...I have had too many of them break where the nut and expanding rivet shell are joined.
Interesting; I've not had that problem (yet). Trying to think of why that would happen (ignoring of course defective rivnuts) and all I can come up with is rivnut intended for thicker metal and thus being "over-folded"? Does this happen with just-installed rivnuts? I can imagine they might tend to corrode where they fold but that would take a while.
BTW, anyone ever gotten away with using thread locker on the fastener? I've never had the nerve to try it, although obviously one could experiment with scrap metal. And there is a "weak" (purple?) Loctite that might work..
<after a pause for some web searching>
It occurs to me that if you over-crimped a rivnut it would pull the nut part up into the rivet part in a way that would start to thin or cut into the folding part. See pic and imagine pulling with arbitrary force on the threads of the crimped part on the right.
Which raises the question of what is the ideal crimping force and how do you determine when you've applied it? Too little and the rivnut spins, too much and the rivnut fails. I bet the rivnut mfrs have a spec that says if the metal is x thick the compressed rivnut should be y long, or else a pounds-of-pull spec. But with hand tools I don't know how you would measure the latter. IAE not sure how we amateurs are supposed to deal with that, other than experimenting with scrap and sectioning a lot of rivnuts.... With hand installation there is a pretty clear point where everything "goes solid" and that's all I know how to do.
<pause for some more web searching I should have done in the first place>
See the attached document from
Cardinal Components - Your source for Rivnut® engineered fasteners and installation tools. themselves.
View attachment Rivnut_Mechanical_Properties.pdf
It specifies an installation torque for each size. Which reminds me that the actual fastener itself can have the effect of "crimping" the riv nut as well. So (I'm thinking) the thing to do is crimp the rivnut short of the ideal force, and then when you install your fastener, just torque it as stated in the table. That, presumably, does not exceed the ideal crimp while also installing whatever it is you're mounting to the frame. Furthermore, the table says the figures are based on standards for torqueing grade 8 and class 5.5 fasteners, so why would you want to go higher than that? My thought is that if you think you need stronger fasteners and higher torque than that, maybe you shouldn't be using a rivnut.
BTW they make special "HRT Steel" rivnuts for use with class 12.9 hardware. (!).