ValueAnalyst
Closed
This is helpful, but the 5-month logic/theory you laid out assumes Tesla expected to start the final assembly line in July.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
- Status
This is helpful, but the 5-month logic/theory you laid out assumes Tesla expected to start the final assembly line in July.
Yes, based on Tesla's plan to start *production* in July.
HG Wells
Martian Embassy
Wait a minute.
The WSJ is telling us that the car is HAND CRAFTED.
All the marketing I see portrays HAND CRAFTED as being the most desirable.
I have seen many adds over my sixty plus years from luxury car companies emphasizing their HAND CRAFTED luxury.
Perhaps the WSJ is telling us how desirable and valuable the model 3 is.
The WSJ is telling us that the car is HAND CRAFTED.
All the marketing I see portrays HAND CRAFTED as being the most desirable.
I have seen many adds over my sixty plus years from luxury car companies emphasizing their HAND CRAFTED luxury.
Perhaps the WSJ is telling us how desirable and valuable the model 3 is.
graphix25
-=electronboy=-
Price targets over time are a fools endeavor. One thing is certain, Tesla future value will be a direction function of it's growth and profitability. Today it near impossible to predict Tesla value as market opportunity is measured in trillions if you include energy and mobility. Technology advancement's exponential path is unpredictable and almost impossible to project from our current world view. The reason I invest in Tesla is at it's core the company represents a future I wish for my children on how we consume, store and generate energy. It could be the spark that sets in motion the future or become the future itself, nobody knows. Going to be a interesting ride over the next 5, 10 to 20 years.
Tesla could become a trillion dollar enterprise or just another company in the trash bin of the stock market history. Out of the 6100 companies in the NYSE and NASDAQ market place. I find Tesla mission and market opportunity to be one of the most compelling and interesting bets out there. Harnessing, storing and consuming energy is the core of our whole civilizations. A solar powered world seems like the most logical healthy outcome of our industrial age which is exactly Tesla reason for existence.
Place your bets and sit back and enjoy the ride. Don't get hung up on what others think in particular when it comes to price targets.
Tesla could become a trillion dollar enterprise or just another company in the trash bin of the stock market history. Out of the 6100 companies in the NYSE and NASDAQ market place. I find Tesla mission and market opportunity to be one of the most compelling and interesting bets out there. Harnessing, storing and consuming energy is the core of our whole civilizations. A solar powered world seems like the most logical healthy outcome of our industrial age which is exactly Tesla reason for existence.
Place your bets and sit back and enjoy the ride. Don't get hung up on what others think in particular when it comes to price targets.
For a large enough definition of testing, sure. But from an automation engineering point of view there is a real difference between debugging to get a line working versus optimizing to get a line working faster.
Based solely on past performance I would say every timeline from Elon about reaching certain production goals is always extremely aggressive.
I really beg too differe here. Those 15 seconds of footage in itself do not refute the claims from the article you quoted. (They also don't confirm them obviously)
My understanding is that you are an automation engineer. So, recognizing that even you might not be working in the auto industry, you, nevertheless, likely have a view at which point Model 3 lines could be considered as completed debugging, and being worked on their speed optimization. Is this point at a consistent 1k/week output? Higher? Could you give us your wag?
Ha, tell us something we do not know.
The reason I asked the question is that I've read at some point in time on TMC (do not remember which thread and posted by whom) that 6 month is a typical time for a new automotive production line launch from the moment it is assembled to the moment it starts producing at a nominal rate. So I was curious if you can give us some insight and perhaps a benchmark for what is a "typical" time for launching these BIW and final assembly lines.
Last edited:
I think that they refute *implication* of the article that reported manual welding of M3 bodies at some time back in September means that M3 bodies are still assembled by hand because the BIW assembly line is not operational.
T3slaTulips
Member
The thing that bugs me about the 1/10th speed operation of the production line is that it's not 1/11th speed. I would buy a production line that "goes to 11".
ValueAnalyst
Closed
IRIDIUM-3 MISSION | SpaceX
Launching in two minutes.
Edit: First Stage successfully landed on Just Read The Instructions.
