10:02
I agree with this thought. 6 motors is just so easy now, especially if they keep even half of the motors available. If you have the motors I see little reason in not going 6 motor drive.
Tripping the main breaker is a serious concern with a 6 CIM drive.
Andrew Lawrence
17-12-2013, 10:13
I agree with this thought. 6 motors is just so easy now, especially if they keep even half of the motors available. If you have the motors I see little reason in not going 6 motor drive.
Michael, Damp's argument is that he'd rather shift on a light bot than add more motors to the drive.
MichaelBick
17-12-2013, 11:48
Michael, Damp's argument is that he'd rather shift on a light bot than add more motors to the drive.
Ok, I misunderstood. Regardless, 6 CIMs is overkill unless you are running a PTO. Otherwise I would max out at 4 CIM + 2 Mini CIM.
cadandcookies
17-12-2013, 17:07
Perhaps this discussion of acceleration versus top speed should be split off...
Regarding top speed versus acceleration, I think it's important to remember what (I believe it was BJC) said in a whitepaper: the goal of a drive train is to get you where you need to be, when you need to get there. To that end, I'd focus on time to go a given distance: a very nice calculator for this can be found here (http://bl.ocks.org/schreiaj/raw/6604718/81de5905aa072e9666bee0c0e8f06fc207f7a970/), which was written up by Andrew Schreiber (based on the drive train acceleration model, which Ether provides a link to in the next post). It's a little bit out of date but it's good for giving reasonably accurate numbers.
Other than that, it's all about team preference. Obviously there's some math behind it (if it's going to take you longer than 25 feet or so to reach your top speed, it probably isn't worth it), but nothing that hasn't already been mentioned elsewhere.
I'd focus on time to go a given distance: a very nice calculator for this can be found here (http://bl.ocks.org/schreiaj/raw/6604718/81de5905aa072e9666bee0c0e8f06fc207f7a970/).
That calculator was written by Andrew Schreiber1, and was based on an early version (now superseded) of the Drivetrain Acceleration Model2.
The model was updated a few days later to correct a small error and include the effects of reduced motor voltage caused by circuit resistance3. Andrew's calculator has not yet been updated to reflect the updated model.
Full C source code2 is provided for the Model. All you need is a C compiler and you can change the parameters, compile, and run it. Compiling and running takes less than 2 seconds. The output is a CSV file which will open directly in Excel if you set file associations. Columns in the CSV file include time, distance, speed, acceleration, motor current, and motor voltage.
1 See post#13 in this thread http://www.chiefdelphi.com/forums/showthread.php?t=119442
2 http://www.chiefdelphi.com/media/papers/2868
3 See post#23 in this thread: http://www.chiefdelphi.com/forums/showthread.php?t=119442
cadandcookies
17-12-2013, 18:10
That calculator was written by Andrew Schreiber1, and was based on an early version (now superseded) of the Drivetrain Acceleration Model2.
The model was updated a few days later to correct a small error and include the effects of reduced motor voltage caused by circuit resistance3. Andrew's calculator has not yet been updated to reflect the updated model.
Full C source code2 is provided for the Model. All you need is a C compiler and you can change the parameters, compile, and run it. Compiling and running takes less than 2 seconds. The output is a CSV file which will open directly in Excel if you set file associations. Columns in the CSV file include time, distance, speed, acceleration, motor current, and motor voltage.
1 See post#13 in this thread http://www.chiefdelphi.com/forums/showthread.php?t=119442
2 http://www.chiefdelphi.com/media/papers/2868
3 See post#23 in this thread: http://www.chiefdelphi.com/forums/showthread.php?t=119442
Sorry, I didn't mean to post that without credit. I've edited my post to reflect the information you've presented.
At this point that model is something that I use quite often for getting a reasonable estimate of time to a given distance, and it just sort of permanently hangs out in my "Useful FRC tabs" bookmarks group. In the future I'll be using the up to date source code for critical applications.
At this point that model is something that I use quite often for getting a reasonable estimate of time to a given distance, and it just sort of permanently hangs out in my "Useful FRC tabs" bookmarks group. In the future I'll be using the up to date source code for critical applications.
Just remember the old adage "Garbage in, garbage out". The model is only as good as the parameters you use: the more difficult ones being the constant, speed-dependent, and torque-dependent drivetrain losses (Kro, Krv, Kf); and the circuit resistances (Rcom and Rone).
cadandcookies
17-12-2013, 19:14
Just remember the old adage "Garbage in, garbage out". The model is only as good as the parameters you use: the more difficult ones being the constant, speed-dependent, and torque-dependent drivetrain losses (Kro, Krv, Kf); and the circuit resistances (Rcom and Rone).
Unfortunately I have yet to find an online C compiler that will give me the output csv file, so for back of the napkin calculations it will have to do.
I did not fully appreciate how much of an obstacle the need for a C compiler would be.
For those who do not have, or do not care to install, a C compiler, I will look into posting a compiled executable that grabs the parameter values from the environment.
http://youtu.be/lQLp_F88CEg
We had a chance to take the drivetrain for a spin last night. We rang out the encoders and gyro, and all seems to be in good shape.
Our driver for the upcoming 2013-2014 season had his first crack at driving it around, and the initial feedback is very positive.
BrendanB
18-12-2013, 18:38
http://youtu.be/lQLp_F88CEg
We had a chance to take the drivetrain for a spin last night. We rang out the encoders and gyro, and all seems to be in good shape.
