Highest production VIN in the wild

[tomas]

All I know is this. Santa Barbara Tesla lot must have 20 model 3s waiting for delivery. Usually empty. There appears to have been a step change. I spoke to one of the DSs and they are up to their eyeballs.

No I didn’t see VINs, they don’t let you back there.

 

[insaneoctane]

Did Tesla guidance of 2,500/week specify deliveries or just production rate?

EDIT: I answered my own question ...

"We continue to target weekly Model 3 production rates of 2,500 by
the end of Q1..."

 

[Reciprocity]

Did Tesla guidance of 2,500/week specify deliveries or just production rate?

Production.

 

[Barklikeadog]

Did Tesla guidance of 2,500/week specify deliveries or just production rate?

What will be discussed is production I'm sure... it has to be a madhouse in the plant right now. They want to get the largest number possible to say that 'we did xxx'

 

[amoney805]

All I know is this. Santa Barbara Tesla lot must have 20 model 3s waiting for delivery.

Can Model 3's be picked up there? I thought Fremont and costa mesa were the only options?

 

[Waiting4M3]

You're right, unless there was a production rate increase (similar to nhtsa vin assignment rate increase).

I'm hopeful we'll see 14000 assigned early next week.

One more thing to bear in mind is at what stage of production the VINs are assigned out. Our understanding here is that Tesla normally seems to assign VINs after the car is out of production, they match the cars to available buyers and deliver. What they could have been doing in the last few weeks, Tesla could be assigning VINs from cars still in production, just to give themselves extra lead time on finding buyers that can take delivery before the end of the quarter, so that could explain the reduction from 2.5 wks down to 11 days in those 2 examples I gave. So if Tesla now goes back to the old method, we could be looking at 2+ wks lead time again.

P.S. the overall production rate could be increasing, but I don't know if that necessarily scales directly to how much time each car takes to go through the line. I'm envisioning a freeway with speed limit of 65mph, when it gets busier it just has more cars on it, but the speed of each car is still ~65mph. Tesla could just be jamming more cars in the production line.

 

[Reciprocity]

All I know is this. Santa Barbara Tesla lot must have 20 model 3s waiting for delivery. Usually empty. There appears to have been a step change. I spoke to one of the DSs and they are up to their eyeballs.

No I didn’t see VINs, they don’t let you back there.

They have been blocking the VINs with a piece of paper work anyway. Atleast in places I have seen them recently.

 

[mongo]

P.S. the overall production rate could be increasing, but I don't know if that necessarily scales directly to how much time each car takes to go through the line. I'm envisioning a freeway with speed limit of 65mph, when it gets busier it just has more cars on it, but the speed of each car is still ~65mph. Tesla could just be jamming more cars in the production line.

You can only shove a few extra cars into the line until the extra space is consumed. After that, they have to come off the end of the line. It's a one lane road that starts out almost bumper to bumper (and can only hold <300 cars total, maybe).

 

[HAL8950]

VIN 11352 posted on @TheTeslaLife

 

[SlicedBr3ad]

You can only shove a few extra cars into the line until the extra space is consumed. After that, they have to come off the end of the line. It's a one lane road that starts out almost bumper to bumper (and can only hold <300 cars total, maybe).

I completely agree with this. Production rate will directly translate into delivery rate although possibly with a week/two at most delay. It will be telling if we going to continue to see 7-13 car carriers at random times of day in the Fremont parking lot as was posted this week.

 

[Webeevdrivers]

I completely agree with this. Production rate will directly translate into delivery rate although possibly with a week/two at most delay. It will be telling if we going to continue to see 7-13 car carriers at random times of day in the Fremont parking lot as was posted this week.

Obviously the factory doesn't run 24/7 But what is it in terms of shifting. 16/6 etc. ???

 

[Waiting4M3]

You can only shove a few extra cars into the line until the extra space is consumed. After that, they have to come off the end of the line. It's a one lane road that starts out almost bumper to bumper (and can only hold <300 cars total, maybe).

