> Body panels are painted separately

Hasn't this been the source of some quality issues, due to variance between panels?

You can either paint a set of panels together off the car and make sure they go onto the same car, or you can run batches of simple colors with low variances between panels and not care which car they go on.

Yeah but handling painted body panels in production is a non trivial effort

They have not yet implemented this process, so it can't be the source of current issues.

You never paint everything at once anyway, so paint matching is a problem always.

You need a highly consistent process either way. Seems to me a smaller paintshop would be easier to quality control. However more parts going individually threw the process also could increase variance.

I would think they care more about saving lots of money then potential paint variance issues. If they happen they will try to fix it otherwise.

How exactly does it work?

Here it is visualized/explained:

https://www.core77.com/posts/120223/Teslas-Radical-Unboxed-M...

Though while it seems to me at first sight more efficient (in the sense that it may be faster) I cannot see where the 50% savings could be.

I think its economy of scale and time/labor savings.

Right now there is one main assembly line, that almost everyone interacts with during final assembly (especially for the interior of the car). So you might have less efficiency if something goes wrong earlier in the line.

In the unboxed method, it looks like there are five or six main assembly lines running in parallel, and all of the interior work also occurs on these lines for their respective sub-components. Finally assembly in this method should literally just be bolting the sub-components together, attaching wiring adaptors, and you should have a working car.

Yes, but all that can lead to 50% savings?

And at the same time reduce factory footprint by 40%?

>Officials said the unboxed process could cut production costs in half and reduce the factory footprint by 40%.

It sounds optimistic to me.

I don't think they mean 50% saving in the whole car. I think they mean 50% saving of the production cost, outside of the materials costs.

And that I assume would be achieved by having a higher threw-put with the same input in terms of humans and machines, like early RISC processors.

Simpler pain shop would be big savings, in floorspace and amount of robots.

Less moving of the whole car around reduces a lot of big robots.

Humans and robots have easier access to the parts they need to work on, increasing both human and robot productivty.

I think this number also includes the changes to the wiring that Tesla has been working on for a few years now. Wires are one of the most difficult and labor intensive parts of car building, and a frequent source of errors. The traditional process isn't really feasible anymore with this assembly method.

So I assume this number also includes the elimination of those task.

Footprint by 40% seems reasonable to me, saving of 50% seems high I would agree. Depending on what all is included in that number.

Yes but even if the halved costs are production costs, without materials, they are essentially investment in robots/machinery and labour.

I don't think that much can be saved on robots/machinery, so it must be largely labour, and to cut down those costs by 50%, of even 40%, seems to me like very improbable.

About the footprint, it could be reasonable and actually feasible, but that could be a great thing from some viewpoints (less visual impact of the factory on the territory or possibility to build a new factory on an area too small for a traditional plant ), but it doesn't sound like it could lead to relevant savings, the costs of actual land area and buildings should be a minimal percentage of the production costs.

I believe they assemble more of the car interior and components before putting the outside on. One example is putting the seats in place before the external body is put on top. Typically the frame is put together first, then you have to maneuver things into place. The "unboxing" makes assembly easier and you can do several of these in parallel, then bring them together at the end

The "frame" isn't a unified box until much later in the process.

It seems like it trades more "stamping" early in the process for more "welding" late in the process.

The primary issue is likely to be process control. Variance in the subassemblies will have to be compensated for in the late welding stage (which itself has variance).

I suspect this means that you have to be a LOT more controlled with your interior assemblies. This likely wastes the benefit of the chassis assembly savings.

However, one thing that has changed is that wiring is now significantly different and may be the key enabler. You can hook things together in a car with a single power connector and a single ethernet jack. That's a very different approach to wiring harnesses.

I hope Tesla succeeds, but they don't have the best reputation for quality control which this system demands in spades.

What a horribly written article. They barely mention what the unboxed model is. They compare it to lean manufacturing, only to explain that it's a stupid comparison. The only negative argument for this "debate" is that it'll be hard to keep the parallel production lines in sync, as if that's not a factor for the +10k parts that go into cars to begin with.

It’s an interesting concept. From a final assembly perspective it makes a lot of sense if you can pull it off. There is a design trade off that you can’t really weld painted parts so you need to use bolts to transfer loads and you need to beef up structural thickness to accommodate.

Bolted assemblies add lots of complexity with fastener location, shimming and mismatch of angular profiles. If you goof that you will preload the parts and probably crack the paint and or the part.

This will be hard to pull off and even if successful would require significant more expensive manufacturing processes for tolerance control