08-19-2008 02:11 PM

Aren_Hill

Re: pic: Swerve Module

why such a large bevel gear on top? if you shrink the top bevel gear you can make your sideplates closer and make the thing more compact. And your probably gonna want some standoffs as low down between the plates without hitting the wheel, because that housing is gonna take alot of side stress when you get hit by another bot.

And really put alot of thought into how force is being transmitted from the module to whatever it mounts to, because right now i assume you have 1 bearing on the vertical shaft, and if your not putting a disc at the bottom 2 is a must.
I got away with one this year but i had a disc being supported at the bottom of the module to take the stresses

08-19-2008 02:28 PM

rtfgnow

Re: pic: Swerve Module

i note that the two gears on the left are abouth the same size, do you want a 1:1 geear ratio on those?

08-19-2008 02:49 PM

=Martin=Taylor=

Re: pic: Swerve Module

What kind of a wheel is that? Is it from a scooter? Always wanted to use an orange one of those on a robot.

I think your going to need some more side supports, and some sort of setscrews or e-clips to keep the gears in place.

Nice work though.

08-19-2008 03:11 PM

NickE

Re: pic: Swerve Module

If you put a 1:1 reduction on the bevel gears, you could shrink the whole module, making it stronger. If you need to put a reduction in the swerve module, you could do that in the gears on the side. Also, standoffs should definatly be placed as low down on the module as possible (without hitting the wheel).

Another thing to consider is using chain and sprockets instead of gears on the side of the module.

08-20-2008 07:24 AM

sgreco

Re: pic: Swerve Module

We do want a 1:1 ratio. If we use gears, we sort of need a one to one ratio because as one gear gets smaller, the other has to get bigger, in which case one of the gears would be sticking off the module either up or down depending on which gear needed to be bigger.

This brings me to the idea of chain and sprocket, I have an open mind to it, but I as well as many of my fellow team members believe that the chain will be a problem. We built a swerve drive during the build season, it had the the cim attached to the module and 25 chain bringing power to the wheel. The chain would constantly pop off, and we had a lot of trouble with it. So we build a new one without chain for our actual robot.

I'm guessing chain can work if done well (our original swerve was pretty low quality).

My team has been debating over methods of attaching this module to the frame and doesn't really like any thing we've come up with. Does anyone have any advice for attaching it?

08-20-2008 10:12 AM

tubeheaD37

Re: pic: Swerve Module

Its not that we have anything against the idea of chain. Its more that we have just had bad experiences with tensioning and alignment every time we try to use chains (not that many times, we are pretty new).
It always seems like it should be pretty easy but just never works out.

Even with a tensioning system I am pretty sure that a chain-based power transfer will be lighter, more efficient, and simpler than a gear based alternative given the tools we have access to and our level of skill with them.

I think it is a worthy time investment for us to develop and manufacture a couple generic chain tensioners before the season starts just to have some practice doing it well.

08-20-2008 10:39 AM

RaMoore

Re: pic: Swerve Module

You should look into 118's design. They use #35 chain without needing any additional tensioner and as far as I know have not had any issues with the chain.

http://picasaweb.google.com/Team1817...32808731775490

http://picasaweb.google.com/Team1817...21920054374706

http://picasaweb.google.com/Team1817...30986730337954

08-20-2008 11:51 AM

Aren_Hill

Re: pic: Swerve Module

We used chain on our swerve module this year, and tensioned by simply putting a cam follower in a slot that ran against the chain. It ran quite smoothly and they stayed put as long as a lock washer got in there.

Mcmaster pt# 6721K2

08-20-2008 12:38 PM

sgreco

Re: pic: Swerve Module

I don't know if I'm misinterpreting your post, so I appologize if I am. How would you use an e-clip and a key on the same shaft to hold in a gear?

The other thing I could do is use hex shafts and a snap ring to prevent the gear from sliding.

Does anyone one know if they make bearings for hex shafts?

08-20-2008 02:02 PM

NickE

Re: pic: Swerve Module

If you have access to a lathe, you can turn down the hex shaft on the ends to fit in a standard size bearing.

08-20-2008 02:04 PM

M. Mellott

Re: pic: Swerve Module

Instead of a hex shaft bearing, your best bet would be to turn down a section of the hex on a lathe to a round shaft, then slide on a standard roller bearing. You could even turn down most of the shaft, leaving only a small length of hex for your gear/sprocket.

As for the e-clip & key slot on the same shaft, there shouldn't be a problem having both--the e-clip groove will run around the circumference of the shaft, while the key slot will run parallel to the axis of the shaft. Even if the two intersect (forming a T-shape), it should be OK--just make sure the piece of key stock is long enough that it won't fall out if it works into the e-clip groove.

08-20-2008 02:07 PM

Madison

Re: pic: Swerve Module

Bearings for hex shafts are made for farm equipment, but they're considerably larger than anything we'd typically use on a FIRST robot.

08-20-2008 03:55 PM

sgreco

Re: pic: Swerve Module

This is a fairly low quality model, I threw it together as a smaller version of what I had, so I could reduce weight. I shrunk down the bevel gears and eliminated the actuation sprocket for the time being. I added snap rings to both sides of the spur gears.(although you can't see the groove or the key that is hiding behind the gears)

The top is 2/2 inches and it is 4 inches high. the module is 3/16 of an inch thick.

The old one was 1/8 of an inch thick, its top was 3/3 and it was 5 inches tall.

This one should be more compact and a little bit stronger.

08-20-2008 05:07 PM

Aren_Hill

Re: pic: Swerve Module

you can get 2x2/.125"wall tube pretty easily and then just cut a slot in it for your wheel, that would give it all the sideways integrity it'll need, as opposed to worrying about having the flanges bend.

