Tesla and Uber Unlikely to Survive (Vol. 2)
Discussion
TameBritishMuslim said:
So because some videos are more popular than others, you're implying something is up with Sandy Munro? He doesn't control how the YouTube algorithm works but clearly, it shows that people are hungry for Tesla content and that in itself is pretty amazing brand power considering they do not really advertise.
I am absolutely not a conspiracy fan, but please do not underestimate the power of the algorithms to steer, direct and influence - just track the stuff coming out from Facebook to get a glimpse on how they operate. Yes, there is an appetite and I am not denying that. Tesla, and more so Musk, have done a great job of manipulating the mechanism - hence no PR or real advertising - but dont underestimate how much opinions are manipulated like this. Its insidious and having an open mind and awareness of biases is essential more than ever.TameBritishMuslim said:
With respect to his rant about the invasion of Chinese vehicles, that's his opinion and he is far more qualified than you and I to give his opinion. However, we can only see what materialises - and yes, just like Americans switched to Corollas and Camrys, it is possible.. but we'll see.
Yes, and both the Korean and Japanese 'invasions' are used as examples of how things will dramatically change. This could happen, and we will see a big shift in coming decades. But, to draw lines to what has happened in the past to that this will happen again isnt understanding the wider picture. Japanese manufacturers had huge success with a confluence of a number of dramatic changes - but the bottom line is that they created cars that lasted, were affordable and ultimately reliable. Same with the Korean manufacturers - but to expect that this situation will occur again doesnt see how the big 3 have shifted or changed also. And lets be 100% honest here, do we really expect the Korean and Japanese to just sit back? TameBritishMuslim said:
Personally, I don't want legacy manufacturers to die, I want them to change and move forward but many are appearing too greedy to milk ICE for all they can and at the same time they're making it harder to transition. I suspect there will be many mergers like Stellantis and some may go bust and be bailed out like with the laughable US Infrastructure Bill giving huge subsidies for tiny batteries and unions.
I agree - and this is what makes it all interesting. There will be a massive shift. Hell, Chrysler is now a brand that makes a people carrier and Dodge just produces 2 cars (and 1 SUV) - both based on the aging E-class chassis! You could say that its a shadow of itself, and in some way it is. But Jeep & RAM are killing it! Things shift. Saw a great breakdown of Fords use of their 'escape' platform - massively flexible and instrumental in new models like the Bronco Sport, Maverick and others to come. Interesting stuff.... And yes, lets not get onto the politics side - its BS and really badly communicated! I live in California and they couldnt even keep the power on last year (no power cuts this year.... yet). The shift to EV's is here, but jeez, they need to fix the political and infrastructure issues too. Ah crap, now you have got me started.... damn it, taxes are coming... :-(
TameBritishMuslim said:
Yeah, I see the nuance there - i.e. nothing with any factual basis.
It's very simple. His primary business is to sell teardowns of market leading products to competitors who are trying to catch up. Many of his videos, even those which explicitly criticised Tesla, need to contain some information which encourages his primary audience e.g. competitor businesses, to want to buy their full report. That is the context of the original point I was making, which clearly passed you by.EddieSteadyGo said:
TameBritishMuslim said:
Yeah, I see the nuance there - i.e. nothing with any factual basis.
It's very simple. His primary business is to sell teardowns of market leading products to competitors who are trying to catch up. Many of his videos, even those which explicitly criticised Tesla, need to contain some information which encourages his primary audience e.g. competitor businesses, to want to buy their full report. That is the context of the original point I was making, which clearly passed you by.Fair enough.. and if those who buy the report subsequently think that it's b
k$, i'm sure they're not going to stay quiet about it.skwdenyer said:
And the chances of producing a well-designed and thoroughly-debugged vehicle are dramatically improved if the part count is reduced.
For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
Actually it’s completely the opposite For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
Making a single, large complex casting places more emphasis on that part being correct in all aspects of design and is much harder and the barrier to change the casting is enormous. So if there is a deficiency in the design, you simply have to live with it or spend millions changing the casting machine.
It’s not the part count that matters, it’s the complexity of the entirety and that doesn’t seem to have changed. It’s almost the complete opposite to agile development which is how Tesla do all the software.
TameBritishMuslim said:
EddieSteadyGo said:
TameBritishMuslim said:
Yeah, I see the nuance there - i.e. nothing with any factual basis.
It's very simple. His primary business is to sell teardowns of market leading products to competitors who are trying to catch up. Many of his videos, even those which explicitly criticised Tesla, need to contain some information which encourages his primary audience e.g. competitor businesses, to want to buy their full report. That is the context of the original point I was making, which clearly passed you by.Fair enough.. and if those who buy the report subsequently think that it's b
k$, i'm sure they're not going to stay quiet about it.Heres Johnny said:
Actually it’s completely the opposite
Making a single, large complex casting places more emphasis on that part being correct in all aspects of design and is much harder and the barrier to change the casting is enormous. So if there is a deficiency in the design, you simply have to live with it or spend millions changing the casting machine.
