Nonadrive Concept

[Sam Skoglund]

Hello CD Community!

I am relatively new and didn't know how to post pictures to chief Delphi, but here is a link to a Google drive folder with pictures and details about a nonadrive concept that I designed.

For those of you who may not know, a nonadrive has nine wheels. 4 of them are traction wheels and 5 are Omni in an h-drive configuration.

Any feedback would be much appreciated!

https://drive.google.com/drive/folders/0B2gZV0m2wg2LTjFZMDhKeTJBRjA

[ollien]

Might wanna take a page out of 148's book ;) (https://www.youtube.com/watch?v=6fLf71xlVhE)

Jokes aside, this looks awesome. I really like the idea of the folding bumper mount a lot.

[Cothron Theiss]

I really recommend you check out the Grasshopper Hybrid drove that 624 initially showcased.

https://www.chiefdelphi.com/media/papers/2968?

Also, just search for Grasshopper drives on here and you'll find several threads with a lot of great information. I'll comment on a few things, but a lot of those threads have information that would help you as well.

Your speeds are high. Are those the theoretical or adjusted values? If you aren't, I'd suggest you use the JVN Design calculator to determine your actual speeds when driving. Remember that your "Weight on Driven Wheels" will not always be 100%. That can make a big difference.
Also, it looks like you've got a fixed strafing wheel in the middle. With n-wheel or actuating wheel drives, you especially need to have some way to actuate your strafing wheel so you can actually get a decent amount of weight on that single wheel.
I'm not a big fan of the way you've attached your actuators to the modules. It looks like there could be a lot of flex in the system, especially if the clevis rod doesn't slide along the slot as smoothly as you expect. Were that to bind mid-shift, you could end up damaging the modules if the pneumatics have enough power.
I like the profile of your gearboxes. Compact gearboxes are a wonderful thing.

Keep working on this design! It's a great way to test out a lot of concepts needed for building great robots.

[Sam Skoglund]

I really recommend you check out the Grasshopper Hybrid drove that 624 initially showcased.

https://www.chiefdelphi.com/media/papers/2968?

Also, just search for Grasshopper drives on here and you'll find several threads with a lot of great information. I'll comment on a few things, but a lot of those threads have information that would help you as well.

Your speeds are high. Are those the theoretical or adjusted values? If you aren't, I'd suggest you use the JVN Design calculator to determine your actual speeds when driving. Remember that your "Weight on Driven Wheels" will not always be 100%. That can make a big difference.
Also, it looks like you've got a fixed strafing wheel in the middle. With n-wheel or actuating wheel drives, you especially need to have some way to actuate your strafing wheel so you can actually get a decent amount of weight on that single wheel.
I'm not a big fan of the way you've attached your actuators to the modules. It looks like there could be a lot of flex in the system, especially if the clevis rod doesn't slide along the slot as smoothly as you expect. Were that to bind mid-shift, you could end up damaging the modules if the pneumatics have enough power.
I like the profile of your gearboxes. Compact gearboxes are a wonderful thing.

Keep working on this design! It's a great way to test out a lot of concepts needed for building great robots.

Thanks for the reply! The speeds are theoretical, but not from the calculator. I will take a look at the calculator. Also it is hard to see in the pictures but the entire cim/gearbox/wheel strafing assembly pivots, and would be pressed against the ground with springs or rubber bands.

[Ginger Power]

I think it's a great start! 80 lbs seems pretty heavy for just a drivetrain plus electronics. There definitely looks to be a lot of weight that could be cut. Is this something that your team is looking to build, or just a thought experiment?

[Sam Skoglund]

I think it's a great start! 80 lbs seems pretty heavy for just a drivetrain plus electronics. There definitely looks to be a lot of weight that could be cut. Is this something that your team is looking to build, or just a thought experiment?

Our team lacks a lot of manpower and doesn't have a lot of access to good machining tools, so it was just more of a design idea. That being said, if the game this season fits this style of drivetrain well, we might revise it and attempt to build one.

