Gearing for 20 FPS?

[Marc S.]

Originally Posted by NuclearNK

I'm no physics buff, and could be wrong, but wouldn't this have a pretty slow acceleration?

Not if you design it correctly. 973 was geared for 17/7ftps and excelerated faster in high gear. And while we did use 6 motors in drive for half the season, it was designed to work with only the 4 cims. This is why the gearbox calculater is such an essential tool for custom drivetrains because it allows you to know for sure if it will work or not.

Sidenote: 148 in 2008 was originaly designed for 30fps, but later reduced because it was too uncontrolable.

[Brandon Holley]

Originally Posted by Marc S.

Sidenote: 148 in 2008 was originaly designed for 30fps, but later reduced because it was too uncontrolable.

Is this actually true? John? Other 148 members? That seems like something even 148 wouldn't attempt.


Our drive this year was geared for 18.5 ft/s in high gear after accounting for losses. Before losses, the theoretical free speed was something like 22+ft/s.

I discussed this in another thread, but controlling the robot at that high rate of speed is not something to take lightly. Here is the other thread- Post #s 8, 10 and 12 were my comments on driving at that speed: http://www.chiefdelphi.com/forums/sh...42#post1085342

Some teams routinely can handle the high speed, but it is certainly not for everyone. My main point is don't assume you can drive that quickly and be able to just control it like you could a slower robot. Make sure you have good reasons to go for it, you've done the math, have a solid design and a good strategy for control and you will be ok.

-Brando

[thefro526]

Originally Posted by Brandon Holley

Is this actually true? John? Other 148 members? That seems like something even 148 wouldn't attempt.

Post Here: http://www.chiefdelphi.com/forums/sh...2&postcount=11

Originally Posted by JVN

This is true. We are geared at about 20 ft/s free-speed. During the build season we were geared as high as 30 ft/s free-speed, but this turned out to be foolish.

[Brandon Holley]

Originally Posted by thefro526

Post Here: http://www.chiefdelphi.com/forums/sh...2&postcount=11

Thanks!

Now as I'm remembering back, I do vaguely recall that. That would have been something to see that little nine-sided robot cranking around the field at 30 ft/s.

-Brando

[thefro526]

Originally Posted by Brandon Holley

Thanks!

Now as I'm remembering back, I do vaguely recall that. That would have been something to see that little nine-sided robot cranking around the field at 30 ft/s.

-Brando

Something tells me that it may have been underwhelming. (Really fast in a straight line but hard as heck to handle making it painful to watch)

Considering how that swerve was driven (RC Car controller) if any robot was going to do 30FPS well it would've been 148... But with John saying that it would've been foolish, it makes me believe that 30FPS is just too fast to be successful on an FRC field.

I remember seeing a few 25fps-ish robots in 2008 and most of them seemed to be under driven or barely controllable.

(Though, at 30 FPS, they'd have been running laps in what? 5 seconds?)

[Brandon Holley]

Originally Posted by thefro526

Considering how that swerve was driven (RC Car controller) if any robot was going to do 30FPS well it would've been 148... But with John saying that it would've been foolish, it makes me believe that 30FPS is just too fast to be successful on an FRC field.

Yeah I agree with that. If any robot in any game was going to move that quickly, a round swerve drive robot driven with an RC controller definitely seems like the one to do it. I imagine taking non-banked turns at that speed is what started causing some issues (not to mention just the shear reaction time required at that speed).

-Brando

[thefro526]

Originally Posted by Brandon Holley

Yeah I agree with that. If any robot in any game was going to move that quickly, a round swerve drive robot driven with an RC controller definitely seems like the one to do it. I imagine taking non-banked turns at that speed is what started causing some issues (not to mention just the shear reaction time required at that speed).

-Brando

You're dead on Brando. Check out their 2008 video: http://www.youtube.com/watch?v=6fLf71xlVhE

Looks like 148 could've risked rolling over at 30fps, they were getting up on two wheels (Maybe one? Don't know how a 3WD swerve handles) during hard cornering.

