30-06-2014 18:39

Tyler2517

Re: pic: New serve mk.1-2517

http://imgur.com/a/H6Qcm

30-06-2014 19:05

magnets

Re: pic: New serve mk.1-2517

That's a really neat swerve drive. I don't think anyone's done something like this before.

A few questions/comments-

I'm assuming the really big green thing that goes in the middle of the wheels is a bearings. If not, what is it? If so, why is it so big?

Those bevel gears- where are they from? Are they strong enough? They're at the end of the reduction so they're going to try really hard to get away from each other. The little one is going to push up so you'll probably want a thrust bearing/washer there too.

The green angle to support the sides has the mounting holes too close to the edge. Making it a little bit wider around the bolt holes is cheap strength.

How is torque transferred between the bevel gear and the wheel?

What type of wheels are those?

Overall, a pretty cool iteration on swerve.

It's pretty neat to see these offseason swerve designs come back. There's usually some pretty cool ones like this!

30-06-2014 19:33

Tyler2517

Re: pic: New serve mk.1-2517

The green thing is a spacer that lets us translate the power from the wheel with the bevel to the wheel with out. It has 12 threaded holes in it so that we can offset the bolts so the load from each one of the wheels so there is a lot less shear force(each wheel has 6 holes.)

6 holes on the gear.

The bevels are from sdtp. We are not using the stock materials and will custom make them our selves if the ones are not strongthat we can buy.
We are looking at making the bigger of the gears out of a titanium alloy most likely Ti-6Al-4V Trading a week of my summer to clean there shop in trade for the machine time and resources to make 20ish on there 5 axis cnc. We do not want 2 of the same material gears so we don't have to grease them so carpet cant get stuck in the grease.

Custom wheel 2.78 in diameter if i remember.
https://drive.google.com/folderview?...sp=dri ve_web

30-06-2014 19:49

cadandcookies

Re: pic: New serve mk.1-2517

That is a very nice deal.

30-06-2014 20:01

Chris is me

Re: pic: New serve mk.1-2517

I'm having trouble picturing this so I can't be sure, but doesn't this just mean each bolt is loaded in single shear instead of double shear? I think 12 bolts going through both wheels would be stronger since the bolts are in double shear. Both are probably overkill.

Why does having two different materials negate the need for grease? Why does having two of the same material necessitate it?

30-06-2014 20:12

Tyler2517

Re: pic: New serve mk.1-2517

I suck at explaining.

The spacer in the middle will have 12 holes threaded in to it. each wheel will have 6 bolt holes they will be threaded in to the middle spacer every other hole so that we don't have large amount of shear force going through the green spacer.
The wheel is really small and trying to get room for 12 bolts is both supper heavy and over kill for a 2.78 inch wheel. We were trying to have no nuts/bolt heads on the out side of the wheel so that we can get the walls of the moduler as close as possible.

Having the 2 different materials will cause much less wear. Just like running 2 steel gears on each other will cause rapid wear. Having the titanium gears running on a brass alloy gears will let us put very little grease on it with very little wear. That way the carpet fibbers will not get cough in the gear grease causing us not to bind our gears like last season.

30-06-2014 20:25

T^2

Re: pic: New serve mk.1-2517

Are you sure you want to be loading screws in shear at all? Why not replace half of them with spring pins?

30-06-2014 20:26

kk052

Re: pic: New serve mk.1-2517

actually the wheels are custum 2.75 inch, and the bevel is from SDP/SI part:
A 1B 3-Y32016B
A 1B 3-Y32064
the grean spacer is threadded on the outer holes, six bolts from the right wheel, six though the left
(im the CAD lead, i would know)

30-06-2014 20:29

kk052

Re: pic: New serve mk.1-2517

also the brass on titanium is to make the brass take all the wear, so we dont need to replace the titanium

30-06-2014 20:45

EricH

Re: pic: New serve mk.1-2517

All the wear, as in "you better replace 'em often". Brass isn't known for wearing well--in fact, it's better known for deforming when harder materials apply force to it (to the point where when a hard material is getting beat up, and brass is suggested as a replacement, engineers start asking "Are you sure about that?" and "Is there another material that will work?"). Try bronze instead--it's a little harder, and to some extent can be made somewhat self lubricating (bronze bushings, NOT brass bushings). Titanium is already really hard, harder than most FRC gears are, which will tend to increase the wear on the brass gears even more than, say, aluminum/aluminum interactions or steel/steel interactions.

