Re: Chassis Connections
 

Here is an FSAE car chassis that I welded together my senior year of college:



Frame weight of 58lbs with all brackets included, welded from 4130 steel. I could imagine a FIRST chassis made in this manor weighing less than 15lbs. Just food for thought for you welding-intensive teams.

Re: Chassis Connections
 

If I recall correctly Kettering's Formula Zero team (hydrogen powered car) is using a similar chassis. Ours is a little smaller due to race restrictions though. Something tells me 1501 used welded 4130 in 2007 for their chassis.

397 has used the KOP frame the last couple years, we just bolt it together. There is something to be said about having a simple chassis together in one day. We have welding capabilities but only use them in situations that require it. No reason to over complicate things.

2337 used box tubing and some custom made blocks. These blocks were basically caps on the tube. Personally I feel this was heavy and labor intensive.

Re: Chassis Connections
 

Based on Billfred's experience and recommendation with 1/4" pop rivets, we've used them for 2 years with excellent results.

Re: Chassis Connections
 

15 pounds is a bit much for a drive chassis, we usually end up with a welded aluminum frame around 7 pounds.

Re: Chassis Connections
 

A very easy method for creating a chassis would be to use 1"x1"x1/16" aluminum tubing with Brunner Connectors. In order to secure the pieces together, we use 1/8" thick aluminum gussets with sheet metal screws.

This method allows easy modification of your chassis without having unweld or cut away pieces.

Re: Chassis Connections
 

...which is less than 15lbs last time I checked ;)

Do you have a picture of this 7lb aluminum frame?

Re: Chassis Connections
 

A 7-lb Al chassis versus a 15-lb steel chassis--that sounds about right, steel being roughly 2-3x the weight of aluminum. (And stronger, so you can use thinner wall thicknesses, further reducing weight, such that for equivalent strength, 2x is closer.)

Re: Chassis Connections
 

2009 robot, 6.9 lbs.

2010 robot, 7.2 lbs.

2010 preseason prototype, 7.0 lbs.

Weights are from memory, but I'm sure they're accurate within .5 if I go back and Check the CAD. Weights not including bellypan, which varied from 1.5-2.5 lbs each year iirc.

I didn't point out the weight difference to be harsh or brag, but to get people thinking. We don't do anything new or innovative with our frames (still ripping off Glenn Thoroughmen's fantastic designs) and we've never failed one.

Re: Chassis Connections
 

A bit off topic, but is the 2010 preseason prototype a 254/968 style with cams/diamond bearing blocks?

- Sunny

Re: Chassis Connections
 

All three posted bases are of that style, except the 09 didn't use cams.

Re: Chassis Connections
 

They look good. If you've got access to the CAD file would you mind switching the material to 4130 steel and making the wall thickness either 0.035 or 0.045? If that's easy to do I would be curious about the weight.

Re: Chassis Connections
 

That's impressive, thanks for sharing. Our first aluminum welded chassis (last year) was 12.365lbm in Inventor. This year's is 19.647, and we've bent it pretty noticeably, though it still works just fine. Do we just put more on the central weldment, or is there something else to it? Exactly what kind of stock are you using?

Re: Chassis Connections
 

Just something to think about, regarding steel vs. aluminum: steels are about 2.5 to 3 times as stiff as aluminum alloys, but both vary greatly in strength and in the ways they respond to welding.

For aluminum, all tempered varieties will exhibit reduced mechanical properties in the heat-affected zone of a weld. As a rule of thumb for aluminum, expect it to have the properties of the O temper (as in AA 6061-O) in the region of the weld, and to retain the properties of the original temper (e.g. AA 6061-T6) elsewhere. This won't change the stiffness appreciably, but will change the strength dramatically. Additionally, some varieties of aluminum need to be welded with a dissimilar filler metal that will have different (sometimes greater, sometimes lesser) properties.

In steels, this effect can vary greatly depending on the alloy being welded, the hardening processes already applied to the workpiece, the filler material being used and even the type of welding process. In the case of things like AISI 4130, AISI 4140 or other chromium-molybdenum steels, you'll actually find that they're supplied in an annealed (softest) or normalized (slightly stronger) state. That means that welding them isn't going to diminish the properties of these steels much further.

Failure of most structures can be said to occur when the part is permanently deformed (yielded)—so the yield strength is relevant. Normalized AISI 4130 has about the same yield strength (436 MPa) as tempered AA 7075-T6 aluminum (505 MPa—yes, "stronger than steel"), but if you weld them (producing something near the annealed condition), the steel will only weaken to 361 MPa, while the aluminum will drop all the way down to 105 MPa yield strength. So, if you can fasten the aluminum without welding it, choose the weld locations so they're under less stress, or in rare cases, weld it and then heat-treat the whole structure, you may elect to choose aluminum because of the weight advantage. If you need a welded structure with good strength even at the joints, then an alloy steel is probably a better choice (if you expect to see this sort of loading).

Cheap structural steel (what buildings are made of) is the worst of both worlds, however: low strength and high weight. For example, ASTM A36 is only rated for 250 MPa yield strength. (Actually, structural steels, like the old-fashioned ASTM A36, and the newer ASTM A572 and ASTM A992, have very variable properties—the minimum yield strength is specified, but the steel could actually be significantly stronger. In fact, nowadays, the majority of small-section steel shapes sold as A36 actually meet A572.) It does have low cost going for it, though.

