4607 Butterslide Drivetrain V2

[Ginger Power]

Here (https://www.chiefdelphi.com/forums/showthread.php?t=149307&highlight=4607) is the first iteration. I decided to go a different direction this time given the suggestions received in the linked thread.

Here (http://imgur.com/a/YgJ7t)* is the second iteration.

Specs:
68 lbs including pneumatics, and all electrical components. This number is before any pocketing is considered.
Everything can be built with simple power tools plus a mill (lots of milling...).
10 Wheels including six 4" Omni wheels and four 2.5" custom wheels fabricated from Delrin.

In omni mode: the drivetrain has a 14:64 gear ratio which equates to about 16.5 ft/s Adjusted Speed according to the JVN Calculator.

In traction mode: There is an additional 18:30 reduction which, along with the difference in wheel size equates to 6ft/s Adjusted Speed.

The strafe wheels are pneumatically actuated, powered by a cim motor, and geared at 12:64 which is around 14 ft/s Adjusted Speed based on the JVN calculator.

Improvements over Iteration 1:
- The strafe wheels are now pneumatically actuated so as to main in constant and consistent contact with the ground
- I hid basically everything, maximizing space for other components
- The new frame is much easier to mount things to
- No weird moments to worry about as nothing is cantilevered... or double cantilevered...

Huge shout out to Aren Hill and 3928 Team Neutrino. A lot of this design was inspired by their 2013 drivetrain.

What am I missing? Where am I going wrong? What can I improve? Thanks for the feedback!

*Sorry for the lack of a quality render. It's 3AM and I'm bad enough at rendering when I'm wide awake.

[Ari423]

A quick calculation with the JVN calc says that even with a conservative estimate of the CoF (I can't imagine it's below .6) I get 70A per motor in omni mode. According to the snap action breaker spec sheet (http://files.andymark.com/MX5SpecSheet.pdf), it should hold for somewhere between 2.2 and 9.2 seconds before cutting off. Unless I'm missing something, you might want to take another look at that.

[DaveL]

The gearing seems fast, especially for the transverse wheel.
Otherwise very nice.

Looking forward to hearing how it works out.
Dave

[Ginger Power]

A quick calculation with the JVN calc says that even with a conservative estimate of the CoF (I can't imagine it's below .6) I get 70A per motor in omni mode. According to the snap action breaker spec sheet (http://files.andymark.com/MX5SpecSheet.pdf), it should hold for somewhere between 2.2 and 9.2 seconds before cutting off. Unless I'm missing something, you might want to take another look at that.
I believe this calculation is for when the motors are stalled. With a butterfly drive, there's no reason to get into a pushing match in omni mode. Any pushing would be handled with traction wheels which will take much longer to pop a breaker. With that said, I know 16.5 fps with a single reduction and 4 cim drive is pretty fast, but is it really too fast?

[Ari423]

I believe this calculation is for when the motors are stalled. With a butterfly drive, there's no reason to get into a pushing match in omni mode. Any pushing would be handled with traction wheels which will take much longer to pop a breaker. With that said, I know 16.5 fps with a single reduction and 4 cim drive is pretty fast, but is it really too fast?

True that 70A is for when the motors are stalled, but don't forget that the motors are stalled whenever they are moving and the robot is not. That includes when the robot starts moving from a standstill and after collisions. The motors are in super-stall when you go from full forward to full reverse.

It ultimately comes down to a question of your factor of safety. I personally wouldn't go above 50A because I don't trust my drivers to not crash into things when they're driving around at 16ft/s or to not go from full reverse to full forward when they're under heavy defense (when you would be using the slide drive). You may trust your drivers more than I do, that's up to you. If you do choose to go with those fast speeds, I would highly recommend some serious voltage ramping and code that auto-shifts to traction drive when the robot starts then shifts back once it gets moving.

[notmattlythgoe]

Any worries about having traction issues on the slide wheels while having 6 wheels in contact with the carpet at the same time?

[GeeTwo]

Any worries about having traction issues on the slide wheels while having 6 wheels in contact with the carpet at the same time?

If you've built a slide drive, it should be because you do not anticipate doing a significant amount of strafing - no pushing wars, no driving a great distance sideways. If you do anticipate such heavy use, you should use a more isotropic drive system, such as mecanum, kiwi, or Killough (or raise and take it to crab or swerve).

Note that with this setup, you will not be able to rotate cleanly with the strafe wheels engaged (meaning on the carpet and either powered in the same direction or in brake mode). If you have not already planned for this, I suggest controlling the strafe cylinders separately from the tank cylinders. This is a minor advantage of a single strafe wheel near the CoG vs one at the front and one at the rear.

Edits:
But, how about the question I asked...
Either I didn't understand the question, or you didn't understand the answer. How was my first sentence not a solid "no" to your question?
have you ever seen a 148 butterfly/nonadrive drive?
Not in action. If I read your tone correctly, it sounds like an even more solid "no" to Matthew's question.

[notmattlythgoe]

If you've built a slide drive, it should be because you do not anticipate doing a significant amount of strafing - no pushing wars, no driving a great distance sideways. If you do anticipate such heavy use, you should use a more isotropic drive system, such as mecanum, kiwi, or Killough (or raise and take it to crab or swerve).

Note that with this setup, you will not be able to rotate cleanly with the strafe wheels engaged (meaning on the carpet and either powered in the same direction or in brake mode). If you have not already planned for this, I suggest controlling the strafe cylinders separately from the tank cylinders. This is a minor advantage of a single strafe wheel near the CoG vs one at the front and one at the rear.

But, how about the question I asked...

[bkahl]

If you've built a slide drive, it should be because you do not anticipate doing a significant amount of strafing - no pushing wars, no driving a great distance sideways. If you do anticipate such heavy use, you should use a more isotropic drive system, such as mecanum, kiwi, or Killough (or raise and take it to crab or swerve).



have you ever seen a 148 butterfly/nonadrive drive?

[Ginger Power]

But, how about the question I asked...

I guess I wasn't terribly concerned with traction, but maybe I should be? I think it should be fine, but I'll definitely check into it, thanks!

[notmattlythgoe]

Either I didn't understand the question, or you didn't understand the answer. How was my first sentence not a solid "no" to your question?


My questions was more directed at the ability to even strafe. It doesn't matter how much you plan on strafing, when you want to do so you still want it to be efficient at it.

[GeeTwo]

My questions was more directed at the ability to even strafe. It doesn't matter how much you plan on strafing, when you want to do so you still want it to be efficient at it.

Then apparently a bit of each...

Having six wheels on the ground can be better for strafing than if you just had five, because you can put a larger portion of the robot weight on those wheels easily and consistently.
As an example: Assuming roughly equal loading on each of the wheels - 1/3 of the robot weight will be on strafe wheels for the 6 wheel vs 1/5 for a single strafe wheel, allowing 66% better strafing traction, other factors equal. As you increase the proportion of weight on the strafe wheels, the ability to drive forward/back may be impaired, especially if the CoG is not directly over the strafe wheels, but to the left or right. Unless there were some unusual game/strategy requirements, 1/3 seems a reasonable maximum strafe loading for slide; if you need much more, you should probably consider a more isotropic drive train*.

Of course, if you only put 1/10 of the robot weight on the strafe wheels (no matter how many or how few), expect those wheels to spin if you step on the throttle.

* [edit] Not to say that a great team couldn't make slide work with a much higher loading, but for most teams, I would anticipate an exercise in frustration above this level.