Regarding the screws: You're loading the screws in shear. (Quick check: Do you know what shear stress is, in broad terms? If not, we'll be happy to explain.) What that means is that the screws are going to want to "cut" themselves apart, probably at the spacer. The tendency is actually less if there are multiple locations for the shear stress to act at, hence the suggestions to run the bolts through, giving double shear instead of single shear. If you're concerned about clearance, there are ways to deal with that, including sinking the holes for the screw heads and nuts so the heads drop right out of the plane of the side plates.

30-06-2014 20:47

magnets

Re: pic: New serve mk.1-2517

Why would you want titanium? It's a worse material for gears than steel. The weight savings would be well under a tenth of a pound. The "good stuff", grade 5 titanium, isn't cheap either.

If wear is so bad that it is an issue for an FRC robot, then you'll experience wear on both gears, regardless of material.

You will need grease regardless of the material you make your gears out of. If you run without enough grease, you'll wreck the gears in no time.

TBH, I haven't done the math, but a cantilevered 32 DP brass bevel gear at the output of the transmission, where the torque is the highest doesn't seem like a good idea.

If you're doing this on CNC anyway, I would get rid of the green thing and try something like the vex pro versakey pattern, or something that's similar to the dog clutches used in FRC gearboxes.

30-06-2014 21:47

Tyler2517

Re: pic: New serve mk.1-2517

http://imgur.com/6orZXtp - The green spacer 12 holes.

http://imgur.com/TskU7rT - wheel with 6 holes.

We will probably change it over to 6 single bolts going all the way through there was a really good reason we did not do this in the first place and now i cant remember why.

For gears steel on steel i think is what we will run for the initial prototype types.

30-06-2014 23:07

asid61

Re: pic: New serve mk.1-2517

I thought screws were meant to be in shear load? 10-32 screws have ratings in the hundreds of pounds range IIRC.

30-06-2014 23:23

EricH

Re: pic: New serve mk.1-2517

It depends on the screw and arrangement. Generally, an individual screw is meant to handle tension loads. However, in the right arrangement, they can handle pretty high-shear loads.

I broke out my machine-design textbook--it's really handy for a lot of things with robotics, including a general screw-bolt-nut overview. (Also one of the few chapters I didn't get into in college.) The answer there is that generally, screws are for tensile loads, but in structure use, they're often used to resist shear. BUT, those shear loads are taken when the screw is already in tension. The suggestion from the book is to use dowel pins in machine-type shear situations to assist the screws.

I've seen what sometimes happens when a lone screw takes a shear load. It's not fun to hear if it's a good-sized screw. Come to think of it, I don't know if we ever did find that one screw/bolt head...

30-06-2014 23:24

T^2

Re: pic: New serve mk.1-2517

Load rating =/= intended usage. Screws will misalign little by little when used as the sole locating feature.

01-07-2014 01:18

DampRobot

Re: pic: New serve mk.1-2517

Would a machine design textbook be a good buy/read for a prospective ME student with little to do this summer? I've heard a lot about this mythical textbook from mentors and on CD, and it sounds like it contains a lot of "real" engineering information. Any suggestions on where to buy it, or what edition (assuming there are multiple editions)?

01-07-2014 01:31

aldaeron

Re: pic: New serve mk.1-2517

The "old one" I have is Kent's Mechanical Engineers Handbook. The copy gifted to me by my first boss is the 12th Edition (1950). It is a good reference, but not an exiting read cover to cover. It has a section on gearing stress and fasteners and many other mechanical things.

A "more modern" design book would be Mechanical Engineering Design by Shigley and Mischke. Those two have a ton of editions of mechanical design. Thanks to indecisive profs I have both the 4th and 6th editions. They are now on the 9th edition according to Amazon. Again, not a cover to cover read, but this book has a section on gearing stress and fasteners.

I caution you that these are 3rd or 4th year collegiate level texts. There are a lot of support classes that will help these equations make a lot more sense (statics, strength of materials, etc). Reading them cold could be very difficult.

-matto-

01-07-2014 02:32

T^2

Re: pic: New serve mk.1-2517

A former mentor gave this version to our team a few years ago. I have no idea how it compares to others, but it's been handy as a reference and learning resource for new designers. All of the math is understandable with the equivalent of Calc BC and Physics C Mechanics knowledge.

