Tesla rethinks the assembly line
assemblymag.com
assemblymag.com
https://www.consumerreports.org/media-room/press-releases/20...
Tesla's problem is not scale. Their problem is that they cut every corner known to man in pursuit of scale. As a result, their cars are the lousiest shitboxes on the road.
https://www.motor1.com/news/669135/tesla-model-y-worlds-best...
https://electrek.co/2023/02/28/tesla-model-y-work-of-art-toy...
Be careful with conflating 'popularity' with 'quality.' Tesla should be wary of their recent success. There's the potential for huge backlash when people start to realize Tesla's support is poor if your issue lies outside the bounds of the 'typical' issues they're expecting.
Bad news travels fast.
lousy shitbox > perfect corolla
Where to start with this article?
Here is a good spot I guess.
First off, Moravy is wrong.
If anyone thinks that auto manufacturing and the "vehicle assembly process" has not substantially changed in the last 100 years, then they are totally ignorant of the industry and of the exacting details associated with something as complex as automotive manufacturing.
The other vital thing that this article fails to mention at all is how manufacturing is shaped by the larger concerns of the product lifecycle - which is (or should be) the actual "product" that leaves the factory.
"The car" is just a hunk of metal that embodies the product lifecycle - which can be competitively unique from manufacturer-to-manufacturer.
One cannot talk myopically about "costs" and whatever happens on the manufacturing floor without bringing in the total concerns of the product lifecycle (i.e. service, end-of-life, market requirements).
That is difficult to do in an article because the total size and complexity of each automaker's product lifecycle is immense (and largely unknown externally from the automaker in question), but it must be done.
> If something goes wrong in final assembly, you block the whole line and you end up with buffering in between.”
Which is how, fundamentally or in part fundamentally, the Toyota Production System works - and it is difficult to argue with the quality results at Toyota.
Honestly, I am not seeing much of a difference here overall.
There are various component assembly lines that do run outside and "in parallel" with the General Assembly lines at incumbent automakers.
I am not even sure how this is debatable.
> “However, there are some quality-related risks involved, such as potential gaps in fit and finish,” warns Pischalnikov. (snip) “The reason that’s always been done is for color consistency, to ensure that there’s a perfect match between the doors and the rest of the car body,” Prasad points out. “By not having to assemble, disassemble and reassemble vehicles, you can reduce production costs and eliminate waste.
Which are quality control aspects that Tesla still seemingly struggles with, near as I can tell.
I am all for encouraging automakers to explore new methods of automotive manufacturing and BEV production will present significant opportunities to do so, but this article from Assembly Magazine is, at the very least, incomplete.
1) Tesla customers seem extraordinarily willing to overlook production defects and servicing/repair issues compared to customers of other OEMs. This lets them get away with lower manufacturing quality tolerances than they normally would, as noted in the article mentioning the water ingress issues. Seems like they're going to further capitalize on this customer tolerance with the paint process changes.
2) As a result of only building BEVs with no legacy support requirements, Tesla is able to design the manufacturing process and vehicles themselves to be more efficient to assemble. It's a definite competitive advantage today and that's another thing some of their intended changes here will try to capitalize on. The question to me is how long it will be until the traditional OEMs catch up here.
While automakers reuse factories, they have no qualms about building new ones and closing old. I have family that works on assembly lines and every couple years they get an offer to move to New Mexico, Arizona, Tennessee, Georgia or wherever the new factory is being built.
What is the legacy albatross that hangs around ICE manufacturer necks?
> "The car" is just a hunk of metal that embodies the product lifecycle - which can be competitively unique from manufacturer-to-manufacturer.
The thing is, Tesla doesn't have a concept about product lifecycle once the car rolls off the line. They don't care about tuners and tinkerers, they don't care about aftermarket sales (e.g. people realizing they might want a trailer hitch), they don't care about people ending in accidents (or why else does it need months for spare parts for a body shop), they don't care about maintenance (because let's be real, unless you get a lemon car, all you'll need to do for 10-15 years is brake and tire changes!) and no one forces them to do so either, so they do what makes the most profit for them: easy assembly trumps everything, and not having much of a dealer/service station network means you don't have to invest money into building it and schmoozing up dealers' arses for incentives.
Their entire structure is fundamentally different from conventional car makers. Add on top what the Chinese are doing, and the conventionals are headed for some really dark times.
But I have a few decades of insight into software development. And from that I can say that it is actually possible that a whole industry can be insanely inefficient without anybody noticing.
So I wouldn't be surprised if the same holds for cars and huge efficiency gains are possible.
Certainly something was wrong with the entire industry, and nobody seemed to pay enough attention, at least among its captains.
