Tesla is facing issues with the bare metal construction of the Cybertruck, which Elon Musk warned was as tricky to do as making Lego bricks
Tesla is facing issues with the bare metal construction of the Cybertruck, which Elon Musk warned was as tricky to do as making Lego bricks
These tolerances are very possible to hold while machining, but speaking from my perspective having been a machinist by trade for 20+ years, holding those tolerances for every single part on a vehicle is going to get prohibitively expensive really fucking fast.
He’s probably hyper self conscious about people ripping into Teslas over their clearances (with inconsistencies measured in millimetres). But, no, instead of saying “VW can produce stuff that doesn’t look like it fell from a truck and you will figure it out, too” he’s going overboard.
If we’re going overboard, why bother with cars at all? Just use this cheesy blueprint, make it work and solve all of humanity’s problems! This is what California should invest in instead of trains.
Micron, as in 0.00004"? Yes you COULD hold it, with second ops and temp control.
The article states “sub 10 micron”, which I interpreted to be +/- .0004" in practice
The total tolerance is .0004". In equally disposed bilateral tolerancing it will be ±.0002".
Eh, if someone tells me to reduce a tolerance from 5 to 10 thou at work, it’s understood that it’s +/-5 and 10. I don’t think I’ve ever heard someone use the full range of a tolerance in conversation. If the tolerance isn’t bilateral, it would be said like plus 5, minus zero. Anyways, +/- .0005" is our standard tolerance on the span of all dowel hole pairs.
Bilateral tolerancing is a Machinist’s first introduction to tolerancing so it’s no surprise to run that as default. And I suppose GD&T is not heavily used where you are.
If you’re given a parallelism tolerance of 10 micron are you assuming that to be ±10 micron? True position? Angularity of 5 thou? Etc… The only feature control that could be interpreted as bilateral by default is profile and it’s still communicated by its total tolerance.
Simple ± tolerancing isn’t the industry standard anymore. And if Tesla prints are anything like spaceX ones… It’s basically all GD&T and minimal title block tolerances.
I use GD&T on all my drawings, including 100% of my hole callouts. However I’m one of the more enthusiastic adopters of ASME Y14.5 at the place I work. Therefore, I get what your saying regarding the tolerance range, but since most of my coworkers are still relying on block tolerances, I’ll refer to a .010" positional tolerance as a “+/- .005” equivalent" in conversation so there is no miscommunication. I can see how this is not the norm.
On the dowel hole point, just measuring this stuff is going to take at the bare minimum an automated and purpose built CMM, which will drive the cost up even more. If we are to assume +/- 5 microns for every single part - we are talking about the level of manufacturing that Mitutoyo or Starrett have. This will be a multi-million dollar vehicle that noone would buy.
Just curious… what does Starrett have that Tesla will need?
Starrett isn’t known for quality precision metrology.
Well, in terms of for real metrology, you are correct. A better comparison would have been Brown and Sharpe. However, Starrett has more than enough reputation of everything that they produce being of a very high standard- primarily layout tools like calipers, precision levels, etc.
ETA: This could very well be my bias as an American showing. I know from experience that the fit and finish of a high end pair of Mitutoyo calipers have what I consider to be subpar to the Starrett equivalent in terms of fit and finish. There is also a $500 ish price difference which could also be a subconscious bias.
Elon: “Stacking” tolerances? No, we will not tolerate anything less than micron precision on every aspect of the design.