It’s not the part count that matters, it’s the complexity of the entirety and that doesn’t seem to have changed. It’s almost the complete opposite to agile development which is how Tesla do all the software.
LolMaking a single, large complex casting places more emphasis on that part being correct in all aspects of design and is much harder and the barrier to change the casting is enormous. So if there is a deficiency in the design, you simply have to live with it or spend millions changing the casting machine.
It’s not the part count that matters, it’s the complexity of the entirety and that doesn’t seem to have changed. It’s almost the complete opposite to agile development which is how Tesla do all the software.
EddieSteadyGo said:
TameBritishMuslim said:
EddieSteadyGo said:
TameBritishMuslim said:
Yeah, I see the nuance there - i.e. nothing with any factual basis.
It's very simple. His primary business is to sell teardowns of market leading products to competitors who are trying to catch up. Many of his videos, even those which explicitly criticised Tesla, need to contain some information which encourages his primary audience e.g. competitor businesses, to want to buy their full report. That is the context of the original point I was making, which clearly passed you by.Fair enough.. and if those who buy the report subsequently think that it's b
k$, i'm sure they're not going to stay quiet about it.
That's all I am saying too! RobDickinson said:
Lol
Your insightful commentary knows no bounds. So explain how a part which contains all the element of a number of historic sub-components makes it easier to ensure all those sub components are designed correctly? Tesla make continual running changes to the design, how’s that going to work if the area that shows up as having a design weakness needs altering and is part of a massive assembly? You work on the premise that Tesla gets things right first time which has been proven time and time again not to be the case because of their design methodology of fix in the field. The single piece casting just means they’ll reliably make the same design mistakes in every car they make.
Heres Johnny said:
skwdenyer said:
And the chances of producing a well-designed and thoroughly-debugged vehicle are dramatically improved if the part count is reduced.
For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
Actually it’s completely the opposite For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
Making a single, large complex casting places more emphasis on that part being correct in all aspects of design and is much harder and the barrier to change the casting is enormous. So if there is a deficiency in the design, you simply have to live with it or spend millions changing the casting machine.
It’s not the part count that matters, it’s the complexity of the entirety and that doesn’t seem to have changed. It’s almost the complete opposite to agile development which is how Tesla do all the software.
My design experience is on aircraft, not cars (penalty for failure is rather higher in the air...). Minimising the number of joining operations is critical to reducing the scope for failure / maintenance problems. You do not design 10 panels riveted and glued together when you can instead have 1 large casting / composite part.
Of course you have to get that part right. But by reducing the part count, you get to spend all your time on getting it right, and less on chasing all the potential tolerance and assembly issues related to multiple parts with multiple joints.
And that is precisely the point Sandy was making in his various Tesla tear-downs - too many parts creates too much cost, too much weight, too many places for potential quality and durability problems to creep in.
But, no, I haven't worked on designing for anything as low-cost and low-risk as a car, so my experience may not be relevant here
skwdenyer said:
And the chances of producing a well-designed and thoroughly-debugged vehicle are dramatically improved if the part count is reduced.
For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
Not if an extra part reduces fatigue overall. Your comment is incredibly simplistic.For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
jsf said:
skwdenyer said:
And the chances of producing a well-designed and thoroughly-debugged vehicle are dramatically improved if the part count is reduced.
For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
Not if an extra part reduces fatigue overall. Your comment is incredibly simplistic.For a given programme length / budget (a typical constraint), reducing the part count is going to provide a major advantage.
But in my experience (YMMV), increasing the part count with additional joining methods tends to increase the chances of fatigue problems. It is far easier to ensure the finished assembly more closely approximates simulations if the number of variable is reduced. To take the example we're discussing here, the large rear casting has a number of variables (mould filling, material consistency, etc.) that are (relatively speaking) easy to constrain, test, and standardise. Whereas producing the same assembly from a number of stamped, hydroformed or what have you parts all connected by a mixture of self-piercing rivets, bonding, welding and so on leaves one much more exposed to a whole bunch of tolerances and sources of variance that can quickly stack up in ways that are incredibly hard to test for.
As regards your specific comment about fatigue, I don't think anyone is suggesting ignoring the S-N curves for the sake of reducing part count
As before, of course, this means you have to do the design work properly. There seems to be an assumption here that that can never be done and, therefore, that being able to iterate in production is mandatory - which is IME a sign of too much complexity for the available design / development time.
Heres Johnny said:
Your insightful commentary knows no bounds.
