I have grown to love steel mandrel aluminum rivets (as Tom Line suggests) and PEM nuts or RivNuts. We often use double-threaded aluminum standoffs instead of passing a bolt through a long-ish spacer. All of these are great for reducing weight.
Warping and bending and not square - get good fixturing, design self-fixturing parts, and manage your heat input more effectively.
Blows through thin wall - aluminum cans and razor blades can be welded together. Knowing the materials, a good welder setup, and using a helium-mixed shielding gas and most people can weld thin-walls with a few hours of instruction. When I taught welding I started my students on .040in aluminum, and they picked it up quickly.
Can’t be fixed - simply not true. We’ve had parts welded, or welded parts ourselves, at several events. Not to fix broken welds mind you, but as general repair work when other things have failed. Even without a welder present, the rivets and adhesive you mentioned can make great scab repairs. Though, frankly, I can’t recall a single weld failure on one of our robots…
It sounds like you have had awful experiences with welded construction, and that is pretty lame. However, just like anything else in FRC, it can be great when executed properly.
If anyone wants tips or advice on how to effectively weld, I’d be happy to help!
Thanks for the input. We have a few mentors with welding experience and our school’s new shop has a welding booth, but we weren’t able to move in until November of this season, so there was no welding training - i.e. no welded robot. For us, it’s something of a lost, ancient technology - the last robot from us to feature any welding was in 2015.
Adding onto what others said, welding is something that is really helpful for us in our designs. My team did light welding around 13, 14 I believe, but in 2016 went heavily welded. The past 2 years I’ve welded myself, and actually cutting back on what we weld together. We have had issues in the past of welding too large of structures without bolts, making changes during season nearly impossible without taking a saw to it.
These are all true, however there are other reasons I avoid specifying welded aluminium joints both inside and outside of FRC when possible. The most common aluminium alloy used by FRC teams, 6061-T6, is a heat treated alloy. That means the material was heated and cooled at very specific temperature and time intervals to increase it’s strength. When you heat treat 6061 to the T6 temper, as almost all the aluminium you buy is, it will have an ultimate tensile strength (UTS) of 276 MPa. Untreated 6061, also know at 0 temper or full soft, has less than half the UTS at 124 MPa. Why does this matter? Because when you weld aluminium the quick heat cycle gets rid of almost all the heat treatment in the material. The area immediately around your weld will be softened all the way to 0 grade. The welded part can be heat treated all the way back to T-6 but the equipment and knowledge to do this is out of the reach of most FRC teams. There are non heat treatable alloys of aluminium that don’t lose their strength when welded, but these are more difficult to acquire, especially from FRC vendors.
Does this matter to the average FRC team? Maybe not, most FRC parts are conservatively designed and don’t go anywhere near their ultimate strength. The 120 lb weight limit is high enough that the robot structure doesn’t need to be very well optimized. However if I’m designing a joint between two pieces of aluminium I often find it’s easier and cheaper to reach my target strength with epoxy/gussets/rivets/bolts than with welding. Your mileage may vary but it’s definitely worth reading up on the subject is you’re planning to go into any sort of structural engineering.
I’m going to assume that the bulk of your post is directed to general audience and not at me.
Commonly available materials, such as 5052 Al, or many low-carbon or alloy steels, retain a larger portion of their strength than 6061 T6 after welding. While what you said about 6061 T6 is true, but any FRC team worth its salt with use more weld-friendly alloys if they pursue welded construction. Thus the point I made about knowing the materials that you work with.
The strength lost from welding is important in strength-limited cases. Often in FRC, and in real life too, we have stiffness-limited situations. In stiffness-limited situations welding is of no structural detriment.
As an example: from 2010-2013 we dropped 3lbs+ from the KoP drivetrain by welding the chassis c-channels together. It was easy to fixture these parts squarely, required about 20in of welding, and the material was 5052.
A more recent example: in 2017 we made many of our gear handling parts from welded 5052 sheet metal. The human station collector used a tab-and-slot self-fixtured design that saved us 9 sheet metal bends and 30+ fasteners per assembly. The gear manipulator parts had two seams tack-welded closed, saving 2 bends and 4+ fasteners per part.
My point is this: if welding is available, the designer must consider if it makes sense to weld, if they can use weld-friendly materials, and if other techniques provide worth-while advantages. No one is suggesting to just weld anything and everything together with reckless abandon.
Steel mandrels are stronger than aluminum mandrels and thus seat the rivet with more preload. They are also rated for a ~20% higher shear strength than aluminum mandrel blind rivets. Steel mandrel rivets are also ~60% of the cost of aluminum mandrel rivets.
Practically speaking we found that steel mandrel rivets held tight more effectively than aluminum mandrel rivets. We had few, or no, operator errors when installing steel mandrel rivets. The steel mandrels did not deform as much as the aluminum mandrels and thus ejected from our rivet setters without jamming.
When installed properly there is no steel to drill through when removing the rivet. If you’re having mandrels break off high (we did at some points) review your installation tools and procedures, it’s likely that something is amiss.
The advantage of aluminum mandrels is a lack of corrosion/galvanic corrosion potential, and a lower installation effort. Neither one is a compelling reason for us to use them.