And what size bevel gears we're you planning on using?

08-20-2008 06:02 PM

Joe Johnson

Re: pic: Swerve Module

Can I beg you not to do this?

Bevel gears are so seductive, but please please, fill your ears with bees wax and do not listen to their Siren song!

The rocky shores of competitions past are teeming with whimpering pit crews and grinding gearboxes whose over confident designers under estimated these bevel gear beasties...

For more bad metaphors and dire warnings about bevel gears, read Inhale this, but do not touch....

Consider yourself warned.

Joe J.

08-20-2008 07:47 PM

sgreco

Re: pic: Swerve Module

I'm not entirely sure yet, the ones in the drawing both have a 1 inch diameter and a .3 inch face width.

I guess I have to re-think bevel gears though...

08-20-2008 08:36 PM

Aren_Hill

Re: pic: Swerve Module

Okay, just to dull down the loud screams of "don't do it", me in the preseason to 2007 designed a swerve drive, with a randomly selected bevel gear. I constructed the modules on a manual mill, when i wasn't that good at accuracy.
I put the thing together, it ran, it was noisy but it ran. And the bevel gears never failed us. We've now done 3 swerves to date 1 prototype and then the 07 and 08 bots, we have never destroyed a bevel gear.

I did bevel gear spacing and alignment before i ever did spur gear spacing and alignment, and it worked, just choose a more forgiving gear, and make a prototype.

i used mcmaster part#6529K14
the first two times with keyways and the third time hex, hex is alot nicer.

if you have the time and resources i'd say build a prototype and get used the them.

08-24-2008 04:25 PM

Joe Johnson

Re: pic: Swerve Module

My experience from inspecting and judging FIRST robots for a dozen years now is that for every 1 teams that successfully implements bevel gears in their drive train, there are 10 or more teams that get the pleasure of watching their 6 weeks of work drive in circles during a match.

It can be done. 148 had them last year and they are part of the 2008 Championship Alliance. But, John V. is not a rookie robot designer and even when he was a rookie, he was better than most will ever become...

I doubt I will ever use one on any robot I have anything to do with, but others are welcome to their own opinions.

Joe J.

08-24-2008 05:24 PM

RMS11

Re: pic: Swerve Module

You can use a bearing at the top where the shaft is and at the bottom a lazy susan works quite well. There are also some more clever, lighter, cheaper, harder to build things. I love winnovations method, here is a picture. http://www.chiefdelphi.com/media/photos/28596

08-24-2008 11:30 PM

Cory

Re: pic: Swerve Module

That method was actually pioneered by Wildstang (correct me if I'm wrong, someone else may have done it that way before 111).

If I was making a swerve the ring at the bottom is one thing I would definitely copy.

08-24-2008 11:37 PM

Aren_Hill

Re: pic: Swerve Module

yep, i used lots of aspects of wildstangs swerve and ditto with the beast's version, then after that 118 and craig hickmans drawings

08-25-2008 07:28 AM

RMS11

Re: pic: Swerve Module

What is that ring made out of by wildstang? Does the module just sit in there?

08-25-2008 03:46 PM

Cory

Re: pic: Swerve Module

http://www.chiefdelphi.com/media/photos/28444

I thought there was a better picture of the ring itself, but I can't find it.

It's got a cross section that looks like a "z". It mates up with the profile on the piece of sheetmetal it mounts to, and the ring around the module.

08-25-2008 07:14 PM

Aren_Hill

Re: pic: Swerve Module

that and they said it was just delrin, that was it

09-01-2008 10:27 AM

joeweber

Re: pic: Swerve Module

Try two wheels instead of one. you can then run the shaft straight down between the wheels and use one set of bevel gears or one set of helical gears. This would eliminate the extra gears, simplifi your design and also give you more traction. We had thoughts of this but did not go farther. Draings http://www.team1322.org/ideas.htm#Rotate%20Drive click Rotate drive.

09-01-2008 12:09 PM

francistexas

Re: pic: Swerve Module

My only suggestion is to use spacers that are either flanged, or wide enough to rest against the housing and the bevel gear. Our spacers rested directly onto the bearings, and over the course of the season, we found that our bearings were being destroyed. The forces on the bevels actually made the spacers compress the two sides of the bearings together.

Good luck to your swerve! They're...'fun'.

09-01-2008 05:26 PM

sgreco

Re: pic: Swerve Module

This sounds good the only thing is traction is friction and friction means resistence to turn, which means we need a higher torque motor to actuate the module. My team used window motors last year and was planning on FP's for our prototype. Do FP's have enough power to turn a module, like the one on your website, with two wheels in it?

Even if the FP's are strong enough, I would wonder how efficietly they can do it.

09-01-2008 08:24 PM

AdamHeard

Re: pic: Swerve Module

Either way, adding wheels isn't going to increase the traction; However, making them wider will make it harder to rotate the wheels (If you think about it, the farther the tread is from the center of rotation, the more torque required to rotate them). The FP motor is well powerful enough to turn the modules, 1625 used a single (I believe) FP to turn all 4 modules, but I have seen more teams using a pair of globe motors to turn the modules.

I wouldn't try to use a helical gear as it forces the center of your wheel off the center of rotation of the module.

09-02-2008 08:28 AM

Aren_Hill

Re: pic: Swerve Module

1FP definetely has the power the turn all 4 modules as long as you take care with where and how load is being applied. The biggest part of swerve drives is foreseeing all of the weird loads on different parts just due to the nature of the drive.