So explain how a part which contains all the element of a number of historic sub-components makes it easier to ensure all those sub components are designed correctly? Tesla make continual running changes to the design, how’s that going to work if the area that shows up as having a design weakness needs altering and is part of a massive assembly? You work on the premise that Tesla gets things right first time which has been proven time and time again not to be the case because of their design methodology of fix in the field. The single piece casting just means they’ll reliably make the same design mistakes in every car they make.
That's a fair point. But equally reducing the part count means it is much easier to get the design right, because there are far fewer potential manufacturing variances, far fewer tolerances to stack up, and so on.So explain how a part which contains all the element of a number of historic sub-components makes it easier to ensure all those sub components are designed correctly? Tesla make continual running changes to the design, how’s that going to work if the area that shows up as having a design weakness needs altering and is part of a massive assembly? You work on the premise that Tesla gets things right first time which has been proven time and time again not to be the case because of their design methodology of fix in the field. The single piece casting just means they’ll reliably make the same design mistakes in every car they make.
And that is rather the point of why it has been done.
If we take your view, things like say BMW i3 shells would be the worst of all possible worlds, because running changes are just not that easy to achieve. Or are you trying to say this is a Tesla-specific issue, and that Tesla is not allowed to iterate its own design methodology?
skwdenyer said:
Of course it is harder to fix a poorly-designed part once you've invested in all the machinery to make it. We weren't talking about iterative design in real-time using production parts like some sort of piece of Silicon Valley software.
And thats a concern for me. You can argue about the large scale pressing stuff - I cant, as I simply dont know anything about that! But, its the simple stuff that worries me. Door handles! I dont know if they are an in house or external supplier part, but the stubborn commitment to particular components (especially if they have heavily invested in in house manufacture) means that they dont fix some stuff. Model S door handles being one of them. Nice idea, but even on their forth iteration, they still cant get it right! Its a bloody door handle! How hard can that be? There were constant complaints for the v1 and v2 versions. I understand that v3 was better, but even on the Plaid, owners are complaining that they dont open correctly, fail (STILL, seriously, they still keep failing? WTAF?) and now owners are complaining that the lighting for them is inconsistent and need constant fixing!
And thats before we get to other simple things like the Model X falcon doors, Model X boot lid alignment, window and door seals etc etc.... when you are invested into in house manufacturer, you are less inclined to make changes, look for incremental small things and given the drive to profit, are actually incentivized to not do things. Oh, we will fix that post sale.... well thats nice.....
While nice to issue bug fixes and changes, thats all good. But the damn thing should work from the beginning - not after patch 3. And of course, the same applies to other manufacturers too!
skwdenyer said:
Of course it is simplistic - this is a PH forum thread - I'd be delighted to take this off into a new thread discussing the engineering of BIW, but it seems like it might derail this thread a bit.
But in my experience (YMMV), increasing the part count with additional joining methods tends to increase the chances of fatigue problems. It is far easier to ensure the finished assembly more closely approximates simulations if the number of variable is reduced. To take the example we're discussing here, the large rear casting has a number of variables (mould filling, material consistency, etc.) that are (relatively speaking) easy to constrain, test, and standardise. Whereas producing the same assembly from a number of stamped, hydroformed or what have you parts all connected by a mixture of self-piercing rivets, bonding, welding and so on leaves one much more exposed to a whole bunch of tolerances and sources of variance that can quickly stack up in ways that are incredibly hard to test for.
As regards your specific comment about fatigue, I don't think anyone is suggesting ignoring the S-N curves for the sake of reducing part count
As before, of course, this means you have to do the design work properly. There seems to be an assumption here that that can never be done and, therefore, that being able to iterate in production is mandatory - which is IME a sign of too much complexity for the available design / development time.
I spend my life working on racing cars where often the designer screwed up and you have to add structure to eliminate a problem, Lotus F1 cars are hilarious where you have single components used for multiple jobs, my favourite is rear brake calipers as suspension mount points. But in my experience (YMMV), increasing the part count with additional joining methods tends to increase the chances of fatigue problems. It is far easier to ensure the finished assembly more closely approximates simulations if the number of variable is reduced. To take the example we're discussing here, the large rear casting has a number of variables (mould filling, material consistency, etc.) that are (relatively speaking) easy to constrain, test, and standardise. Whereas producing the same assembly from a number of stamped, hydroformed or what have you parts all connected by a mixture of self-piercing rivets, bonding, welding and so on leaves one much more exposed to a whole bunch of tolerances and sources of variance that can quickly stack up in ways that are incredibly hard to test for.
