Rather than hijack the discussion about Art’s very nice sheet metal chassis here, I thought it might be better to have a new thread about driving Colson wheels.
These wheels are available from a few sources, including McMaster-Carr (pg. 1347) and robotmarketplace.com and are well regarded for their wear characteristics. They come in a variety of sizes.
Can people talk about their coefficient of friction with carpet as compared to roughtop or wedgetop tread?
They appear to come with a polypropylene hub. Are the delrin bushings they’re furnished with from McMaster-Carr removable? How about the bearings?
I don’t know as much about these wheels as I should and want more options for wheels available to me. I’ve been considering what it’d take to, say, press this hub from AndyMark into this wheel with its 1 3/16" bore. You could then match drill and bolt on a sprocket.
We used Colson wheels from 2005 through 2008 of varying diameters and widths with great success. The standard design was to machine our own hubs (slightly oversized with the outside surface knurled) and then press them into the wheel bore. The hub bore was for 1/2" shaft and broached for standard keyway. (http://joemenassa.com/images/robotics2006/day10/images/IMG_3452.jpg) Never had a problem with the hubs or the wheels (other than after plenty of field time, they will wear down, so make sure you have enough clearance underneath). We used to always run live axles with this setup, so i’m not familiar with concocting a setup to run these wheels dead axle.
I really have no experience with the combination you are proposing. This is just my opinion.
Without any modification, you would have an interference fit of .0125". That will be a tight fit, with a penetration depth of approx. .82", that is just 55% of this wheels width. This might be sufficient, but will it be enough? I am not familiar with polyolefin, so I don’t know if it could handle a press fit of that nature.
This would be a great experiment to try in the off season, but I’m not sure I would try it during build season unless you have a tried and true, bullet proof solution, as a backup.
If you do decide to try it, would you please share you findings? It sounds quite interesting.
Theoretically dead axle should be even easier to setup because you don’t need to worry about your knurled insert stripping out.
I would take that same insert you guys make, but add a flange on the end with a hole pattern (very similar to the AM hub Madison linked to), and then just make sure you can press bearings into that hub. If space starts to become a factor, you could use needle bearings to ensure they fit inside. The flange then allows you to mount whatever sprocket/timing belt setup of your choice to the wheel. Slide the whole assembly onto a dead shaft and your done.
I like your idea to install an OTS bearing housing, this will eliminate the need for a load-transmitting press fit. You could then install the wheel/bearings on a shoulder bolt (my personal favorite) and shim it into position with OTS spacers.
McMaster PN 91259A636 and page 3232 respectively. You could also get larger OTS spacers or shaft collars if so desired.
Definitely! That hub was just a design by one of the team leaders and we never bothered changing it because we never had problems. I’ve gone to dislike live axle setups though. Especially when to replace a wheel required us to nearly take apart the entire chassis. A bit of poor planning perhaps on our part, but eh…oh well. That’s what makes them the good ol’ days.
We’ve been sticking with the AndyMark wheels for the past few years…I’ve found them just plain easier to work with. But with a new game comes a new design, so you never know…
Unmodified Colson wheels have a coefficient of friction of about 1.1-1.2. In last 2009, we did some tests with Colson wheels and determined cutting thread in them increased their coefficient of friction by 30% on a similar carpet to the FRC playing field (I want to do better tests to verify these results, but subjective evidence from watching our 2010 robot drive around confirms this). We really like these wheels, and will likely continue using them whenever possible.
A hub diameter of 1.20 is pressable into the wide Colson wheels if you have a beefy arbor press. We’ve always used live axle setups with the 1.5" or 2" wide Colson wheels on 228, since the sides of these wheels aren’t very flat and I’d be worried about the drill bits wandering. When I was on 190, we used 7/8" wide Colson wheels, and those were much easier to drill a three/six hole pattern into.
The one time I used Colson wheels was on this robot. I made hubs similar to the above, but I put a keyway and key between the hub and wheel also. I thought I had them pressed in, but I got a few that started sliding out. They weren’t knurled though. It drove really nicely and smoothly on those wheels. However, it got pushed around by robots with wedgetop incline conveyor belting tread. In my experiences, the coefficient of friction on carpet was more in the range of 0.8 to 0.9, if I remember correctly (it was nearly 5 years ago). I think they were the 7/8" wide by 3.5" diameter ones.
I once tried machining the hub down flat on a Colson 6" wheel, and found air-pockets/voids inside the plastic, that were not visible from the outside prior to machining
The problem with the Colson wheels and other variants is that the outer skin is dense. As soon as that outer layer wears away the under lying material is softer and wears fast. At a certain point the wheel needs to be replaced. AndyMark plactraction wheels offer a good alternative with the tread being replaceable. However we have found that we can’t get through a regional with out replacing the conveyor belt treads. It’s not desirable to have to replace tread in the heat of competition. The larger size Andymark wheels have been Known to break under high impacts. Like crossing the bump aggressively. This fall our team looked at different better wearing treads for the 4" plactraction wheels we use. We have found that bike tires work well. The key is finding the right tread compound and tread pattern. There are so many different tires that finding the right tire is not easy. We have also found a good way to mount the tire treads to the Andymark wheels. The raw tires are too thin and too stretchy to mount directly to the wheels. No matter what tire a team chooses remember that too grippy a wheel is just as bad as on with out enough traction.
For those that have used Colson wheels:
The hub bore size is listed as different sizes on mcmaster than on therobotmarketplace.
What i belive to be happening is, mcmaster inclues bushings pressed in, and robotmarketplace does not. Are the bushings from mcmaster removable in the event you wanted to create your own?
Then, will the hub bore be the same as listed from therobotmarketplace?
