How much backlash is recommended for gears? I am using AM gears (20 dp, 14.5 pa) and we have the option of cutting the side plates on either an old mill or a waterjet. The mill is by no means state of the art and I would guess that the tolerance is .005". The waterjet tolerance as listed on their website says .005" but our mentor assures me that they can get at least .003" if not much better.
How much backlash would you recommend between two gears? I was thinking .002" but am not really sure if that will be enough.
Also, what are some good places to get hole reamers for .005" under 7/8" and 9/8"? I am planning on having the waterjet cut a smaller hole than needed for the bearing and then pocketing that hole with a reamer so I can press fit an unflanged bearing in there.
thanks, Vivek
EDIT: err… It seems that reamers are meant to go all the way through the hole. What is the best way to make a hole like that? I am planning on using .25" aluminum and leaving a lip of .13-.15" for the bearing to sit against. Gimme a break, I’m still a n00b.
You have a couple other options–all of them involving second operations on a mill. You can use an endmill and interpolate a pocket for the bearing to sit in. If you wanted to use a manual mill, you could use a boring head and bore the diameter down to the flat.
It’d be a heck of a lot easier to use a flanged bearing and a through hole.
as for reams, check McMaster, Enco, KBC Tools, etc. Pretty much any industrial supply/tooling company will have perfectly suitable reams.
You don’t want 0.005 under either. More like 0.0005. a 0.005 interference fit is WAY too tight. you want a max of about 0.002 interference. You might be ok going 5 thou under in aluminum, but you could actually deform the bearing in a harder material.
Honestly the best place I have found to buy bearings is on ebay. The store is VXB wholesale I believe and they have very cheap prices compared to most places.
I might just stick with the local hardware store’s bearings since they are cheaper. With the run time of a FIRST bot I doubt there will be problems. It is not really a high load situation anyways.
Anyhow, I never did get an answer on how much backlash to put in. I was thinking like .01"? The gear teeth are .1" long and the machine we are using to cut the plates will have .003" tolerance at worst.
It is ok to put in none and then just run them in. At least I’m pretty sure it’s ok. It is not the “best” method but it accomplishes the same goal and you don’t have to worry about to much backlash and power loss.
I would speculate that you want your center to center distance increased by at least the amount of tolerance of your machine to ensure your gears do not bind. If your machine can only hold .003, I’d say either look for a better machine, or increase your center to center distance .005 over nominal. I’d be weary of going with any more than that though.
For the bearing fit in aluminum, you certainly don’t want any more than .001, with something in the range of .0005 being optimal. More than probably .0015 and you’ll start to excessively load and potentially even damage the bearing. There are formulas to calculate this, but this is a basic rule of thumb based on experience.
Make sure you use precision ground ABEC rated bearings. If you don’t, you’ll certainly have problems. Unground and/or non-precision bearings are not rated for anywhere near the speeds, loads, or tolerances of precision ones, by a factor of up to several-fold. I too would recommend through holes and flanged bearings. The most common bearings to use in FRC gearboxes are FR8ZZ and FR6ZZ. These can be obtained from a variety of sources, including Small Parts at about $4 each, under part numbers BRF-06 and BRF-08. There are also cheaper sources. Check local suppliers.
yeah, our machine does atleast .003" (better from what our mentor tells me). We are trying out a waterjet from our sponsors for the first time. They list the tolerance as .005" on their website but my mentor assures me that they can do much better than .003". I made my previous designs for .002" of backlash but it is no big deal to change.
Ok, I will look into flanged bearings. They certainly would make life a lot easier. (Come to think of it, I have half the bearings I need from a toughbox I [slash]demolished[/slash] err… modified :))
If you really want to learn what the right demensions are use a machine design book (norton or shigley) or you can buy an old machinist Handbook. Since you are using spur gears any old addition should have the right info.
As far as tolerance goes, almost any mill that has functional lock on it can hold a tolerance tighter than 0.005 learn how to work against slop and practice.