Launching in two minutes.
Edit: First Stage successfully landed on Just Read The Instructions.
Last edited:
I think that it does not really show that a portion of the BIW line is operating at 500/week, because 1/10 of the speed on the video does not imply continuous operation at this speed.
It seems that the line is in the debugging stage, and a run at 1/10 of the speed could be followed by line (or a portion of it) stopped to analyze the produced bodies, verify what needs to be adjusted, or, perhaps which tools to be changed, etc. this process could take hours or days, we do not have enough information to guess this.
Last edited:
There is a particular problem using 'typical' for automotive, for that matter any production process. That problem is one of definition.
Two particular issues cause constant confusion:
1. What is 'new'? The vast majority of automotive production processes are, from a manufacturing process perspective, modest changes from prior cases. It is vanishingly rare to have truly all-new. Engines, transmissions, ancillary systems, running gear, third part parts and much more tend to be the same as others already used.
2. How much vertical integration? Nearly all existing auto manufacturing has third parties supplying most major components, often designing those components (i.e. Tier One suppliers). Those suppliers often provide entire drivetrains (e.g. LG for Chevrolet Bolt, Yamaha for the old Ford Taurus SVO, Peugeot/Mini JV for engines, and on and on).
Tesla is always a bit more difficult to evaluate because they have extensive vertical integration, but also have (for S and X) some absolutely standard parts such as switchgear, coolant pumps, brakes and so on. Even the vaunted BMS, as special as it is, is clever repackaging of industry standard pieces. Because the non-standard parts are so very complex from an assembly process and Tesla is attempting to use almost unprecedented levels of automation we can safely say nobody is very well equipped to guess how long it will take to reach the planned 10,000 Model 3 per month.
Still, Model X experience is instructive. After horrible FWD problems they've finally resolved the issues and there are very few problems. So, how did they do that? One major clue is offered during Fremont factory tours. There is a single automated line assembling doors for Models S and X, front, rear, FWD, self opening and not. Imagine that? Now look at the probable number of innovative assembly and manufacturing techniques used for Model 3. Much we do not know but we do know there are some serious innovations in materials bonding within the BIW. We also know that the recent all-glass roofs have been adopted in part because it makes interior automated assembly easier.
Frankly, there are approximately 500 robots in the Model 3 line, nearly all of which require individual "parameter setting" (I don't say "programming" because for most of them this in line parameter setting is not actually programming. It will take painful months to get all of that right. Once they do we'll see high quality high volume assembly happening. As with the prior door example nobody has enough analogous experience to know exactly how long this will take.
All of us and the markets too want to know exactly how long it will take and when the major worldwide deliveries of all variants will begin. I want to know too. As it happens we'll know when it happens but not before.
(Disclosure: I am not technically qualified to make these judgements. I have visited a number of manufacturing plants that have been state-of-the-art at the time I visited, including plants that built industrial robots. From a process management perspective I recall none of them happened just the way they were planned. Some things went better, others worse. Eventually they were brilliant advances. I conclude that doing things in a new way is less predictable than is the old, tested way. That always leads to uncertain timing. Despite my lack of engineering knowledge I have learned that new things require more patience than do old ones.
New white paper from Sam Korus (ARK) on EVs and trucks..
Electric Vehicles | The Future of the Automotive Industry — ARK Invest White Paper
Electric Vehicles | The Future of the Automotive Industry — ARK Invest White Paper
While I do work as an automation engineer in the auto industry, that's really above my pay grade. I work on a very specific type of equipment and therefore have a rather micro rather than macro view. That said, I think that many here have too much a 'Charlie Chaplin-Modern Times' view of automation development. Like "The line starts, the line stops, it goes at 10%. Made boss Elon barges in and tells to run section 5A-3 at double the speed'. That's not really how development progresses. Personally, I consider a process 'debugged' if a regular crew can run it without me being there. I really have too little insight to tell you at what point that would be for the M3 line. Hell, I can't even put a number like that on my own projects since it varies so much. My most recent project had a 23s cycle time and a 7 minute buffer. No margin of errors there. But I also remember one where we had to replace old equipment that had a 12% 'uptime', rest was wasted due to tooling. I could have done such a sloppy job just by virtue of working with faster, newer equipment that even a major crash every shift would still have been 'good'.
