

Reliability counts. Inconsistency makes great robots lose events. Murphy's law knows no bounds.  ratdude747 [more] 



Thread Tools  Rating:  Display Modes 
#1




Chain Tensioning
Okay, I have been trying to use 1114's method of not using chain tensioners. In their "Drivetrain Design" presentation it states if you use the biggest sprockets possible that you do not need to tension the chains IF the distance between the the centers of the two wheels is a multiple of the chain pitch. BUT, I have gone online and used some chain distance calculators (http://www.botlanta.org/converters/d.../sprocket.html and http://www.rbracingrsr.com/chainlength.html) and when I put in a multiple of .375 for a center to center distance, it does not come up with an even number of chain links as 1114's presentation stated. Is this correct? Should I use what the calculators are giving me or what 1114 stated in their presentation? I emailed Karthik already and he forwarded the message to their design lead and I haven't got a response yet. I am very confused on which center to center distance to use. The calculators or 1114's concept of multiples of .375. Well If anyone can help, Thanks in advance!

#2




Re: Chain Tensioning
OK, let's imagine we have a chain that has a pitch of 0.375 inch. That means each link is .375 apart, as are teeth on a sprocket...but there are even and odd links (or 'innie' and 'outie' links). So really, a full "link" is 0.750. Get that?
So let's say you want an Integer (whole, not fractional) number of links. If we have a sufficiently large sprocket, we can say that: 1. The number of links on 1/2 of the sprocket is an integer (one lnk per tooth, right?) 2. If we imagine a point at the very top of the sprocket, it should coincide with the center of the sprocket. So for any chain going from top to top of two sprockets, if the centers are an Integer number of links apart, there must be an Integer number of links between the tops. And the bottoms. 3. If we have an integer for the sprocket half and an integer between the tops and bottoms, that all adds up to an Integer. See it? I went to the first calculator and used 2 sprockets of 40 teeth each and a center distance of 37.5 inches. Using the above, I would expect 2 * 20 links on the sprockets and 100 links top and bottom, a total of 240 links...and that is what the calculator says it is. So the calculator seems to be working for me.... Last edited by DonRotolo : 06232013 at 05:19 PM. 
#3




Re: Chain Tensioning
In addition to what DonRotolo said, the guideline you're referring to doesn't work unless the two sprockets have the same number of teeth, like in a wheel to wheel connection. Are you looking at a chain that goes from a smaller sprocket on a gearbox to a larger sprocket on a wheel? If so, you can still use your center distance calculator to figure out a suitable center distance that gives a chain length that is a multiple of 0.750".
Last edited by Nemo : 06232013 at 05:22 PM. Reason: worded incorrectly 
#4




Re: Chain Tensioning
Quote:
I get what you mean though. So if I were to use the .750 for "pitch" in the calculator the number of links needed would be accurate If I am using #35 chain if I am trying to not use tensioners? 
#5




Re: Chain Tensioning
Quote:

#6




Re: Chain Tensioning
We don't tension chains either. We use the center to center calculator in inventor to give us a starting point, then as the chains stretch, we put spacers underneath our gearbox to tighten the chains. We played > 35 matches this year on the competition bot, and we never had to adjust, but after a weekend of driver practice, it helps to put some spacers under the gearbox. Try to keep the chain runs as short as possible.

#7




Re: Chain Tensioning
Quote:

#8




Re: Chain Tensioning
Here is a tool from Dr. Joe that you could use. We used it for our arm actuators* and did not experience any problems**. The interface is not beginner friendly though, and unless you figure how to use the goal seek feature in excel, you will be playing a game of guess and check to get a good distance.
Alternatively, you can use this tool from Paul Copioli. It is much more user friendly, but isn't as powerful as Dr. Joe's tool. Also, the only person who has confirmed it actually works (as far as I know) is Paul Copioli himself, but I personally would be inclined to trust him. Or, if you are too macho for these wimpy excel sheets, Paul Copioli provides the actual formula he uses in the his calculator: Quote:
Not so macho now, huh . *given that the cc distance was only ~4in and we were using it at 8:00PM on bag and tag day, so we weren't that picky. ** except for that 1 time where a bolt head got into the path of the chain and the master link eventually snapped, but I doubt that this was a problem with our cc distance. 
#9




Re: Chain Tensioning
OK, 15t and 32t, no problem.
If we assume a "long" chain, the wrap on each sprocket is about 50%, and the formula still works. Let's assume a 'short' chain though, cc around 4" maybe? 15t gets 7 links, not 7.5, because it doesn't wrap all the way around. YOU can calculate the exact number. If you don't know the geometry, try doing it graphically  draw it to scale on paper. 32t gets just a hair more than 16, call it 17. 7+17=24, a nice even number. If we pick a distance as a multiple of 3.75, it would almost be perfect, except the chain is at a slight angle, so the chain needs to be a hair longer...or the cc needs to be a bit shorter... than normal cc. So if I pick 3.75", I'd go with 3.7 and accept a little looseness or 3.65 and struggle installing the chain. Again, YOU can do the math to figure exactly how much longer 10 links will be at an angle of (need to figure that too...) The whole point is, if you can do the math, it's just geometry and trigonmetry. If you can't, or need help, draw it out to scale on paper. (a 15t sprocket can be depicted with a (15*.375)/Pi diameter circle). Measure the angles or calculate them, both are perfectly acceptable methods. And, in your head, know the answer will be just a bit short of (N*.375), so if your method gives such a result, you can know you are spot on. 
#10




Re: Chain Tensioning
Team 1640 uses this calculator by our head mentor, which automatically finds the actual correct wrap around each sprocket (or pulley), and finds the closest correct (integer link) distance for any given input. Just put in your 2 tooth counts and a ballpark of what you want the CC distance to be, and it'll spit back the true answer.
It's been great for us, though being swerve drive we don't use long chain runs very often. We've never had a CC chain issue on our swerve modules in 4 years, and we've just shy of 90 matches this year to date. It also works perfectly for timing belts, which we have used over long distances. 
#11




Re: Chain Tensioning
Quote:

#12




Re: Chain Tensioning
Quote:
If you go exact cc, you're actually going to be under that number by whatever slop exist in all the holes and shafts in the system. 
#13




Re: Chain Tensioning
Quote:
From my experience, the actual calculated CC distances are exactly where you want the chain to be, but due to slop between the bearing and the frame, the hex shaft and the bearing, and the hex shaft and the hex bore in the sprocket, you can have a decent amount of slop that reduces the effective CC distance. 
#14




Re: Chain Tensioning
You could always consider using belts. Much less headache that way, and they're more efficient/make less noise, to boot!

#15




Re: Chain Tensioning
My only real issue with belts is that they're thick. They take up a lot of valuable horizontal space on the robot that can be used for mounting electronics and such. Also, Vex doesn't (at least yet) have GT2 profiles for VersaPulleys (I'm absolutely addicted to the VersaHub). We could probably get them manufactured, but I'd rather use our machining resources for other stuff.

Thread Tools  
Display Modes  Rate This Thread 