Our driver for the upcoming 2013-2014 season had his first crack at driving it around, and the initial feedback is very positive.
Wrong video? The cars are pretty sweet though!;)
Wrong video? The cars are pretty sweet though!;)
Try clicking it again. There was some YouTube weirdness going on with the link.
Andrew Lawrence
18-12-2013, 18:42
http://youtu.be/lQLp_F88CEg
We had a chance to take the drivetrain for a spin last night. We rang out the encoders and gyro, and all seems to be in good shape.
Our driver for the upcoming 2013-2014 season had his first crack at driving it around, and the initial feedback is very positive.
Looks solid. Something I noticed is your CG is between your back two wheels, so all of your turns spin around that back point, negating one of the useful effects of the 8wd and also preventing you from turning in place with almost no displacement of the robot.
Akash Rastogi
18-12-2013, 18:48
Looks solid. Something I noticed is your CG is between your back two wheels, so all of your turns spin around that back point, negating one of the useful effects of the 8wd and also preventing you from turning in place with almost no displacement of the robot.
I'm quite certain 610 understands how weight distribution works :p Also, check again where the raised wheel is.
Edit: One can only assume this was quite intentional, for whatever reason.
Andrew Lawrence
18-12-2013, 19:05
I'm quite certain 610 understands how weight distribution works :p Also, check again where the raised wheel is.
Edit: One can only assume this was quite intentional, for whatever reason.
Ah, didn't see the raised wheel part. That makes much more sense. I was confused for a minute as to why 610 would have their CG in the back on a standard center dropped 8wd.
BrendanB
18-12-2013, 19:08
Looks solid. Something I noticed is your CG is between your back two wheels, so all of your turns spin around that back point, negating one of the useful effects of the 8wd and also preventing you from turning in place with almost no displacement of the robot.
With the smaller robot size placing your gearboxes on one end allows you to keep a nice wide electronics area open in the robot without having to break it up. Having a small square robot you won't notice this off center turning as much and depending on what is build on top of it, the weight can distribute out. Our 2013 was weighted towards the back due to where our hanger was connected to our robot and we had both our gearboxes and battery as far back as we could. Even then when we hung on the tower we were the front end of our robot hung lower. Our driver had no issues or complaints and it was one of the smoothest 6wd robots I have driven. Also, when using gears to drive your robot, adding in wheels 7 & 8 is extremely easy so as an off-season robot it makes a nice test platform.
For those who do not have, or do not care to install, a C compiler, I will look into posting a compiled executable that grabs the parameter values from the environment.
Beta release posted (http://www.chiefdelphi.com/media/papers/2868). Constructive feedback would be appreciated.
Nathan4567a
18-12-2013, 23:51
Fantastic job! I love the nice and neat cable runs. Our team needs to have clean electronics like that this year. Good choice with the talons too. I love the fact there is no annoying fan on them
z_beeblebrox
30-12-2013, 22:42
^Reported
This is pretty cool. I agree with the whole shifting is overrated thing, and the rear-turning 8 wheel drive thing is pretty neat. Where do you get the fiberglass for the bellypan?
on a totally unrelated note-that spam post 2 above this one is crazy. It has the word movie in it 445 times.
Tem1514 Mentor
31-12-2013, 08:59
I sure that you guys will show off that bot during the quick build as it will give all the new teams a good idea of what may be done outside of the KOP :)
I have my bag already packed with those must have tools ready to help all our new teams get a good start on the season with a working drive train.
I sure that you guys will show off that bot during the quick build as it will give all the new teams a good idea of what may be done outside of the KOP :)
I have my bag already packed with those must have tools ready to help all our new teams get a good start on the season with a working drive train.
Yes! I actually plan on using this drivetrain for the Java Workshop at Quick Build. I'll be doing code for the robot from the ground up starting from a bare laptop with no tools installed.
The goal is to progress to a driveable base, and a simple autonomous by day's end - and to touch on any additional topics the Rookie Teams are interested in learning about.
brennonbrimhall
01-01-2014, 20:24
Yes! I actually plan on using this drivetrain for the Java Workshop at Quick Build. I'll be doing code for the robot from the ground up starting from a bare laptop with no tools installed.
The goal is to progress to a driveable base, and a simple autonomous by day's end - and to touch on any additional topics the Rookie Teams are interested in learning about.
Would you mind recording it? Something like that would be an awesome resource for all.
Would you mind recording it? Something like that would be an awesome resource for all.
The Quick Build workshop is really a group of mentors helping Rookie teams get their tools setup, programming and downloading their first program to a robot, etc. It's over 8 hours long, and is just a "get work done" session where we veteran mentors help out. It doesn't make for good recordings, nor is it all that effective to watch the recordings after the fact.
We've tried to record it in the past I believe, and I don't think anyone was willing to watch 8+ hours of a lot of different teams working on their robots.
Sorry!
brennonbrimhall
02-01-2014, 15:16
The Quick Build workshop is really a group of mentors helping Rookie teams get their tools setup, programming and downloading their first program to a robot, etc. It's over 8 hours long, and is just a "get work done" session where we veteran mentors help out. It doesn't make for good recordings, nor is it all that effective to watch the recordings after the fact.
We've tried to record it in the past I believe, and I don't think anyone was willing to watch 8+ hours of a lot of different teams working on their robots.
Sorry!
Thanks for explaining -- I obviously misunderstood. Best of luck at the workshop!
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.