One of the closest projects I can see is the retooling of Audi Belgium in prepartion of the eTron. That's also a working factory that has to accomodate production of a new model. The first physical changes to the factory started in 2016 and earliest production is for 2018. Another data point : it took NedCar 18 months to retool from Mitsubishi to BMW. So 6 months for a brand new line within an existing factory running with minimal interruption. I'd consider that very aggressive. In fact, I was a little bit surprised when we had reports of crate unpacking spring of this year. My assumption was for Tesla to have been at that point already second half of last year instead. When we started to see a clear run up in VINs second half of September, I was surprised again. Positively this time. But given what we know now, this was likely due to some creative production methods.
jhm
Well-Known Member
Right, but my basic point is that the slowest bottleneck sets the pace for the whole line. That pace may be slowed by lots of inspection and diagnosis as you point out.
Thank you for sharing your perspective in detail, it is very helpful.
So to summarize, if I may, Tesla reaching 5k/week run rate some time in Q1 would be a major achievement worthy of automotive automation hall of fame. It would be interesting to see the reaction of OEMs if this comes true.
On a flip side of the coin, if Tesla does not hit 5k/week run rate some time in Q1, I expect a lot of gloating coming from the OEM and potentially a lot of negative market sentiment.
Tesla better deliver.
On a side note, there are no any changes to the VINs registered in NHTSA data base after initial 1100 were discovered registered with NHTSA (as of late yesterday night).
Last edited:
When someone knows what they do not know, I'd think of that person as very competent. This I believe you are.
One example often quoted is Nissan Sunderland.
Two born every minute: inside Nissan’s Sunderland factory
Then there are a few others.
One often referenced as a quick-change expert is Magna, especially Magna Steyr. They do OEM Tier One work for almost everyone and have tooled up to produce specific models, usually low-volume ones, in months. Almost never do they actually have a narrow time schedule in which to build an entirely new vehicle. I mention them because they are also a significant Tesla supplier.
Magna: Customers | About Magna
Then the current reality that BMW and Jaguar both have Magna Steyr building their PUH and some BEV's too.
News Release - Magna Adds New Jaguar E-PACE to Contract Manufacturing Line-up
One thing we know is that your uncertainties about precise timing for introducing a new highly automated process are those of everyone. That both BMW and Jaguar go to Magna makes me think that they'll be on time to market and that they will not have major technical advances.
I really wish we all knew more. I wish I knew more. All the evidence seems to show that only Tesla is right now delivering on serious technical advances.
PR says one thing. The actual automotive supply chain often tells quite a different story.
I don't know if someone has corrected my earlier post. I was wrong about the video and production speed. The video was played at normal speed. Which means that's how the robots are moving at the moment. Once it's confirmed working fine, it will work 10 times faster.
Also if a particular step is done by manual work, the whole production line can still achieve 10,000 cars per week, as long as that step can finish in 1 minute. Or if it takes 5 minutes to finish that step, they can have 5 sets of manual workers to do that step. Anyway, I wouldn't be concerned if there are some manual work in the line.
Turning from 5,000 per week to 10,000 per week will not need to duplicate the production line. They just need to double part of the line. Any robots/stages that can finish their work in 1 minute don't need to be doubled.
Also if a particular step is done by manual work, the whole production line can still achieve 10,000 cars per week, as long as that step can finish in 1 minute. Or if it takes 5 minutes to finish that step, they can have 5 sets of manual workers to do that step. Anyway, I wouldn't be concerned if there are some manual work in the line.
Turning from 5,000 per week to 10,000 per week will not need to duplicate the production line. They just need to double part of the line. Any robots/stages that can finish their work in 1 minute don't need to be doubled.
ValueAnalyst
Closed
That makes perfect sense and explains why going from 5,000/week to 10,000/week costs much less than going from zero to 5,000/week.
- Status
Similar threads
- Sticky