On the other hand, if your structure isn't going to see levels of stress approaching the yield strength, maybe you're interested in designing for sufficient stiffness at minimum weight. In that case, because steels have a stiffness advantage roughly equivalent to their weight penalty (steel is about 2.75 times the density of aluminum), either one can be a good choice. At that point, if welding is possible for your team, you might as well go with the steel—you'll get similar stiffness and weight, but you'll be able to weld it rather than worry about more complicated joints. Indeed, it's rare for unwelded joints to be as stiff as welded ones, so welded steel will come out ahead, especially for complex shapes like that FSAE car pictured above. (Imagine if it was made from unwelded aluminum: as a practical matter, they would have been forced to use a stressed skin to stiffen it, or otherwise add more material. And if they'd welded it from aluminum tubes, it would have been weak at the joints.)

But what if you can handle the assembly of a stressed skin? In that case, you can probably get better results out of aluminum—sufficient stiffness by using thin webs between structural members, and sufficient strength by using high-strength aluminum structure. A steel structure constructed the same way would be infeasible, because of the difficulty of producing, fastening and maintaining paper-thin sheets of steel (compared to three-times-thicker aluminum sheets of the same stiffness and weight). (This is a big consideration with aircraft: you usually see aluminum planes, rather than steel ones. Only when there's another constraint—like air friction heating due to high-speed flight in the XB-70 and MiG-25, or cost as in some kit planes—do you see steel construction in aircraft.) As far as FIRST goes, stressed metal skins aren't typically very impact-resistant, and as a result, I'd be cautious about choosing them for the main structure of the robot. (I did have very good success in 2010 using polycarbonate sheets as the structural skin of a robot—but that was only possible because neither strength nor stiffness were major priorities in that part of the robot.)

Now, despite this talk of stiffness, sometimes it doesn't actually matter much. Have we ever known the field in Atlanta to be flat? (No, not a chance—those stupid tiles are worthless.) In that case, for some drivetrains, a bit of flexibility is actually very beneficial. (Holonomic ones that need all wheels on the ground are the best examples.) For the last few drivetrains I've designed (a 4WD wide-base and two 6WD long-base designs), we built them stiff longitudinally (so that they could maintain wheel and gearbox alignment), but flexible laterally, to account for variations in the field and to reduce the amount of structure. They weren't perfect for climbing (a crooked approach to an obstacle caused the frame to flex), but the largely open interior allowed us to keep the robot compactly packaged without worrying about structure.

Realistically, though, if you're willing to overbuild a little, you can ignore the conclusions of this sort of analysis and still make a perfectly serviceable robot. (It's probably for that reason that we see teams being successful with all sorts of designs that aren't necessarily constructed in the most technically optimal way.) Equally, different teams have different priorities when it comes to design. Some favour a rock-solid robot, others want one that can be taken apart easily. There is something to be said for getting the robot done quickly, and building it strongly enough that there's no chance the frame will need repair. That's something else that teams will have to think about—what's the tolerance for risk, and how capable are you of making a field repair to each of the possible designs?

To answer the original question: rivets. Lots of aluminum rivets, inserted into all manner of structural components. I've used aluminum extrusions, kit frame rails (sheet metal), aluminum-plastic composites, polycarbonate and high-density polyethylene as structural materials over the last couple of years. Typically, the way this works is frame rails are positioned at the outermost extents of the frame, and riveted to structural sheets on perpendicular edges. The sheets form the walls of the frame, and stiffen it somewhat, but the principal loads are borne by the frame rails themselves (to avoid drivetrain deflection).

Re: Chassis Connections
 

Last two years we have used welded frames. Mostly 6061 Al MIG welded.

Before that, we used bolts.

Reasons for the shift to welding:
1) Weight! No connecting pieces. Reduced nuts & bolts.
2) Strength - a well-welded frame is nearly as strong as the native aluminum.
3) Durability & Maintenance - Bolts work loose. Our welds no longer break. There's a learning curve here.
4) Design - As head of the Design Team, I am well aware that it is easier and faster to design a welded frame than a bolted frame. Especially if you are paying attention to detail (and the devil always hides in the detail).

The other benefit is that frequent welding makes us better welders, and yields welds less likely to fail. It also becomes easy and fast. We are now able to weld on the fly.

Downside is that when a critical weld fails, you need to work out a field fix. This can be easy. This can be damned near impossible.

Still, for us, the benefits of welding (applied to the right places) overcomes the costs and dangers.

Re: Chassis Connections
 

I would like to add that 4130, when welded with the proper filler material, can be stronger after it is welded. I developed and got AWS certified in a pulsed-TIG process for welding 4130 tube frames. My process used ER-80S-B2 wire, the 80 standing for "80ksi yield strength" in the filler, where annealed 4130 is about 75ksi (I like english units for welding as designations start to make sense: A36 = minimum yield strength of 36ksi, etc). The resulting welded material was stronger than the base material, but still ductile enough to be run through a tube-bender (and anyone who has done it can tell you that bending 4130 is very difficult).

Clem1640: watch out, your 6061T6 is becoming 6061T0 around the welded areas!