I think the weight could be cut down quite a bit from some sort of lightning hole pattern, I just wasn't motivated to CAD model such a thing!

[Ginger Power]

Our team lacks a lot of manpower and doesn't have a lot of access to good machining tools, so it was just more of a design idea. That being said, if the game this season fits this style of drivetrain well, we might revise it and attempt to build one.

I think the weight could be cut down quite a bit from some sort of lightning hole pattern, I just wasn't motivated to CAD model such a thing!

I've put a lot of time into nonadrives/decadrives or as I like to call them, butterslide drives. I will be presenting about FRC drivetrains at Jumpstart at St. Cloud State University. It'd be awesome if your team could make it there!

I think it would be really cool if you modified this design to be built with your team's resources. Adding constraints like that make these designs much more challenging and in my opinion fun.

[CalTran]

Any reason for Nonadrive over Octacanum?

[Sam Skoglund]

Any reason for Nonadrive over Octacanum?

I chose the omni wheels over the mecanum because it would allow for faster and smoother speed in all directions. I designed this in the summer as sort of a "best possible" all around drivetrain. If we decided to build something like this, I would reconsider the Octanum option because it is easier to manufacture.

[echin]

Some thoughts:

I'll echo the above comment about actuating the center wheel to get more down force. My FTC team is currently using an H-drive and having a center wheel with variable down force has proven to be very important for smooth driving.

Also, it might be a good idea to reinforce the drop down wheel modules a little more. They look like they might bend easily from a side hit with the traction wheels down.

Overall, this looks really good. I'm curious to see how it develops!

[dradel]

interesting setup. would love to see a similar concept based on a kop chassis

[DaveL]

Nice Job.
As others have posted, you might run into problems with your pneumatic cylinder. I would mount the cylinder with a bolt as a pivot to the back (bottom) of the cylinder. That way it can align to the wheel strut's movement path.

Can anyone talk about how a bot like this turns? For example, are the turns sharp and crisp or wide and slushie?

Dave

[roboruler]

interesting setup. would love to see a similar concept based on a kop chassis

The KOP chassis is designed to be a simple and reliable system for teams. Trying to do this with the KOP chassis would be far more work then it is worth, just use tube/Versaframe.

[dradel]

going to have to disagree... there are many teams that use the kop chassis, and have done different drive types using the chassis.

plus the inspiring of teams to think outside of the box is priceless! let alone the experience gained in cad by doing such a project

[Knufire]

The KOP chassis is designed to be a simple and reliable system for teams. Trying to do this with the KOP chassis would be far more work then it is worth, just use tube/Versaframe.

The AM14U3 is significantly more versatile that one might assume. A slide drive shouldn't be too difficult with a couple TB Nanos and some extra inside rails.

[dradel]

The AM14U3 is significantly more versatile that one might assume. A slide drive shouldn't be too difficult with a couple TB Nanos and some extra inside rails.



Our team did a H drive using the kop chassis for recycle rush. We actuated the center wheels with pneumatic cylinders so the robot would be able to cross over the scoring platform.

[Cothron Theiss]

I took another look at this design since my last reply was from my phone, so I couldn't really see anything then. I'm gonna be a bit more critical this time around so you can correct these things early on in the design process, but if you have an explanation for something or think I'm completely wrong about something, please let me know. I think you've gone in an interesting direction with this, and it could end up being a slick design. You're trying to pack a lot of stuff into a very short profile.

First, I've taken the liberty of calculating your speeds of the specs you provided in the attached document. Screenshots of the spreadsheets are on imgur (http://imgur.com/gallery/n1c5K). I assumed that the omni wheel configuration would be your primary configuration because if your pneumatics were to fail, you would be stuck in omni configuration, though your traction wheels could very easily drag on the ground if your cylinders were to lose pressure. You can fix this with a spring or surgical tubing of some sort to actually lift the traction wheel out of the way in case of some mishap. (Don't forget to bump up the force your cylinders exert accordingly.)

