New AndyMark Products (Swerve & Steer module?)

[Adrian Clark]

Originally Posted by AdamHeard

The long cantilever on the bevel pinion worries me too.

I noticed that too. I did some quick calculations on the bending of the shaft under stall conditions and got .000187 max deflection. But keep in mind that's at the end of the shaft, not the mating surface of the bevel gear. So it seems the long cantilever shouldn't be a problem at all.

For those interested, the values I used to calculate the deflection are as follows:
.75 length of unsupported shaft
26.5 angle of pinon pitch cone
20 degree pressure angle
3/8" 4140 hex shaft
47 lbs radial pinon load under stall
145 in-lbs at pinon under stall
25 and 50 mm pitch diameters (from AM site)

-Adrian

[mman1506]

Wow, I'm confident we could build this faster than our versa-chassis. I really love how it comes fully assembled, the drive train is something I want to touch as little as possible. Being able to purchase a replacement module and just swap it in is awesome.

[asid61]

8.2lbs is a little much, but still realy good with everything on it.
Are the gears steel or aluminum? Are there alternate gear ratios available?
I'm wondering if there's a way to remove the first stge of gear reductions without compromising anything. This could easily be made half a poun dlighter if that were the case.

EDIT: I can't see in the cad how the smaller bevel gear is constrained axially. Is there a shaft collar I'm missing?
DOUBLEEDIT: Same goes for the large bevel gear. It appears to only be constrained by a single e-clip and the wheel. And the wheel is not constrained.

[Jon Stratis]

Originally Posted by asid61

8.2lbs is a little much, but still realy good with everything on it.

I imagine it would be incredibly difficult to get it to weight any less...

PG-71: 1.9
CIM: 2.82 lbs
4" wheel: 0.31 lbs

However you slice it, you pretty much have to have those components for swerve (we wont go into variants that chain together steering... that invites a compromise situation that I don't think these modules were designed for). Those 3 parts alone total up to 5.03 lbs, which is 61% of the weight of the module.

Of course, there's always the discussion on where to spend your weight budget, and how advanced drive trains like swerve play in... but that's probably best left to a different thread

This looks like a pretty awesome solution. Personally, instead of a PG-71 for steering, I would have liked to see an attempt made to incorporate the Denso Throttle Motor... we've been allowed up to 4 of them in the past, yet I think they are the most under-utilized motor in FRC so far! Of course, using them would probably require a complete gear box design as well...

[AdamHeard]

Originally Posted by asid61

8.2lbs is a little much, but still realy good with everything on it.
Are the gears steel or aluminum? Are there alternate gear ratios available?
I'm wondering if there's a way to remove the first stge of gear reductions without compromising anything. This could easily be made half a poun dlighter if that were the case.

EDIT: I can't see in the cad how the smaller bevel gear is constrained axially. Is there a shaft collar I'm missing?
DOUBLEEDIT: Same goes for the large bevel gear. It appears to only be constrained by a single e-clip and the wheel. And the wheel is not constrained.

8.2 is fine.

Its a COTS swerve for $325. Thats a different set of design constraints than a team. Very difficult.

[pwnageNick]

Originally Posted by AdamHeard

A sensor on the back of the 775 with that much reduction after is going to be pretty sloppy imo.

The long cantilever on the bevel pinion worries me too.

I'm gonna second both of these concerns, especially the second.

[asid61]

Originally Posted by AdamHeard

8.2 is fine.

Its a COTS swerve for $325. Thats a different set of design constraints than a team. Very difficult.

I love the design. $325 is absolutely incredible. Even custom swerves can cost more than that (for sure), and after seeing how many machined components are in this one I'm shocked they're not selling it at a loss.
8.2lbs was okay a few years ago, but we have seen lighter swerves from all over CD in the past few months (including a shifting version similar to this weighing in at 7lbs, and that one would survive a season as it was designed by PWNAGE) with this same basic design. Combined with the extra high COG of this module, I feel like it could be improved quite a bit. If a COTS swerve drive could compete with WCD in terms of weight, then it would be a pretty big improvement IMO. Of course I love swerve drives, so I'm more than a little biased here.
I'm looking at the CAD right now to see if there are any cheap improvements that can be made to raise the caster box up a level. It looks like it's filled with custom parts already (seriously, just look at it) so some different lightening patterns could really hack off a lot of weight. Plus, there's an extra plate that doesn't need to be there under the top plate.