That being said our team made AL side-plates for an AndyMark gearbox. It took two tries. The first shot I merely explained what to do with the students. They made a couple big set-up errors and the plates wouldn’t work. I worked closely with them to make the second set and it worked out better.
Rule of thumb on press fits 0.001" per 1" diameter. This has almost always worked for me in the past. Gearboxes are a good test of skill. They hate being too tight and will destroy themselves if too loose.
I hope to learn more about machining over the summer. Our mechanical team knows how to run the mills but to different extents (and we use a digital readout that holds .0005" of tolerance, although not sure how accurate it is).
I will probably switch to using those flanged bearings.
I am using hex key turned down at the ends for all of my shafts and was wondering how good of an idea it is to turn down hex shaft like that. I would think that all those edges banging against the bit can’t be good for it (although if I came in from the side that would be solved. This is what I have done before but the AM hex shafts already had an end turned down onto them.).
thanks,
-Vivek
EDIT: Is an arbor press absolutely necessary for press fit bearings? Could the same effect be achieved without too much trouble with a vice or some blocks of wood and a hammer? I am not against getting one, it seems like they would be quite useful. It is just that we can’t really afford luxuries right now :).
You can make a pocket in your gears on a lathe, and tolerances to .001 or better should be easy with just a little care. .001 interference in aluminum is a pretty tight fit, .0005 for a 1.5" OD bearing would be plenty. You can fiddle this a bit with sandpaper or a file.
For clearance between teeth, you must be sure there is no negative clearance (too close together). The higher the speed, the closer the tolerance, the harder the material, the closer they can be. For moderate speeds (1000 RPM-ish) in aluminum, .0005-.002 is fine, be sure to run the gears in for several hours with light lubrication, in both directions and light load.
When making intermittent cuts (like turning hex to round) don’t use carbide bits, regular tool steel will manage that better. Carbide tends to shatter, especially the ones with the little triangular inserts.
An arbor press is convenient, but not required. The reason to avoid using a hammer is the shock loads being transmitted. Better use a vice or some clamps with metal blocks than a hammer, but if unavoidable, go slowly with the hammer.
Cheap bearings are OK for low speeds (600 RPM-ish) and sloppy tolerances. ABEC bearings are great for higher speeds (10,000 RPM) and longevity, as well as tighter tolerances. I buy bearings on eBay all the time, but if you have a bunch of useable bearings, use them.
Old trick if the pockets are accidentally a loose fit: Make a plug for the hole from scrap metal, then drill three (or more) evenly-spaced 1/8" or so holes in the edge of the bearing pocket, with the holes protruding maybe 0.010 into the pocket (and hole plug). Fill the 1/8" holes with metal rod, file them down on the inside a little bit to get your press fit.
I believe the only bit we have ever used for our lathe is a carbide. There must be some tool steel ones laying around somewhere…
That is a pretty nifty trick about the hole plugging! Never would have thought of that.
Our mill has a digital readout that goes to .0005 (with the last digit being either 5 or 0). Now that is 25 times what I estimated the tolerance to be (I believe the digital readout is a Bausch and laumb). I will have to find out the actual tolerance next time I go to our workshop.
If you want an alternative to flanged bearings you can make your own support for the bearing out of button head screws. Here is a picture of how I do it. http://www.chiefdelphi.com/media/photos/16362.
We have typically used a 0.08 mm increase in gear center distance with our custom-made geartrains. My guess is that this is on the high side compared to some others who make their own boxes, but these are typically wire-edm’d gears whose surface finish is a little rough before run-in. Our gears are not very wide either.
We run them in at no load (i.e., not driving the chassis around, but just running the wheel while it is off the ground) for an hour or so. We have been using the same tube of Amsoil grease (apparently its “racin’ grease” :)) for several years now…
If I was using purchased, nicely finished, wider gears, I think something in the 0.05 mm range would be my choice for the amount that the center distance is increased.
The looser distances like I am suggesting here make the boxes easier to assemble and get low geartrain friction, in my opinion.
