21-01-2015 14:14

Re: pic: 6 week swerve

Looks nice. How much pre-season testing have you done?

21-01-2015 14:27

orangemoore

Re: pic: 6 week swerve

If you read what it says it sounds very much like they have no experience.

21-01-2015 14:28

lukedude43

Re: pic: 6 week swerve

None, programmer made some swerve code for fun and I had been toying around with the idea in the offseason but everything was started fresh on everything when the season started

21-01-2015 14:35

Re: pic: 6 week swerve

Out of curiosity (if you don't mind me asking) - what was the thought process behind deciding to choose swerve for the first time in-season this year?

21-01-2015 14:36

AustinH

Re: pic: 6 week swerve

I wonder if the game this year is at all conducive to taking these sorts of risks. We've been able to spare some individual team members for tackling quite a few of these "don't try this during build season" projects. For some reason, it seems like a more favorable risk/reward scenario compared to previous years in case we don't end up checking off all the boxes.

Looking forward to seeing how this turns out for you guys!

21-01-2015 14:41

Christopher149

Re: pic: 6 week swerve

Sorry, but I can't resist:

21-01-2015 14:48

lukedude43

Re: pic: 6 week swerve

Like a lot of other people we thought omnidirectional drive would be useful, swerve was best for our method of dropping off stacks of totes, and we honestly just thought it would be fun.

21-01-2015 14:55

Navid Shafa

Re: pic: 6 week swerve

*Looks good Luke, we've got 4030 on our Fantasy FIRST alliance so I'm really hoping you pull it off

21-01-2015 14:55

lukedude43

Re: pic: 6 week swerve

Haha this is beautiful

Something important I forgot to mention is that we have a set day in week 4 where we will consider if we have actually made it to a point where we think our swerve gives us any benefit over a traditional tank drive. When subsystem integration is fully cadded it will include a fully fledged tank drive configuration so if we decide to ditch swerve we should be able to have a running tank a few days later.

21-01-2015 15:49

asid61

Re: pic: 6 week swerve

If that bottom is a single piece, you will want to add a radius on the inner edges on the top.
How thick is the bottom plate? It looks pretty thin right now.
How are you constrianing the bevel gears on that shaft? How are you constraining the other gears?
Your top plates are too thin. It looks like it's 1/8" at most. Make it 1/4" to be really safe.
I LOL'd at that gear on the top. I assume that's a versakey gear with 3/8" hex bore like I ised on my drill press verson.
You need some thrust bearings on the module to support the vertical loads. This year more than ever it's going to be turning under a lot of load.

All things considered I would not do a swerve drive fresh out of the box. Please consider buying something like a premade team 221 revolution module and just belting some motors to it.

21-01-2015 16:14

lukedude43

Re: pic: 6 week swerve

Probably should have posted cad cause it would clear up a lot of your questions/considerations but my naming on some parts is embarrassingly bad (I blame sleep deprivation) but this should definitely help http://imgur.com/QfTX6i1

The bottom is being milled out of a single piece of 4x6x.25" 6061 channel

the plates are .125, but this is designed to use the chassis to beef up the outer edges and the gray squares are actually bearing blocks holding 1" ID 2"OD bearings. these are bolted straight through, counterbored on the top bearing block, with nut clearance on the bottom. We think this should be fine for preventing the plates from warping.

I didnt see your drill press swerve but yea it's a 3/8 bore 44 tooth versakey gear. Great minds think alike (and fools seldom differ)

there is a 1" ID thrust bearing and accompanying washers in there.

21-01-2015 23:29

nathannfm

Re: pic: 6 week swerve

Put your module into the CAD of the field right at the transition between the carpet and the scoring platform. Rotate the module as it would if the steering motor was rotating it, will that low ground clearance gear run into the ramp in any orientation? Looks to me like it will.

22-01-2015 02:33

lukedude43

Re: pic: 6 week swerve

Thank you for pointing that out, just updated to fix it http://imgur.com/eowq8F1

22-01-2015 02:36

asid61

Re: pic: 6 week swerve

As long as the wheel is bigger than the gear, it should be hard for it to crash unless the whole robot tips.

22-01-2015 02:42

lukedude43

Re: pic: 6 week swerve

Unless we want to do any sort of rotation on the step, that could cause the gear to hit the step and that would obviously be bad. The solution was to go from a 30:50 gear reduction to a 40:40 which increased the max fps from 11.5ish to 15ish but since pushing isn't an issue the lost torque from turning down wheel speed in programming shouldn't be an issue

22-01-2015 08:41

JamesCH95

Re: pic: 6 week swerve

The lost torque will be manifested as a loss of low-speed control, loss of acceleration (especially when loaded with game pieces), increased current draw from your drive motors, and a general loss of 'responsiveness' from your drive.

Do not dismiss such gearing changes so flippantly.

22-01-2015 13:28

lukedude43

Re: pic: 6 week swerve

Which is why we're playing around with last years robot in high gear with max motor speed turned down in code.