[AdamHeard]

Originally Posted by asid61

I love the design. $325 is absolutely incredible. Even custom swerves can cost more than that (for sure), and after seeing how many machined components are in this one I'm shocked they're not selling it at a loss.
8.2lbs was okay a few years ago, but we have seen lighter swerves from all over CD in the past few months (including a shifting version similar to this weighing in at 7lbs, and that one would survive a season as it was designed by PWNAGE) with this same basic design. Combined with the extra high COG of this module, I feel like it could be improved quite a bit. If a COTS swerve drive could compete with WCD in terms of weight, then it would be a pretty big improvement IMO. Of course I love swerve drives, so I'm more than a little biased here.
I'm looking at the CAD right now to see if there are any cheap improvements that can be made to raise the caster box up a level. It looks like it's filled with custom parts already (seriously, just look at it) so some different lightening patterns could really hack off a lot of weight. Plus, there's an extra plate that doesn't need to be there under the top plate.

Please design something lighter that could be made and sold for the same price then.

[AdamHeard]

Originally Posted by Adrian Clark

I noticed that too. I did some quick calculations on the bending of the shaft under stall conditions and got .000187 max deflection. But keep in mind that's at the end of the shaft, not the mating surface of the bevel gear. So it seems the long cantilever shouldn't be a problem at all.

For those interested, the values I used to calculate the deflection are as follows:
.75 length of unsupported shaft
26.5 angle of pinon pitch cone
20 degree pressure angle
3/8" 4140 hex shaft
47 lbs radial pinon load under stall
145 in-lbs at pinon under stall
25 and 50 mm pitch diameters (from AM site)

-Adrian

Don't forget mechanical slop, harder to model.

The slip fit in the being multiplied times that length is more appreciable.

Its probably fine, AndyMark knows their stuff... but it looks scary at first glance.

[Scott Kozutsky]

This is going to be a blessing and a curse. I foresee many teams not prepared to do swerve doing swerve.

[wilsonmw04]

Originally Posted by Scott Kozutsky

This is going to be a blessing and a curse. I foresee many teams not prepared to do swerve doing swerve.

we found out the limiting factoring to swerve is not making a module, but having the programming to control all 4 at once.

[Ether]

Originally Posted by wilsonmw04

we found out the limiting factoring to swerve is not making a module, but having the programming to control all 4 at once.

What aspect of the programming did your team find troublesome?

[mrnoble]

Originally Posted by wilsonmw04

we found out the limiting factoring to swerve is not making a module, but having the programming to control all 4 at once.

1339 will be staying away from swerve for the time being, even though so much awesome stuff has been released this year. We just don't want to get in over our heads and build a robot we can't compete with.

[Darkseer54]

Originally Posted by Ether

What aspect of the programming did your team find troublesome?

Our team has very little programming support, and struggled with tuning a swerve. We understand that the math for it is online, and we use your math for our mecanum, but sensor integration and PID have been very difficult for us.

[asid61]

Originally Posted by AdamHeard

Please design something lighter that could be made and sold for the same price then.

The thing is, I can't gauge the cost of making this. Please just take a look at the CAD; there are many custom parts that need to be CNC'd or even shaped.
Look at it this way: if you want something cheaper, then you can just take any old swerve drive and CNC all the parts from stock. Heck, might as well just make the encoders too, just for the lols. Sure, it's cheaper- technically. But there's no way of knowing how much their machine time is worth, and how long it would take to manufacture some of the more complex parts.

AM's design is fantastic in that it's also bulletproof, which adds to weight as a tradeoff. But I think it's possible to make it lighter without increasing the number of non-COTS parts, by using concepts laid out in (for example) Aren's recent swerve. I'll let you know if it's possible to remove that middle plate. My goal is to add one more part (made by a team or something) but decrease the weight by removing some of the other parts.