I think of the production line as a one lane road going through a series of toll collection points where it forks into multiple toll booths then merge back together into 1 lane. In production line typically people measure 2 different kinds of times, queue time and process time. In the toll booth analogy, the queue time is the time it takes to drive from one toll point to the next, and the process time it takes to pay toll at each toll point.

The queue time could be determined by both the limit on the conveyor carrying the cars (think of it as road speed limit), and also by the congestion at each "toll booth". Since the limit now seems to be the robot speed, I think we're "bumper to bumper" due to the congestion at the "toll booths", so when ramping the process time, it would also help reduce the congestion, so the queue time should also improve.

I hope the sharp increase through the middle of the "S" curve will translate to quicker VIN assignment after VIN registration. I think when we reach 3-4k/wk production, we should start to approach the steady state of VIN registration to assignment delay, and can back out a production TAT, which will be useful in calculating Tesla's cash flow situation.

 

[Waiting4M3]

Obviously the factory doesn't run 24/7 But what is it in terms of shifting. 16/6 etc. ???

It's seems to me that they are actually running 24/7 if you believe what people report on the internet. I did some digging and discussed it in the Market Action thread.

 

[Webeevdrivers]

It's seems to me that they are actually running 24/7 if you believe what people report on the internet. I did some digging and discussed it in the Market Action thread.

So if they hit this 300 per day number they are shooting for that is about 2000 per week allowing for some line maintenance. It falls short of the 2500 in the guidance. Is the plan to use a second line to get to 5000? I would think what ever number they report for weekly production better begin with a 2 (eg 2100 2150 etc) to calm investor and market jitters.

The cash burn rate will probably looked at this quarter as well. It has to be dropping at this point although if they need to buy a second line that could still indicate big bills coming.

Fun to follow. Wishing them the best.

 

[mongo]

I think of the production line as a one lane road going through a series of toll collection points where it forks into multiple toll booths then merge back together into 1 lane. In production line typically people measure 2 different kinds of times, queue time and process time. In the toll booth analogy, the queue time is the time it takes to drive from one toll point to the next, and the process time it takes to pay toll at each toll point.

The queue time could be determined by both the limit on the conveyor carrying the cars (think of it as road speed limit), and also by the congestion at each "toll booth". Since the limit now seems to be the robot speed, I think we're "bumper to bumper" due to the congestion at the "toll booths", so when ramping the process time, it would also help reduce the congestion, so the queue time should also improve.

I hope the sharp increase through the middle of the "S" curve will translate to quicker VIN assignment after VIN registration. I think when we reach 3-4k/wk production, we should start to approach the steady state of VIN registration to assignment delay, and can back out a production TAT, which will be useful in calculating Tesla's cash flow situation.

I think the parallel booth will apply once they increase equipment for 10k, then it will be separate lanes with duplicate tool booths ( possibly short sections, possibly long sections).

I'd characterize the two times as time at each booth (process time) like you mentioned, but the other I would say is the time from the first assembly operation until the car rolls off the line. (or on ramp at start to off ramp at destination).

For example, it could take 3 hours to build a car with the process taking 36 steps of 5 minutes each. The buffer between steps would increase the total time for one car, but the cars would still come off the line every 5 minutes. So the line runs at the rate of the slowest step, and the queue increases the total time a single car is on the line.

 

[mongo]

So if they hit this 300 per day number they are shooting for that is about 2000 per week allowing for some line maintenance. It falls short of the 2500 in the guidance. Is the plan to use a second line to get to 5000? I would think what ever number they report for weekly production better begin with a 2 (eg 2100 2150 etc) to calm investor and market jitters.

The cash burn rate will probably looked at this quarter as well. It has to be dropping at this point although if they need to buy a second line that could still indicate big bills coming.

Fun to follow. Wishing them the best.

The line is designed to do 5k on its own, the slowdown has been in pack supply which should be alleviated with the new GF1 equipment. Fremont will then be continually updating and tuning the line to keep up.
Hopefully we see or hear more at the Q1 call/letter.