22-01-2015 13:37

JamesCH95

Re: pic: 6 week swerve

Good.

However, the principle of limiting power-train speed in code, typically done via limiting the maximum voltage from each motor controller, is very bad. This artificially cripples the drive train, which is likely the most important sub-system on your robot. You might consider using chain, belts, or another lower-profile means of getting the gear reduction you should have inside of the envelope you're limited to.

22-01-2015 13:40

Ether

Re: pic: 6 week swerve

http://www.chiefdelphi.com/forums/sh...d.php?t=132640

22-01-2015 14:12

lukedude43

Re: pic: 6 week swerve

I understand that it isn't a perfect idea, but what it does allow us to do is get 4 modules made and starting testing on 4 modules. If we want to change the top gear reduction to something that will give us a more reasonable top speed we can go back and change the belt reduction to #25 chain. We'd have to redo the actual gearbox plates but those are relatively cheap. 4 module testing will be the make it or break it for the swerve so I see continuing on and limiting motor speed in code to be the best option.

22-01-2015 14:23

lukedude43

Will come back and edit so this makes sense

Realized I forgot to add other important info but the gist of it is 3d printed preliminary pods and having the gears for this gearing change in shop

latest post covers everything

22-01-2015 14:24

Ether

Re: pic: 6 week swerve

Just to be clear: The point James was making is that your testing will not be representative of the fine control you could get by reducing the speed via gearing.

22-01-2015 15:30

JamesCH95

Re: pic: 6 week swerve

Yup.

I can appreciate that you have a hard deadline and you want to prove the drive works before then.

However, consider this: you can make the gearing change, the drive is sluggish and does not perform well when you demo it, but by gosh it never gets hung up on the ramp (which is the whole reason to change gearing, right?) Or, you can leave in a more robust gear ratio, and show that the drive accelerates well, is very controllable, and behaves nicely; and though the gears might contact the playing field under certain rare circumstances, you can solve that problem by making any of the changes mentioned in this thread.

If you're going to argue for swerve drive, which of those situations gives you a firmer position to argue from?

Edit: In the spirit of full disclosure, my team is setting up a chassis with a free speed of around 14 ft/s. We made this decision based upon easily available gear ratios, easily available wheels, and ability to neatly integrate it into our chassis. We are, however, planning on using 4 CIMs and 4 mini-CIMs, which will help offset the losses in acceleration and low-speed control at the expense of weight.

22-01-2015 16:17

Ken Streeter

Re: pic: 6 week swerve

Fine control through appropriate reduction gearing is a huge benefit. In past years, this has been a big reason we have used a 2-speed robot. Typically, people focus on "low gear" for pushing power, but a "low gear" is awesome for fine-control situations.

In 2012 we optimized the drive base design for bridge balancing. We knew we needed very fine control and designed the drive base around that. Teams all year wondered how our driver could balance with nearly any other robot so quickly and efficiently. The real key was the fine control of the drive base. If our driver wanted to move the robot forward a half-inch to finish the balance, she could easily do that - no problem. Our gearing in low gear that year was 2.35fps (after efficiency losses). The real benefit of the low gearing wasn't the power in low gear (although that didn't hurt) but was the fine control for bridge balancing.

Does stacking totes seem like a high-precision task?

22-01-2015 16:55

lukedude43

Re: pic: 6 week swerve

I'm back on a computer and actually have some time so hopefully I can write a response that will make me seem less willfully ignorant

First of, the change in gearing is an admittedly quick and dirty fix, but what I believe to be the best fix just to get some pods printed up and working together. We are 3d printing the first iteration of the pods in the next few nights so that we can have something together and working so our programmer has the maximum amount of time to work on getting the pods working together nicely and field oriented control working nicely. The gearing change is by no means a final fix which is something I something I don't think I ever mentioned (serious mistake on my part), it is just a quick fix to get the test bed together and running.

What i'm leaning towards as a final fix is a cim face mounted to the lower plate with a cim gear going to a gear on some .5 hex shaft. That shaft will have the pulley that was previously on the cim with the same gear reduction. If I do this right that means the wheel pod reduction could actually end up being something like 50:30 (as opposed to the previous 30:50) which should leave more than enough ramp clearance for the wheel gear.

The testing we're going to do tonight is only to see if last years robot (roughly 15 fps) with motor speed limited will give us fine enough control for our collector (In my mind our robot design is split into collector and indexer, collector is the only part that needs fine control). If it is it leaves the current quick and dirty gear change fix as a semi viable option if for some reason everything goes to crap. All of the downsides (motor speed being limited as opposed to proper gearing, barely scraping by on ramp clearance, etc.) will all be taken into account in the final decision matrix that decides if our swerve is actually worth it or if we should throw in the tank drive.

23-01-2015 04:12

lukedude43

Re: pic: 6 week swerve

updates http://imgur.com/a/MYNGL