I would really really really recommend changing the way your cylinders actuate your modules. Anyone can disagree with me on this, but the way the clevis rod is intended to slide along the slot in the angles during actuation SCREAMS like trouble waiting to happen to me. The reason it scares me is, when the cylinder actuates, the force the cylinder exerts on the inside of the slot will be almost perpendicular to the direction you want the clevis rod to travel.
Instead of rigidly mounting the cylinder and expecting the connection to the module to take all the movement, I would find a cylinder with a universal mount and pass a bolt through the side plates of the chassis and through the rear pivot of the cylinder. Then, your connection from the end of the cylinder rod to the module can be rigid and your entire cylinder can pivot. That's just what popped into my head first, but you can fix that issue however you like.

I see 32 hex shaft collars. It'd be cheaper and lighter to drill and tap the ends of your shafts and use a screw and washer to fix that ends, with spacers cut to length to determine spacing. In fact, I'd argue that all of your hex collars could be replaced, but I can see why you'd want the ones constraining the strafing wheel so you can change that wheel's placement.
Similar to that train of thought, why are there six sprockets on your strafing wheel drive and not just three?

You mention your strafing wheel being pressed against the ground with springs or tubing or what have you. Does this mean that you are constantly putting weight onto this wheel? This might be an issue if you get into a pushing match or are going over obstacles. If you're in a pushing match, any force the strafing wheel exerts on the ground is lessening the normal force on each of your traction wheels, which means you'll not get as much traction. Also, your strafing wheel might not like going over uneven terrain perpendicular to its primary axis of rotation. Many h-drives handled the scoring platform just fine, but anything more rough than that might give you issues if you can't retract that strafing wheel. That's just a trade-off, so it's entirely up to you.

I would look for ways to increase the rigidity of your chassis at the exact center plane, the mid-plane between the mini-CIMs. I might be wrong about this, but I imagine if you took a section view of your assembly at this mid-plane, you might see less material than you need.

Is the block between those CIMs your battery? If so, good on you for CADding it early. That's something I've forgotten to do and it cost me. But I'd consider where you're going to route your terminals. The battery's got to connect to something for it to work (no ballast!).

As you reconsider some of the things on your design, don't be afraid to abandon some of the perfect symmetry present in a lot of your design. Symmetry can be an excellent design tool to follow, but understand where it's holding you back or adding extra constraints that you don't even notice.

[Sam Skoglund]

I took another look at this design since my last reply was from my phone, so I couldn't really see anything then. I'm gonna be a bit more critical this time around so you can correct these things early on in the design process, but if you have an explanation for something or think I'm completely wrong about something, please let me know. I think you've gone in an interesting direction with this, and it could end up being a slick design. You're trying to pack a lot of stuff into a very short profile.

Thank you so much for the detailed feedback Cothron Theiss! I will talk about your topics one by one here. If you or anyone else could answer some more questions I have that would be great!

Speed: Our team has a history of building speedy drivetrains, so my view of an "average" speed is most likely skewed. If you think that I have geared it too fast, what speeds would you recommend?

Cylinder Actuation: I agree fully that I need to revise the attachment of the cylinders to the wheel modules. Mounting them as you suggested seems like the best option. Would it work to keep them mounted as they are and connect them to the wheel module plates with a small linkage that pivots on both contact points?

Hex Shaft Collars and Sprockets: My team is addicted to these. I may eliminate some and opt for threaded shafts, but only if it means easier production and or a better final product. Price is not a factor, as we already own well over 32 of these! The extra sprockets are there because we already own that many and if one strips, the chains could be rearranged to new sprockets without disassembling the entire gearbox.

Strafing Wheel: In terms of a pushing match, I don't think it would be a bad thing to have the strafing wheel constantly pushing on the ground. The talons controlling the strafing motors could be programmed to enter brake mode whenever the drivetrain is in traction mode, so that when hit from the side, the strafing wheel would act as a brake or could even be driven to push back slightly. I think the actuation would be something to consider if the field is uneven.

Chassis: I assumed that the sheet metal bottom would provide enough rigidity, is this not correct?

