Fastest Obtainable Speed for an FRC Robot

Before this post gets going, let me clarify that my team is in no way looking to implement any ludicrous drivetrain speeds in an FRC game. Instead we are looking into maximum possible speeds using only FRC Legal parts.

Hi, my name is Noah from Team 6341, and I am here to gather a general CD opinion on a crazy off-season project. My team is looking into building a drivetrain capable of reaching speeds above standard FRC speeds. We are building this as a special request for a sponsor. Here’s what my team has come up with so far - :yikes:

Dual Input 775pro VersaPlanetary Gearboxes with a 3:1 Reduction to drive each individual wheel in the drivetrain. 4" x 2" Colsons for drive wheels. The drivetrain would be a 8 wheel tank drive, resulting in 8 different gearboxes, but 16 different motors. Depending on amps needed, this could be a very bad problem.

Resulting free speed would be 108 fps. Or ~74mph. :eek:

It is not very deep in the design phase yet. Overall, I’m just looking for general thoughts on our design, or anything we maybe could do better! Remember, we can only use FRC Legal parts, so keep that in mind please.

Regards,
Noah Minter

What does the sponsor want to do with it (please say giant RC race):slight_smile:

At very large speeds, you have to consider things like tires (which wear out VERY quickly depending on which you use) and motors (who aren’t designed for long term usage). Not to mention batteries for this thing. If you have to use FRC parts, but not make it strictly FRC legal, you can use more than one battery. Otherwise, you’re VERY limited as to how far / how fast you can go.

One thing to keep in mind regardless, however, is brakes. FRC robots rarely need actual brakes, but at speeds like this you might want to consider them so you don’t turn an alliance wall into an alliance floor.

(side note, by your math you’d be able to cross a field in < 1/2 a second.)

Disclaimer: I am not an enginerd and cannot do math. Just be careful.

I couldn’t find it in the very short time I have, but I believe that this was disused in another thread during preseason.

Here’ what I think I remember:

  1. 6 wheels
  2. 3-4 speed shifting gearboxes
  3. theoretical top speed of 60 mph+
  4. super light weight
  5. 6 CIMs + 2 MiniCIMs
  6. pneumatic wheels?

Remember it has been a long time and I only skimmed the thread so these details may not be totally accurate.

A few ?'s for future discussion

  1. what is your drive surface?
  2. resources available/willing to invest for this?
  3. do you have the need for speed?

See this thread: https://www.chiefdelphi.com/forums/showthread.php?t=138937

You are talking about building a land speed type of vehicle. In that case, you want a low drag profile for the robot, and a lot of gears to shift through to get to the top speed. Tank drive is a bad idea. Use an Ackermann type of steering setup that is used on normal cars. Look at cars that compete at Bonneville for inspiration.

An FRC robot is designed to maneuver around a small field, and that type of drivetrain layout is pretty terrible for making a land speed type of drivetrain.

Even basic Ackermann is limited. At any reasonable speed you’ll start to want things like rake, caster, and other automotive steering/suspension characteristics.

Knowing what the intent of this bot is would greatly help us give accurate advice.

Are you limiting yourself to a 54 foot field?
If so, acceleration will be key.

Dave

The fastest drivetrain I’ve seen can get 154 ft/sec using 10 inch decagon wheels in an omni pattern, each direct driven by a CIM.

https://www.chiefdelphi.com/forums/showthread.php?t=89768

Ah yes. The legend lives on.

(please say giant RC race)

Yup.

Using FRC Legal Parts, could we run two batteries and two PDPs? Would that setup mean we would have to run two RIOs?

Then we have to keep weight down…
and if air were to get underneath the very light robot, some bad stuff could happen.

Also we’eve been playing around in JVN Design Calc, and we can get the 775pros down to 30amps each using the current setup, question is, how much can we pull before we trip legitimately every breaker on the bot.

Before this post gets going, let me clarify that my team is in no way looking to implement any ludicrous drivetrain speeds in an FRC game. Instead we are looking into maximum possible speeds using only FRC Legal parts.

Thought I should restate that point.

Thank You Everyone!

First, if you’re going for the bleeding edge of physics you will want to use Ether’s design calculators. JVN’s tell us what to buy to get us ‘pretty close’ to what we want, but Ether’s help us optimize a design. This calculator is a bit of electrical simulation built on top of an original JVN sheet, and is what we used to simulate all of the conditions for our ‘ludicrous’ speed robot this year. One of the best things it told us was that we did not need to auto-shift from low to high in order to maximize acceleration and minimize time to the target distance.

