Need general drive train expertise

[Caleb Sykes]

Originally Posted by Mike Marandola

Weight does have an effect on robot speed. The heavier it is, the slower it goes.

Can you clarify a little bit more as to what you mean by this?

[Oblarg]

Originally Posted by inkling16

Can you clarify a little bit more as to what you mean by this?

A heavier robot would certainly incur more frictional losses in the drive, and thus would have a lower top speed. Can't say how big the effect is without any data, though.

[Tristan Lall]

Originally Posted by Oblarg

I have never met a single person who believed that shifting with servos was a good idea.

I'd qualify that by saying that there are plenty of commonly-available servos that can do the job, but they haven't historically been FRC-legal. Digital servos for giant-scale model aircraft would be great, for example. Saving the weight and annoyance of a pneumatic system could make it very worthwhile.

(I've tried shifting a DeWalt-based transmission with FRC-legal servos before. It was terrible.)

[Oblarg]

Originally Posted by Tristan Lall

I'd qualify that by saying that there are plenty of commonly-available servos that can do the job, but they haven't historically been FRC-legal. Digital servos for giant-scale model aircraft would be great, for example. Saving the weight and annoyance of a pneumatic system could make it very worthwhile.

Oh, certainly, I've worked with some beefy servos in other applications. I was only talking about FRC-legal hardware.

Unfortunately, FRC-legal servos are so weak that I've yet to actually find an application for them. I'm sure someone, somewhere has used them effectively on their robot, but I haven't.

[RonnieS]

I know 67 has used a window motor configuration in the past because they did not have pneumatics. Using an opposing cam, it pushes or pulls into high or low gear.

Here is the link to their engineering notebook, I recommend any reads all of them anyways

-Ronnie

[JesseK]

Originally Posted by Oblarg

A heavier robot would certainly incur more frictional losses in the drive, and thus would have a lower top speed. Can't say how big the effect is without any data, though.

After writing a front-end to Ether's sims and playing with it a bit, very large swings in weight will have an effect on acceleration from a theory perspective. Torque remains the same, but the mass is larger, thus acceleration is lower. For the same amount of time, it nets a shorter distance traveled. Small changes (5lbs here, 5lbs there) seem to have no discernible difference in acceleration on a FRC field since it's less than an inch or two of difference.

Preview here. Requires Java 8, Win7/CentOS 6+/Who knows what for a Mac. Probably works in Ubuntu 12.04+, but I can't tell what version of GTK it ships with at the moment. The preview is just the latest iteration of layout ideas I'm playing with. For example, I don't think it'll have a web viewer upon release, which is what the current preview has.

[Ether]

Originally Posted by JesseK

After writing a front-end to Ether's sims and playing with it a bit, very large swings in weight will have an effect on acceleration from a theory perspective. Torque remains the same, but the mass is larger, thus acceleration is lower.

Changing the weight will likely affect the rolling resistance of the bot, especially on carpet.

If you change the rolling resistance parameter(s) in the model accordingly, it will affect top speed as well.

[JesseK]

Originally Posted by Ether

Changing the weight will likely affect the rolling resistance of the bot, especially on carpet.

If you change the rolling resistance parameter(s) in the model accordingly, it will affect top speed as well.

When I used to model in Excel, the resistance was modeled something like (1-0.995^n)*<weight> where n = # of loaded bearings. For most robots near the weight limit this falls between 5-6 lbs of resistance since only 4 wheels touch at a time.

This was really so I had some automated way to increase the resistance based upon weight and number of wheels hitting the floor rather than an accurate way to model it. It also provided a correlative model for what happened on our drive trains in 2008 & 2011, where the bearings had some binding due to improper mounting (thus the 0.995 was lower).

[EricH]

Originally Posted by Oblarg

Unfortunately, FRC-legal servos are so weak that I've yet to actually find an application for them. I'm sure someone, somewhere has used them effectively on their robot, but I haven't.

FRC0330, back in '05 and I think in '00.

Both were used as latches, and as I recall the one from '00 had some assistance from a spring. (Used to lock a winch with a pin--might have been replaced by a short-throw cylinder.) The one in '05 was rigged to drop a thin sheet of Lexan. Why a thin sheet of Lexan, you ask? It's really nice to be able to deny someone who is trying to jam you against a goal half of their traction. There was no retract mechanism, though.

[MrBasse]

Originally Posted by Oblarg

Oh, certainly, I've worked with some beefy servos in other applications. I was only talking about FRC-legal hardware.

Unfortunately, FRC-legal servos are so weak that I've yet to actually find an application for them. I'm sure someone, somewhere has used them effectively on their robot, but I haven't.

We used them to make our scale model of Dozer look left or right with his eyes when steering, other than that I've got nothing.

[Alan Anderson]

Servos controlled the pan/tilt of our camera in 2007. Tracking the green light over the rack was vital to the robot's success in scoring during autonomous.

2005's camera also had a pan servo, but a combination of factors led to our abandoning vision tracking that year.

There was a servo-controlled release on our 2004 fabric ball-collecting funnel, but we decided to leave the funnel off after our first couple of practice matches. We wanted to focus on knocking the 10-point yellow ball off its pedestal to drop the 5-point balls early.