01-07-2014 06:08

nathannfm

Re: pic: New serve mk.1-2517

Aren't all of the extremely common dead axle (sprocket/pulley bolted directly to the wheel) drives in FRC putting 6 bolts in shear on each wheel? (we found out what happens when you accidently forget 3 of the bolts...ignore the EI metal )

01-07-2014 09:10

Gdeaver

Re: pic: New serve mk.1-2517

This design effort seams to be focused on mechanical losses experienced in the 2014 swerve module. What specific problems did you have? Have you identified where the losses on the current module are? Have you quantified the total losses? I see some potential problem areas in the current Module. Before I say more, what have you found so far? Know what the 1st iteration problems are before you design the 2nd iteration.

01-07-2014 10:51

Chris is me

Re: pic: New serve mk.1-2517

While this is common and may work, it is not optimal. Screws aren't great in shear due to the stress risers of the threads and the effectively smaller load carrying diameter of the bolt. A better solution is to use 3 bolts for fastening and three shear pins to take the load. Shear pins are better simply because they aren't threaded and thus have a larger effective diameter.

01-07-2014 11:02

Lil' Lavery

Re: pic: New serve mk.1-2517

Why transmit torque through the screws at all? Do something similar to what Vex does with their versahubs.

01-07-2014 11:26

autox_vette

Re: pic: New serve mk.1-2517

Not all of the force is transmitted through the bolts (in an ideal world, none of it would be).

Think about the wheels on a car. They are held on with 4-6 studs in single sheer. Some are taking 500+ HP. The bolts/studs only hold the wheel onto the hub and create a HUGE clamping load between the wheel and the hub. All of the power is transmitted through the friction between the wheel and hub.

So, it depends upon what your wheel and spacer materials are. If you use metal, then less sheer force is transmitted to the bolts because you can get enough clamping force to transmit the power. If you use plastic/nylon, you probably can't get enough clamping force, so some power will be transferred as sheer force on the bolts.

Kenton

01-07-2014 12:41

T^2

Re: pic: New serve mk.1-2517

This used to be the case, and still is for AM products. As mentioned above, VEXPro products rely on the VersaHub to transmit torque; the screws hold purely in tension. Guess which system is more reliable?

01-07-2014 15:03

Chris Endres

Re: pic: New serve mk.1-2517

I understand where you come from with this idea, although the power that the gears with transfer will need grease, no matter what type of material you use. I know that carpet fibers and other junk gets stuck in normal grease, but you can also use dry grease. Machining tools (boring bars/milling disks) use dry grease so chips don't get stuck between important seals. I would look into using dry grease because of the low friction index, such as Dry Graphite Lube (I once used it to win a pine-wood derby).

01-07-2014 15:15

Tyler2517

Re: pic: New serve mk.1-2517

Starting from the top.

Gears- Use steal on steal unless we can prove that aluminum would hold up which i personals will hold reliably.

Screws- switch to having 6 screws going all the way through exerting a clamping force on the spacer in the middle and the gear. This way the screws are not taking a large amount of shear load or no shear load. Would a softer material like nylon work better or would AL6061 work the best?

Thrust bearing on the top bevel gear and not just a thrust washer due to the end conditions of the transmission

On the same manner of thought the reasons for changing designs from last years module. Is primarily to improve year to year in as many ways as possible.

Increasing mechanical efficiency by removing a chain reduction. And maybe reducing weight in the moving parts of the drive transmission(we might end up adding weight not sure on that yet)

Slightly improved traction by increasing suffice area the tread and more importantly 2 leading edges per wheel. This will make our passive defense stronger.

Smaller foot print with in the robot. Alowing the wheels to get closer to the corner still and increasing foot print slightly.(not much but i am a firm believer that a large amount of small changes for the better will eventually make a better product then competition)

Price reduction no longer buying sprockets/gears/bearings. (yes the gears used now might be more spendy but still should be under the price of the old system)

Improved senser feed back adding in primary drive encoders for pid loops. And using a limit switch to auto home the module(things i have not mentioned in this post yet)

01-07-2014 15:16

artdutra04

Re: pic: New serve mk.1-2517

I've used brass gears before in FRC, and I wouldn't ever use them again in torque carrying situations. They were way too soft and wore too easily when interfaced with steel gears - they needed to be replaced fairly often as they quickly turned to dust even when well greased.