Battery: Yes that is the battery. The routing of the battery wires, motor wires, and pneumatic tubing is something I have actually planned out. The battery cables will rest above the strafing module to reach the PDP. There is enough room between the strafing module and the upper sheet metal panel to do this. The pneumatic tubing and motor wires will route through a hole in the main tubing and through the drive chains (will have to protect the from shredding) to get to the electronics.

Questions I still have:
- How many motors and of what type should I use? In the current setup, while driving diagonally in omni configuration, 6 cims and 2 minicims would be running simultaneously. This seems like overkill to me.
- Does anyone know if it is possible to machine custom gearbox plates with little machining equipment? (The best tool we have access to is a mill without DRO)

[lorykzarr]

One thing you may be interested is this white paper by GeeTwo on torque actuated strafe wheels. It's similar to 148's 2014 design in that the only thing pushing the strafe wheels to the ground is gearbox spinning them. It might prove to be simpler then adding some kind of pneumatic or spring system.

https://www.chiefdelphi.com/forums/showthread.php?t=139928

[Cothron Theiss]

Thank you so much for the detailed feedback Cothron Theiss! I will talk about your topics one by one here. If you or anyone else could answer some more questions I have that would be great!

Speed: Our team has a history of building speedy drivetrains, so my view of an "average" speed is most likely skewed. If you think that I have geared it too fast, what speeds would you recommend?

Nope, your speeds are fine. 15 ft/s is fine if you're viewing it as a high gear. 20 ft/s is getting a little risky.

Cylinder Actuation: I agree fully that I need to revise the attachment of the cylinders to the wheel modules. Mounting them as you suggested seems like the best option. Would it work to keep them mounted as they are and connect them to the wheel module plates with a small linkage that pivots on both contact points?

If you can make it work, it'll work. Not meaning to be asinine, but you'll just have to play around with the design and see. One issue you might run into with using a linkage like you described is making sure the distances are always the exact same. You wouldn't want to increase the pressure of your cylinders and have the height change.

Hex Shaft Collars and Sprockets: My team is addicted to these. I may eliminate some and opt for threaded shafts, but only if it means easier production and or a better final product. Price is not a factor, as we already own well over 32 of these! The extra sprockets are there because we already own that many and if one strips, the chains could be rearranged to new sprockets without disassembling the entire gearbox.

Yup, I understand that. I've used a gear as a spacer just because it's what I had on hand and because Week 6 is insanity. But 80 pounds for a drivetrain is WAY too much in my opinion, so I'd be looking for any and every way to cut down on weight that you can. Also, a screw and washer is a more robust solution in many situations.

Strafing Wheel: In terms of a pushing match, I don't think it would be a bad thing to have the strafing wheel constantly pushing on the ground. The talons controlling the strafing motors could be programmed to enter brake mode whenever the drivetrain is in traction mode, so that when hit from the side, the strafing wheel would act as a brake or could even be driven to push back slightly. I think the actuation would be something to consider if the field is uneven.

Yup, it's probably a bad thing to have the strafing wheel exerting force against the ground. Driving the strafing wheel or setting its controller to brake mode will only help when you're being pushed from the side, and you aren't going to win any pushing matches with just one wheel. You'll want to evade and outmaneuver in that situation, not push. If you're being pushed from the front, you'll most likely lose.
Any force your strafing wheel exerts on the ground is weight not being translated into normal force onto your driving wheels. The normal force on the driving wheels and the co-efficient of friction between your driving wheels and the carpet are the two major factors of how much traction you have. If you want to win pushing matches, you need to maximize those two factors.

Chassis: I assumed that the sheet metal bottom would provide enough rigidity, is this not correct?

Battery: Yes that is the battery. The routing of the battery wires, motor wires, and pneumatic tubing is something I have actually planned out. The battery cables will rest above the strafing module to reach the PDP. There is enough room between the strafing module and the upper sheet metal panel to do this. The pneumatic tubing and motor wires will route through a hole in the main tubing and through the drive chains (will have to protect the from shredding) to get to the electronics.

Good on you.