H Drive
 

I Was wondering if anyone could explain to me what an H-Drive is. Also, it would be great if you could add a couple pictures or videos. Thanks

Re: H Drive
 

4 omni wheels in a regular 4-wheel drive configuration, and 1 omni wheel in the center of the robot perpendicular to the other wheels in a simple form.

See:

Re: H Drive
 

Thanks! So all the wheels are directly driven I was thinking using chains on the sides and directly drive the middle, is that a good idea or should I stick with directly driving all of the wheels?

Re: H Drive
 

No problems with that. In fact, I would say that it is better, because if the front 2 wheels lift off the ground, you will still be driving your back wheels with 2 CIMS of power. Same goes for if the back 2 wheels leave the ground.

Re: H Drive
 

Thank you paul i will try and post a picture once I create it in Autodesk to see what you think

Re: H Drive
 

Chaining or belting each pair of wheels with the same center lines automatically distributes the necessary extra torque to the wheel which you're accelerating away from. If you have a CIM for each wheel, it also appropriately redistributes the torque and power of the two CIMS to the two wheels.

Getting the weight load on a strafe wheel correct can also be tricky. This year, we modified a KoP drive train into an H, and ended up pulling the strafe module to minimize weight. Try to make the strafe wheel so that it will carry a defined load, either through a spring or pneumatic cylinder. Using a cylinder also allows you to retract the strafe wheel, which better supports climbing ramps and so forth.

Re: H Drive
 

It automatically distributes the extra torque to the wheel with greater traction, once the one with lesser traction reaches its traction limit. Which wheel has greater traction depends on magnitude and direction of acceleration, location of center of mass, and wheel tread material (if wheels are not indentical).

Re: H Drive
 

Here is a video of us testing our 8+1 wheel H-Drive. We later scrapped it for a different set of wheels altogether, but this was our first time strafing with it. Yup. Nine wheels.

https://www.youtube.com/watch?v=2rAjU9ncg8w

Re: H Drive
 

You can find full CAD files and a Bill of Materials for an entirely COTS H-Drive here:

http://www.vexrobotics.com/vexpro/examples-guides

Re: H Drive
 

It's probably worth noting that to make an effective H-Drive you really need to have the center wheel on some kind of suspension otherwise its very hard to get traction.
We found that out the hard way this year (our first time building an H-Drive), even after dropping the center wheel 1/16" we still only had enough traction to strafe about 75% of the time.

Re: H Drive
 

I don't know much about this drive train but, I'm assuming it isn't a good drive to have in games like Ariel Assist because in that game, drives needed a good amount of traction? Thanks. Just wondering.

Re: H Drive
 

Hello, I'm from 3641, we did a different kind of H-Drive. Our system was called a deca drive, this consists of 2 CIMs driving a center gear in the middle of a seesaw type system with two small omni wheels on the ends, connected by gears to the center gear being driven by the CIMs. When power is applied either way, the system throws that side down into the ground (just from the motor's torque) and spins the omni wheel at the same time. If we switch the direction of the motors, the seesaw simply tilts to the other side. When the system is not being used, the seesaw balances even with the bottom of our robot, allowing us to easily go over the scoring platforms. This deca drive worked very well for us in Recycle Rush, and thanks to our four main drive wheels being omnis as well, we were very maneuverable.

Pictures of ours are attached, and a video of robowranglers with a similar system is right here: https://youtu.be/oRwvQ2D_B-c?t=34s

Re: H Drive
 

Our system, which I explain in my last post, was able to move up and down as far as it needed to get enough pressure to move the robot, whether we were on carpet or foam or solid concrete.

Re: H Drive
 

Thanks for the renderings! I like that the swing is clutched to the cluster axle, which causes the wheel on the side that you're accelerating away from to touch the carpet, which should improve traction. I do have a few questions about some of the details:

• Did you have trouble overshooting your strafes? There does not seem to be any way to properly brake in the strafe direction.
• The gearing looks rather fast for strafing in Recycle Rush; counting of gear diameters, I get about 8:1 or 9:1. Did you actually use strafing as a means of travel, or just to line up for pickups and scoring? Unless you needed the top speed, you could have made the output cluster gear smaller, pulled the wheels pulled in to match, and probably done as well with one CIM.
• How do you control the coupling between the seesaw and the cluster shaft, to get the proper wheel loading? Is there a friction clutch between the output cluster gear and the seesaw, or is it just in the bearings, or what?
• Is there a spring, or does the seesaw return to the horizontal when not driven just based on gravity? Given the amount of friction needed in that bearing, the short offset of the CoG from that shaft, and the large moment of inertia of the seesaw, gravity doesn't appear to be enough.

Re: H Drive
 

Not necessarily; it just depends on how you want to play the game. As you may be able to tell from the pic below, 33 played 2014 with just 4 omnis and did just fine. Didn't even need powered strafe.

They went with the "run from your enemies and be too fast for them to catch you" route.

Re: H Drive
 

I'm not entirely sure how 'correct' our method is, but we ran an H drive this season and had to somewhat scramble to put it together since we were so far behind. Our solution to this was to put a very small drop on our center wheels which we adjusted by turning a few bolts until we were happy

Re: H Drive
 

Sounds similar to what we did. Unfortunately the way we designed the support structure for our center wheel didn't allow a lot of variability in the placement, we basically had to slot some mounting holes and add a shim to get the offset we needed. Lesson learned. :rolleyes:

Re: H Drive
 

1st Answer: No, the robot was heavy enough that as soon as we dropped power to our system, the robot stopped, if we were going too fast that it might have slid, we just slammed the opposite direction for a split second which brought a dead stop.

2: We used it mainly as a lining up tool, but occasionally to actually move far distances across the field, and yes, it was very fast. The system was better used as a quick way to move into perfect lineup with the human station. And about the one CIM thing... I didn't say it my last post, but in order to gain more weight for other things, we did take a CIM off, it does work, however, pretty much the exact same with one CIM or two. The only difference is in speed, and that's not much.

3. There is no friction clutch in the system, it's just the bearings. I attached different pictures to illustrate.

4. The bearings allow the system to be balanced only on gravity, as the bearing are just a pivotal point. Most of the weight is under the pivot, so gravity is the force used to balance the seesaw.


Thanks for the questions - keeping me on my game during the off season :D I'm happy to answer any more questions you have too.

Re: H Drive
 

Thanks for the answers! By the way, the first and third renderings answered questions 3 and 4 better than your text. I had pictured the module rotating about 20 to 30 degrees to make contact with the floor; the actuality is more like 70 or 75. This greatly reduces the torque which must be applied to the module to achieve a given contact (and therefore traction) force.

Also, to put a new subthread about H-drives, or more particularly about omni wheels: The renderings shown here depict classical omnis with ~18 nearly cylindrical rollers per duallie wheel, which are available from both AndyMark and Vex, and probably others. AndyMark came out with a "Duraomni" wheel this year, which has eight longer, more (american) football shaped rollers for an integrated dualie. 3946 used two non-driven 6" idle omnis for Aerial Assist, and five 4" Duraomnis for Recycle Rush, but we didn't do anything that could serve as an "apples to apples" comparison. Did anyone do any comparison tests between the different types and is willing to share results?