As regards your specific comment about fatigue, I don't think anyone is suggesting ignoring the S-N curves for the sake of reducing part count
As before, of course, this means you have to do the design work properly. There seems to be an assumption here that that can never be done and, therefore, that being able to iterate in production is mandatory - which is IME a sign of too much complexity for the available design / development time.
off_again said:
skwdenyer said:
Of course it is harder to fix a poorly-designed part once you've invested in all the machinery to make it. We weren't talking about iterative design in real-time using production parts like some sort of piece of Silicon Valley software.
And thats a concern for me. You can argue about the large scale pressing stuff - I cant, as I simply dont know anything about that! But, its the simple stuff that worries me. Door handles! I dont know if they are an in house or external supplier part, but the stubborn commitment to particular components (especially if they have heavily invested in in house manufacture) means that they dont fix some stuff. Model S door handles being one of them. Nice idea, but even on their forth iteration, they still cant get it right! Its a bloody door handle! How hard can that be? There were constant complaints for the v1 and v2 versions. I understand that v3 was better, but even on the Plaid, owners are complaining that they dont open correctly, fail (STILL, seriously, they still keep failing? WTAF?) and now owners are complaining that the lighting for them is inconsistent and need constant fixing!
And thats before we get to other simple things like the Model X falcon doors, Model X boot lid alignment, window and door seals etc etc.... when you are invested into in house manufacturer, you are less inclined to make changes, look for incremental small things and given the drive to profit, are actually incentivized to not do things. Oh, we will fix that post sale.... well thats nice.....
While nice to issue bug fixes and changes, thats all good. But the damn thing should work from the beginning - not after patch 3. And of course, the same applies to other manufacturers too!
As regards the door handles, I agree with you - I think all these pop-out handles are ludicrous. The "it's a bloody door handle" should preclude all that stuff. Also I'm clearly getting old, as I thought that things like EuroNCAP required doors to be openable after a crash. How does that work if the electrical systes are destroyed in the impact? I've seen a tear-down of a Tesla door handle; I don't recall seeing any pyrotechnic "pop out devices" to present them in an accident.
jsf said:
I spend my life working on racing cars where often the designer screwed up and you have to add structure to eliminate a problem, Lotus F1 cars are hilarious where you have single components used for multiple jobs, my favourite is rear brake calipers as suspension mount points.
And how much simulation and testing went into designing those parts? I thought Chapman felt any part that wasn't on its last legs during the cool-down lap was over-designed skwdenyer said:
As regards the door handles, I agree with you - I think all these pop-out handles are ludicrous. The "it's a bloody door handle" should preclude all that stuff. Also I'm clearly getting old, as I thought that things like EuroNCAP required doors to be openable after a crash. How does that work if the electrical systes are destroyed in the impact? I've seen a tear-down of a Tesla door handle; I don't recall seeing any pyrotechnic "pop out devices" to present them in an accident.
Yeah, I agree. There are obvious areas in which we can innovate and bring convenience. But this shouldnt be at the excuse basic functionality. Its not because you are getting old, I am frequently finding more and more people who prefer simple things - door handles are a great example. skwdenyer said:
And how much simulation and testing went into designing those parts? I thought Chapman felt any part that wasn't on its last legs during the cool-down lap was over-designed
Plenty, i work on all sorts of kit including very current race cars, it's an ongoing theme. The update list on some of the kit i work with is massive, despite the best designers in the business working on them.There are plenty of updates on Aircraft too, as i am sure you know.
On production cars you cant beat thorough harsh environment testing, no matter what you design models tell you.
skwdenyer said:
And how much simulation and testing went into designing those parts? I thought Chapman felt any part that wasn't on its last legs during the cool-down lap was over-designed
How much simulation and testing went into the software release(s) they had to recall? Or the seat belt mounts? You know... safety critical stuff like that.Honestly, they are just a car company - and quite capable of messing up like any other car company. Things like 'bullet proof' glass that isn't are not really a big deal, but they should be a reminder that they are not above criticism, or above making very silly mistakes.
Tuna said:
How much simulation and testing went into the software release(s) they had to recall? Or the seat belt mounts? You know... safety critical stuff like that.
The seat belt mounts were extensively designed and tested. AIUI it wasn't a failure of components or design engineering; it was instead a failure of fastening installation (or more specifically a problem of exception-triggered inspection).The software is fundamentally different. I don't have to agree with everything Tesla do to recognise that their ability to iterate software is rather different to its ability to iterate hardware.
Tuna said:
Honestly, they are just a car company - and quite capable of messing up like any other car company. Things like 'bullet proof' glass that isn't are not really a big deal, but they should be a reminder that they are not above criticism, or above making very silly mistakes.
And I've never said they're not. I'm just a little bemused by this constant knocking of them when they attempt to evolve stuff that was ripe for evolution. If Mercedes, say, had come out with a large-scale cast component, I've no doubt it would be viewed as a step-change in production engineering.Gassing Station | EV and Alternative Fuels | Top of Page | What's New | My Stuff