Mechanically, fewer is better in nearly every regard. If you have the capability to make a custom gearbox, do it. Put all 775 PRO’s on it. Start with 3-4 motors, and cut weight out of your chassis to make up for lower power. Personally, I’d go with 3 wheels, not 4. 2 wheels in the front, and one in the middle of the back. All wheels are linked to a single transmission via 25 chain & long aluminum axles. If it’s for a race course, then perhaps make the rear wheel a custom giant swerve module (or linkage steering?) and drive it independently of the front wheels (which would still be linked via an axle, but would also require a differential gearbox between the axles).

Also, you’re talking about running this on an imperfect surface - a road (or still a gym?). Even airport runways have deviations. So unless you’re running it around one of Cape Canaveral’s VAB’s (the flattest floors in the world), you’re going to want some sort of suspension for high speeds. At least put bicycle tires on it ;).

Actually, it seems that weight may be a good idea, sacrificing the minimal extra speed for the not toppling over when it reaches the maximum speed without the weight

Since the chassis will require aerodynamic consideration already, this can be accomplished with spoilers, such as what is seen on the front fins of F1 cars.

When will this die?

I would definitely say do not use versaplanetarys, they are not made for drive trains and the gears will easily strip at high torque applications, as well as strip when moving from forward to reverse. My team made a custom 4 CIM gearbox that was direct driven to both rear wheels, this thing could hit about 20-30 mph but had little to no control. Hypothetically with crazy gearing insane speeds could be reached, but your biggest struggle will be to keep a low center of gravity for tight turns as well as mentioned above the wheels are key. You can have all the power in the world, but it means nothing if your tires just do burn outs (burnouts are fun though). I would definitely suggest pneumatic wheels with high surface are and a lot of grip, the WCP wheels are pretty good, but the rubber does wear down rather quickly. Good Luck!

Never. You can’t kill a legend.

This is always a fun thought experiment, and I would love to see some group of teams decide to make off-season robots to race on something like a school track. Decent tracks are about as good a flat, soft-ish surface as most of us can get access to.
As for design: I would strongly advise against using VersaPlanetaries as your drive gearboxes… There probably isn’t any gearbox sold by the FRC dedicated suppliers that fits your needs, unless you had some way of putting two shifters in a series to create a 4-speed gearbox. But your best bet on gearboxes is probably going pure custom. AndyMark and Vex both have decent selections of dog gears. (You could also go the CVT route. I’d love to see a CVT made from FRC parts.)
As others have mentioned, pure tank drive (referred to as skid steer outside of FRC) is not optimal for speeds nearing 20 mph. But at 60 mph? Trying to turn might be disastrous, for the robot and any bystanders. In my opinion, the easiest and most successful method would be to copy automobiles. Power two rear wheels and use some sort of Ackermann steering for the front two unpowered wheels. Pneumatic wheels are the way to go, but I’d be wary of using any of the wheels sold be actual FRC suppliers. (I don’t have any data to back this up, but my gut tells me the pneumatic wheels sold by AndyMark just aren’t designed for this application.)
As for how sophisticated your steering has to be and whether you need a full suspension or not, I’m not sure. That would increase the complexity and amount of custom machining needed by quite a bit, but it also would reduce the chance of your robot driving straight through a turn or worse, doing a barrel roll when you attempt a hard turn.
Just as a side note, you need to be using round bearings at every point on this robot. As convenient as hex shafts are, getting certified high speed bearings is important. FRC bearings aren’t designed for racing.
As mentioned, you’ll need to be aerodynamic. You’ll have to make some short of shell to go over the robot(/car? What do we call it at this point?). If you need more downforce to facilitate turning, it’s better to add downforce via spoilers or airfins than just adding weight. The lighter you can make this, the easier it will be to reach your design goals.

*aileron roll

In the end, it will really come down to how light can you get the robot and how much energy can you pull out of an FRC battery. The faster you go, you either lose acceleration or battery life. Designing this (and any) robot will be a balancing act of these variables.

Bingo - max obtainable speed is going to be a function of energy in the battery and the equation of kinetic energy. (E_batt = 1/2mv^2) so solving for V gives us:

sqrt(2*E_batt/m) = V [1]

So there’s your theoretical limit. Obviously we can’t do that with 100% efficiency so there’s some losses. But we can see that it’s important to get your mass as low as possible.

Now, if you wanted a practical answer, it gets more complicated - what’s your area you can accelerate in? Material you’re on? Do you need to decelerate at the end? (more than just plowing into a wall, I suggest the answer is yes)

Need more constraints.

[1] This ignores things like going downhill or tailwind cuz physics is hard.