There are some alloys of bronze that would be a lot better, but I suggest that you use two of the same material gears and lubricate them well. You do not need to use the gooey grease, there are some good lubricants that form a fine film like this one that won't attract carpet dust:
http://www.mcmaster.com/#8710t35

01-07-2014 19:05

EricH

Re: pic: New serve mk.1-2517

I'll give that a Yes. And, like any college-level course, there are multiple authors and editions, which may be used by various colleges.

I use the same one that T^2 linked, down to edition. If you happen to have some ideas as to where you want to go to college, I'd suggest trying to get one from one of the college bookstores.

01-07-2014 19:08

AdamHeard

Re: pic: New serve mk.1-2517

I'd suggest avoiding purchasing anything at a college bookstore. Much better deals on amazon (especially used), ebay (used/internationl), and smaller companies that have used books.

01-07-2014 19:11

EricH

Re: pic: New serve mk.1-2517

At least looking to see what they have. If they're changing editions, they might be dumping the older edition as fast as possible--read "discount".

03-07-2014 16:16

kk052

Re: pic: New serve mk.1-2517

thanks for the advice ill see about replacing half the bolts with shear pins

03-07-2014 16:26

kk052

Re: pic: New serve mk.1-2517

it was designed to be smaller, lighter but stilll be made cheaply (-ish), the lower number of transfers is a result of deciding to remove the chain and first axle of a traditional swerve,(because steel chain is heavy)
im no expert on mechanical efficiency (that's Tyler's job) but the system should be more efficient if we took out an axle and a chain (once again i could be wrong, im not too good with theoretical mechanical efficiency, if it stalls we just throw another moter on it and/or lower the ratio)

04-07-2014 02:57

Tyler2517

Re: pic: New serve mk.1-2517

Here are the 3D printed prototypes. Super small we made them really dense so we could do performance testing with wheel tuning and gearing on that.
Supper small the whole bottom yoke is smaller then your standard 4 inch wheel.
http://imgur.com/a/spJtM

On a side note you could put this on a ftc robot.... with ease they sell plastic gears identical to the ones we would use and the 3d print with some fine tunning could print something i would put on a robot.

04-07-2014 16:45

CENTURION

Re: pic: New serve mk.1-2517

Doesn't look like anyone has mentioned this yet, so I will. You're going to have a lot of trouble turning this module.

If I understand correctly, both wheels are locked together? When you try to rotate the module while the robot is stationary, one wheel needs to spin one way, and the other needs to spin the opposite, because the center of rotation is between them. And when the robot is moving, and you attempt to rotate the module, one wheel will have to turn faster than the other.

What you need is some sort of differential, so that when turning, the wheels are able to rotate independently, but you can still apply drive power to both.

Here's a great old video that wonderfully explains how differentials work (at least in the context of cars)

08-07-2014 15:34

kk052

Re: pic: New serve mk.1-2517

it was argued about in the design process about the use of a diff, we decided not to to save space and make the unit simpler, and the less things to go wrong the better, also if there are any ramps in the 2015 game and one side of the wheel gets on it, all the power will be given to the other wheel. we are going to have to use a lower ratio to turn it however

09-07-2014 11:25

asid61

Re: pic: New serve mk.1-2517

As long as you use a beefy turning motor (not a window motor) you should be fine.

09-07-2014 15:51

kk052

Re: pic: New serve mk.1-2517

a bane bot 775 (i think) on a 132:1 ratio is what we used this year with no problems, and most likely what we will use on the prototype, if we have any issues we may use a different gearbox, or find a way through code that the wheels never pivot in place, only when the wheels move forward/backward will the assembly turn, but this is what we are prototyping to find out.

09-07-2014 15:54

Tyler2517

Re: pic: New serve mk.1-2517

We used a 430 on a 132/1

09-07-2014 16:09

R.C.

Re: pic: New serve mk.1-2517

A 430?

09-07-2014 16:30

Bryce Paputa

Re: pic: New serve mk.1-2517

Nope I was wrong. 430 is probably a typo, there isn't a banebots 400 series as I assumed.

09-07-2014 16:34

Andrew Schreiber

Re: pic: New serve mk.1-2517

That's, kinda, not on the list of approved motors?

09-07-2014 16:48

R.C.

Re: pic: New serve mk.1-2517

http://banebots.com/p/M3-RS395-12

Might be this guy, I believe it was legal at some point in time. But with the current motor allocations teams should just use the 550 or the 775.

09-07-2014 16:49

Tyler2517

Re: pic: New serve mk.1-2517

We used the 540.

We designed initially to use a 400 series moter with a moduler mounting plat to the gear box. Ended up using a 540.