[jspatz1]

Our 2008 lapbot was geared for 23 fps and had very good acceleration. 4 cims, 70 pounds, 6:1 ratio from CIMs to 6" wheels.
http://www.youtube.com/watch?v=52PxEzpm9Us

[Aren_Hill]

Originally Posted by jspatz1

Our 2008 lapbot was geared for 23 fps and had very good acceleration. 4 cims, 70 pounds, 6:1 ratio from CIMs to 6" wheels.
http://www.youtube.com/watch?v=52PxEzpm9Us

I can vouch for that, speedy little bugger

[Ether]

Originally Posted by jspatz1

Our 2008 lapbot was geared for 23 fps ... 6:1 ratio from CIMs to 6" wheels.

That's 23 fps at CIM free speed. Do you know what was the vehicle's actual top speed? (I'm trying to get some data on realistic numbers for speed de-rating)

[Jared Russell]

6 motors geared for 20 fps will draw a LOT of current at stall (when starting, pushing, or turning in a skid steer arrangement). More than I'd want to design for; you risk resetting the cRIO (if you use the original 8 slot model) or the wireless gaming adapter due to momentary drops in supply voltage when drawing 600+ amps of current. You will then be sitting lame on the field for 20+ seconds while the component reboots.

With 4 motors, 20 fps will result in a woefully slowly accelerating bot.

Both of the above statements assume a bot at or near the weight limit plus bumpers and battery - obviously, a robot that weighs 80 lbs soaking wet can be geared faster and still accelerate at a good clip (although since it is soaking wet, maybe not). Maybe those crazy Cheesy Poof guys who design ridiculously fast welterweight bots are on to something

[theprgramerdude]

Frankly, the easiest way to reach >= 20 FPS is, besides throwing on more motors, NOT approach the 120lb weight limit (which, when adding in the bumpers and battery, goes to the limit of about 150 lbs). Since F=MA, and V=int(A(t),t), Shaving 20 or 30 pounds would be a huge boost in acceleration, and thus make it possible to reach those speeds within the length of the field.
Taking our 2011 robot as an example, it only had sprockets+CIMple boxes as a reduction, while weighing about 110 pounds competition-ready, it accelerated faster to higher speeds than our similar robots in the past that hit the 150 pound mark, but had higher reductions to get similar acceleration up to their lower max speed.

[AdamHeard]

If you open up the question to bots that are geared fps free speed, it's not terribly uncommon. Our 2011 comp bot had a free speed of 20.6 fps, and our 2011 prototype Emperor Swerve was geared for 20.7 fps.

Frictional losses obviously bring these numbers down.

Both these robots have shifting to provide a low gear, and our comp bot was somewhat underweight and had very little friction in drive.

Using the oldschool version of JVN's design calc with the drive modeling tab allows you to graph the position and it's derivatives over time, and study what changing mass, friction, power, gear ratio, etc... has on them. It's very interesting how a *slight* friction difference can change speed appreciably.

[jspatz1]

Originally Posted by Ether

That's 23 fps at CIM free speed. Do you know what was the vehicle's actual top speed? (I'm trying to get some data on realistic numbers for speed de-rating)

Sorry no measurements. With a #25 chain directly from each CIM to each dead axle wheel, and its light weight, the friction loss was very low.

[IanW]

Using the oldschool version of JVN's design calc with the drive modeling tab allows you to graph the position and it's derivatives over time, and study what changing mass, friction, power, gear ratio, etc... has on them. It's very interesting how a *slight* friction difference can change speed appreciably.

I just looked at this spreadsheet and I must say, I am floored. As a student who knows very little about DC motor physics (or physics in general), I am amazed at the amount of work that went into creating it. I would be interesting in reading an explanation of some of the math/concepts behind it, if someone would know where to find it. Also, if anyone has any tips on using it now that the motor and gearbox scene is slightly different from where it was in 2004, I would definitely appreciate that too.
Well done JVN, it would be cool to see some of these features re-incorporated into a later iteration of the design calculator.