Hey, I would love some feedback on my design.
15.82 fps free
6.00:1 2 cim 1 minicim
Some questions:
When doing a belt like this, the path length dimension for the whole belt is just the (number of teeth)*(pitch) right?
When the belt is given as a length e.g. 400mm, is that measure from the outside diameter of the belt loop, the inside, or the diametrical pitch?
Just want to confirm: the diameter of each pulley should be increased by the (machining tolerance * 2)?
When looking at what material the standoffs should be made of, do I use the modulus of elasticity for determining if the standoffs will bend? Any recommendations for standoff material/diameter on this one?
Why is the WCD gearbox bearing block 0.98”? Shouldn’t it be 1.00” to reach through the versaframe?
Anyone have any tips for belt tensioning? There was some talk about eccentric bearings and locking collar bearings on the discord but just wondering what do most of you guys use?
Is a snap ring what most people use on the drive shaft in front of the encoder?
If you have done a belt system like this before, are there any thing you wish you had known before building it?
The vex bearing blocks are slightly undersized so they will clamp against the tubing. Too much extra space will bend the tubing, but not enough will lead to slop.
You might want to consider compressing the gearbox so that the CIMs stick out over your frame rail and wheel. With proper bumper mounting and construction you won’t risk them being damaged. This could save you a bit of space in the center of the bot.
Many times they will slot one of the motor mount holes. That looks like the easiest thing to do in your case. You may want to slot one hole on each of the outside cims to make getting the belt off and on easier.
192’s method in 2014 is my favorite way to tension a drive belt. Look up their 2014 drive gearbox.
How did you pull off the path length dimension for you belts? You are exactly right that the length should be teeth*pith.
Why do you believe this is the case? To be clear, I agree, but I like to see discussion on this topic and I’m sure OP would appreciate a more specific explanation to improve their future design decisions.
Is there anything stopping you from making cutouts for the CIMs in the outer (output shaft side) plate and reducing overall thickness? It looks like there is a lot of wasted space between the two plates.
Cool to see more onshape! Im curious how you went about designing your belt path? Without a path length dimension it seems impossible to design something like this 100% in onshape.
I did the path length dimension in Solidworks and transferred the dimensions to Onshape :P. Is 192 using the Cim holes to vary the tension?
Hmm, maybe this could also be used for tensioning? I would be worried about the motor rotating if the only thing keeping it from doing so is the friction between the screw head and the plate.
I was thinking about that, however I was worried about wire strain (If we had to do tight bends) and if I wanted to mount something to that versaframe, however, now looking at it, I might not mount anything there to avoid interfering with the wires anyways.
My feed back is to ask why couple your motors to the shaft using a belt rather than gears? What design advantage does this offer?
At first glance I see some big disadvantages:
The belt is being wound back and forth around multiple small pulleys, which is going to drastically reduce the life of the belt. You are flexing it back and forth, putting first the inner surface, than the outer surface in tension.
There are also many finicky little parts, and lots of cantilevered shafts. Lots of potential points of failure.
The biggest problem I see is that it concentrates all your drive load in a single belt. When three motors use small spur gears to drive a single larger gear, each spur gear mesh carries 1/3 the resulting drive load. Here, the tension in the belt increases as it passes around each of the motor pulleys, then drops very low as it leaves the big pulley. Three motors with spur gears have reaction loads which tend to cancel out. Here, you have large reaction loads on all those little pulleys, and especially your main shaft that require significant structure to keep the gearbox from twisting out of shape.
Did you calculate how much bending load is at the root of the most highly loaded idler shoulder bolt? What is the maximum tension in your belt?
I would also suggest changing some of your mounting points for how the gearbox connects to the drive rail. Right now it looks like you only have 2 10 screws holding the entire gearbox. Also your bolts for what I assume screw into the blue thread spacers are way too long if you do end up threading that spacer. An easy work around would be to drill a clearance hole on the outside face of the drive rail. Then mount a screw head that rests on the inside face which is a clearance hole for a 10, which goes through a clearance hole in your small gearbox plate into the threaded spacer. This will add another mount point so you are not just relying on two fasteners to hold the weight of your gear box.
Check out WCP tutorials on designing with belts- its meant for Solidworks but the offsets you need for the belt thickness is in there. You need to add an offset to account for the thickness of the belt above or below the pitch line of the belt (that’s probably not the right term). If you have an idler on the backside of the belt (for 3mm GT2 belts) this offset is 0.88 mm on the radius. You should check this yourself and not take my word for it. You can dig through the Gates belts catalog and find the specs.
In the past I’ve used the technique of delrin idlers with a bearing and just turning the delrin down until belt tension is awesome and then making the rest of the idlers that dimension. Not the fanciest of design solutions but also a lot easier than path lengths in onshape.
I’m also curious what the intended advantages of this system are compared to gears for all the reasons Todd listed above.
First of all, thank you for the detailed response.
Tell me if I’m wrong, but wouldn’t belted gearboxes be able to have less backlash than a geared gearbox because of the ability to tension the belt, they would also be quieter and more efficient. These were the main advantages I was going for here. Again, I could be wrong about this. I also like how it is able to make more space for reductions below the cims (since they can be farther away from the “spur” pulley on the first stage), this would allow me to be more versatile with gearing (this would allow gearing for both 4" and 6" wheels). Being flipped not only saves space in the middle of the robot, but it also allows 2 pulleys to be put on the output shaft which will be able to go to the front and rear wheels, allowing the gearbox to be mounted in the middle of the robot and saving lots of space.
I have not done calculations for the loading on the belt. Soon I will read over the Gates manual which has some equations and tips for calculating this. Right now the tensioners are in place just so that I could maximize the engagement of the belt with all of the pulleys. I am using a Gates HTD 5mm pitch 9mm wide belt as opposed to a normal GT2/GT3 belt, I would think that it would be able to hold up well, again, I will have to do calculations on this.
I don’t really understand, what reaction loads are you talking about that would bend standoffs?
Tell me if I might be wrong in what I’ve just said. I’m always trying to learn :).
