Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

2017 Investor Roundtable:General Discussion

This site may earn commission on affiliate links.
Status
Not open for further replies.
Slow downs in one part of production does not preclude Tesla from making improvements to other parts of the automation. Certainly there are limits to what you can do when one part is very slow or stopped or missing parts, but I'm sure they are not just sitting on their hands waiting.

On the last call Elon did say that they expected to get to thousands per week by the end of the year. This could be the minimum level where the automation is running but not as fast as theoretically possible. Which brings me to a question of just how you speed it up. Is it all just adjusting the speed and timing on the software? Tesla seems to be allowing for 3 months or one quarter to go from a guestimated 2k/w at the end of your treat to 5k/w at the end of the 1st quarter next year. Elon had mentioned that the speed of the robots was so high that mass is the moving arm was an issue due to inertia and how the supplier has never asked to increase the speed of the robots. I remember a quote about needing a strobe light to see the robots but I think that was related to the gigafactory and not Fremont.

They must have tested some of these processes before making the final decision on the robots as that would be a show stopper. Some is those welding stations have several arms and they must stop long enough to actually complete the weld. How do you make that 2.5x then 5x faster without running into timing issues. You have to imagine they aren't just relying on timing and could infact have a feed back loop and use machine learning and ai and maybe even vision or other sensors to get feed back on where exactly the arm is that is in the way for the next step. It cant be as simple as just turning the speed dial to eleven. Tesla might have even ordered custom robots with higher quality motors and lighter materials on the arm. Tooling could also be special in that they are lighter. This is where Space X advantages come into play.

Simple. Use 5 times as many robots, and let them weld in parallel.

The computer industry switched 15 years ago from a system where 50% more performance meant 10x higher cost, to a system where you just multiplicate the hardware. For Telsa production line, similar arguments can be made.
I don’t want to say that the same robot can’t be reprogrammed to do the same thing 5x faster. I just want to say that it may be easier/cheaper (and certainly more predictable if you have a robot already doing what you want) to just use 5 robots. In my line of business, performance is only looked at if it turns out to be really bad, otherwise throwing hardware at the problem is the most economical thing to do.
 
  • Like
Reactions: ValueAnalyst
More accurate maps?
Better routing?
Better traffic aware routing?
Multiple waypoints?
More options? Shortest/quickest; Avoid toll/highway/etc.

Predictive navigation. Get in your car at 7AM on a Wednesday, it automatically routes you to work. At 6PM, it routes you to recent restaurants you have been to, and comparable restaurants you would like.
 
One of the comments says that one of the changes coming is that you will be able to plan the route in your phone and send it to the car.


Well if the update goes hand in hand with autopilot update, then this would be really cool : the car does all the driving right from the start, and just warn you when you need to take the lead when needed.
 
"
What kind of navigation system can be "light-years away?" I'm not imaginative enough to predict what's coming, but maybe some here are?

Heads up Display?

Right now i have most faith in Google Maps, especially because they use live data for traffic. Wouldn't be like Elon to make friends with another company, so it must be something proprietary. Problem is, not enough users on the road to be as accurate as Google.
 
Last edited:
Anybody here predicting Tesla will hit the elusive 30% gm on model S for q4?

Last earnings call they said they were going to be slightly scaling back on the S &X lines to focus on the 3. I’m hoping they eliminated overtime and what not on the S &X line to increase margins.

It seems like we might actually be set up to hit the 30% gm on the S and hopefully close on the X.

Thoughts?
I think the gigafactory slowdown led them to put m3 resources on S/X production, especially early in the quarter, explaining the big Europe deliveries. Europe deliveries for December are on track for 5000+. Haven't heard from Bonaire or schonelucht on vin count, but Europe data points to over 27000 S & X. Not sure if low interest loans were available in all of Europe or just Norway.
 
  • Like
Reactions: neroden
Slow downs in one part of production does not preclude Tesla from making improvements to other parts of the automation. Certainly there are limits to what you can do when one part is very slow or stopped or missing parts, but I'm sure they are not just sitting on their hands waiting.

I am in agreement with you. My point was that the line could be limited by supplier's rate, not Fremont's. If the pack supply was the bottle neck, and the cell production is good, then the vehicle build rate can jump to the line max once the pack's arrive, a step function vs a ramp.

As for cranking up the line speed (which you pretty much covered), physics is going to be the limiting factor on minimal time for an assembly option. How fast can the robot move repeatably over time and what is the settling time of the material/ process. The video of the welding cell showed large end actuators for the pinch weld. That is going to limit speed and settling time. Reducing mass and adjusting movement profiles to minimize inertia will allow faster movements. Applying a motion profile that counters the natural resonance will reduce the settling time.

2nd level issue will be the material flow path to the assembly line...

Maxing out post topics:
Tesla could have been using the AP fleet as a mapping/ digitizing system. Perhaps the new system is a continuous 3-D vehicle POV system (like a continuous street view). Or linking nav to autopilot for lane changing? Live traffic map from all the other Tesla on the road?
 
Maybe some kind of live views as well ?

Hmm that would be interesting.

Well if the update goes hand in hand with autopilot update, then this would be really cool : the car does all the driving right from the start, and just warn you when you need to take the lead when needed.

Agreed. Level 3? One can only hope...

Predictive navigation. Get in your car at 7AM on a Wednesday, it automatically routes you to work. At 6PM, it routes you to recent restaurants you have been to, and comparable restaurants you would like.

This makes sense. That would be a big upgrade.

Elon is prone to exaggeration.. space ship controls and the like. Probably just higher resolution display and better traffic integration.

Sigh... you’re probably right.
 
Anyone notice the model 3 vin's going up? At least two over 2600 and more over 2000. Still filling in some lower sub 1000, but new vin's seem more common in the 2000 range now, versus sub 2000 last week. This is my new OCD page along with Norway registration site.

Highest I see is 2802, vin assigned, no delivery date, yet.

VIN Assignment

And this

U.S., Model 3 invites to configure
 
Status
Not open for further replies.