 

[iwannam3]

"One more thing to bear in mind is at what stage of production the VINs are assigned out. Our understanding here is that Tesla normally seems to assign VINs after the car is out of production, they match the cars to available buyers and deliver. "

To prevent theft aren't their Vins stamped on multiple places on the vehicle and drive train? Matching numbers is the holy grail on collector cars where the VIN matches the engine number. My '69 Camaro had VINS stamped on the block, inside the air vent between the hood and window, you could see it, but not access it, another one hidden by the heater fan. No way are they building a car and then pop riveting a VIN on. Windshield is in the way of accessing the VIN plate. Roadster has a VIN stamped in the door sill.

 

[mongo]

"One more thing to bear in mind is at what stage of production the VINs are assigned out. Our understanding here is that Tesla normally seems to assign VINs after the car is out of production, they match the cars to available buyers and deliver. "

To prevent theft aren't their Vins stamped on multiple places on the vehicle and drive train? Matching numbers is the holy grail on collector cars where the VIN matches the engine number. My '69 Camaro had VINS stamped on the block, inside the air vent between the hood and window, you could see it, but not access it, another one hidden by the heater fan. No way are they building a car and then pop riveting a VIN on. Windshield is in the way of accessing the VIN plate. Roadster has a VIN stamped in the door sill.

The original comment may have been regarding when a VIN is registered and/or assigned to a buyer, not when it gets paired to a body.

The unassembled vehicle has a fixed VIN once the dash is mated. Given drive unit and pack replacements, there may not be many/ any other locations other than the ones called out in the VIN decoder.

 

[Waiting4M3]

I think the parallel booth will apply once they increase equipment for 10k, then it will be separate lanes with duplicate tool booths ( possibly short sections, possibly long sections).

I'd characterize the two times as time at each booth (process time) like you mentioned, but the other I would say is the time from the first assembly operation until the car rolls off the line. (or on ramp at start to off ramp at destination).

For example, it could take 3 hours to build a car with the process taking 36 steps of 5 minutes each. The buffer between steps would increase the total time for one car, but the cars would still come off the line every 5 minutes. So the line runs at the rate of the slowest step, and the queue increases the total time a single car is on the line.

Yes you can measure total time, and it's the sum of process time + queue time, they are equivalent. Below is a simple diagram depicting 2 steps. With a mismatch of process time of 10 min in step 2 vs 5 min in step 1, the total time balloons from 20 min without congestion to 45 min with full congestion. You can measure the queue time by either subtracting the process time (15 min) from the total time (45 min), or directly measure how long a car sits on the conveyor, both give the same #.

My point here is that if you can speed up step 2 process time by 5 min, you can eliminate the congestion and reduce the queue time from 30 min down to 5 min, it's a huge increase. So I hope by ramping up the robot speed, the total time a car stays in production can be reduced much more than just the improvements on the process time.

 

[mongo]

Yes you can measure total time, and it's the sum of process time + queue time, they are equivalent. Below is a simple diagram depicting 2 steps. With a mismatch of process time of 10 min in step 2 vs 5 min in step 1, the total time balloons from 20 min without congestion to 45 min with full congestion. You can measure the queue time by either subtracting the process time (15 min) from the total time (45 min), or directly measure how long a car sits on the conveyor, both give the same #.

My point here is that if you can speed up step 2 process time by 5 min, you can eliminate the congestion and reduce the queue time from 30 min down to 5 min, it's a huge increase. So I hope by ramping up the robot speed, the total time a car stays in production can be reduced much more than just the improvements on the process time.

View attachment 290467

I think you are making it too complicated. If step 2 takes 10 minutes, you can only make a car every 10 minutes regardless of conveyor size. Once the line is full, the queue times do not impact production rate, unless the conveyor itself is the bottle neck (for example, can only move one assembly at a time, and it that takes more than 5 minutes).

The buffering does help with process synchronization (having material ready when the step is available). In your example, if you don't fill the buffer in the conveyor, each car has only a 20 minute build time. Less queue time, same build rate.