The wording in the article "[t]he breach would violate the GDPR" implies that the act of the whistleblower leaking the information IS directly the violation...
I'm a fan of whistleblower protections, but if there isn't anything reasonably interpreted as illegal on the safety front, and he shared everyone's private information creating a violation of the law, then he should obviously face legal consequences without any extra protection whatsoever.
I am struggling to even see what the whistleblower intention is really. He leaked a seven year period of drive assistance complaints and it was just 4,000 (this sounds like a surprisingly small amount of complaints for a layman like me). Was Tesla hiding this in court cases? Were they not following regulatory reporting requirements?
Floor area of a warehouse is very cheap compared to the ~ $13 Billion worth of cars that a factory will churn out in a year.
The machines you put in a factory, and the people you employ in the factory are the expensive bits. The actual concrete floor area is super cheap.
If someone has looked at the total cost of a factory, calculated the cost per square foot, and then said "if only we could reduce the square footage, we would save money", then they are fooling themselves.
If the truck that delivers window motors can unload at a loading bay near the window motor storage racks, which is in turn near the place on the production line where window motors is installed, then at no point does a member of staff transport any goods along the length of the line.
Cram the line into a tiny floor area, and you are unlikely to be able to achieve that.
Namely, cars are expensive and hard to ship - they don't pack well, and they have tariffs.
Yet car subassemblies in many cases ship really well - you can probably pack 500 doors in a 40 ft container for example, with a shipping cost half way around the globe of $5/door.
And tesla, having factories all over the world, can then manufacture subassemblies in places with low material or labor costs (depending on the subassembly), and do far less final assembly in the destination country, avoiding import tariffs, high labor prices, and loss of efficiencies of scale.
Since they're vertically integrated, they have a lot of pricing power - can can decide if a car door is worth $50 or $1000 as they like as far as 'country of origin of materials' goes.
And they can also build a 'car factory' (ie. subassembly bolting together factory) in nearly every country worldwide, grabbing tax incentives from a lot of governments who would love to have a car factory. They can write "Proudly made in X to a US design" on all cars sold in X - saving on destination shipping, while also being the 'local' car for buyers who are nationalistic.
But yeah, Tesla is the first company to try that.
Apple is the only other company which can claim to be similar in that aspect.
Ford, like the first one, tried vertical integration. You can google how his rubber plantations worked out. Vertical integration stopped being a thing in the 70s, and for good reasons.
I was interested to learn this was also a consideration in apparel manufacturing: dyed fabrics have some perceptible color variation between batches. So, although some complex components are made separately before being combined, these components have to be managed so that all the visible fabrics from the separate components being combined into a unit came from the same dyed batches of materials.
(Our startup had to figure out at which stage of a high-end production line to integrate part of our supply chain integrity technology.)
To get the benefits, you need storage racks for completed subassemblies.
And, when the machine is fixed, you have the opportunity to run just that line with more operators faster to 'catch up' and refill stocks of that subassembly.
No shit, you need storage? Please, do tell me more...
JIT tells a supplier, 'I need X parts available at exactly Y time. If they are late, I will punish you massively. You can't deliver them early either.'.
So the supplier builds the parts early (so that any machine breakdowns/unforseen problems don't cause them to be late). The supplier then stores the parts till the exact time and delivers them. Parts are still stored, just the cost and effort of the storage is borne by the supplier.
You can see this if you take apart a car and look at date codes on, for example, microprocessors. They'll be months or sometimes even years apart.
It's not clear how a damaged casting could even be repaired.
In an ideal world, cars would be driving themselves and not crashing into one another, of course.
One change that people don't appreciate much is that Tesla has gotten much better at making their cars. While other automakers now have models which spec-wise are competitive with the 3 and and Y, only Tesla and maybe BYD can make EVs profitably at scale. This is why they've been able to cut prices and still be decently profitable - a Model 3 RWD costs $40,240, which is about $33,500 in 2018 dollars (when the Model 3 started selling).
I expect legacy car makers will follow Tesla's lead on this process (and switching to 48V) much as they are now following and introducing large castings in their cars.
"Tesla plans to revolutionize automotive assembly with its unboxed production concept, which aims to reduce costs by 50% and factory space by 40%. Instead of linear assembly lines, Tesla will produce subassemblies from large castings and assemble vehicles in parallel. Tesla engineers believe this will enable them to scale production to 20 million vehicles per year. The unboxed process involves delaying 3D assembly for as long as possible to simplify operations and automation. While some experts are skeptical, others think the concept has merit and could transform automotive manufacturing if Tesla can overcome challenges with parts alignment and sealing. The unboxed assembly concept could enable huge gains in the paint shop by compartmentalizing and painting different vehicle sections simultaneously."