Great architectural comparison
- Thread starter UncleDave
- Start date
This amplifies 2 key Tesla advantages:
EV from the ground up; no compromises.
Pure play EV company; no allocation fight over scarce resources.
EV from the ground up; no compromises.
Pure play EV company; no allocation fight over scarce resources.
Yup the perils and costs of electrifying an ICE car vs a ground up approach.
" gotta use those because we bought a ton of them"
Yup I have the answer to the question I asked prior about how deep the Mach E's OTA goes and why numerous issues and fixes require a trip back to the dealer. Syncrhonozing and enabling 50 ECU's to a central piece of code is a nightmare pursuit only worth descending so far into before it becomes hundreds of man years to write an maintain
I look at the design of most car brands and what the designers do and pray their significant other gives them a V.D. that has no cure. Who designs a vehicle that the cab has to be removed for example. Usually the less stuff there is the less can potentially go wrong. Tesla's designers aren't Ford or VW designers. I went through lots of architecture changes what working on the electric forklifts and the Forklifts basically were the same but they went from resistor speed control and mechanical safety [for example is there was a short] it would shut off the power so the forklift wouldn't run away to solid state controls to micro processer controls and MOSFETS for speed and directional control which eliminate the forward and reverse contactors and top speed contactors in one nifty box.
Last edited:
How long will it take for the guys fixing them to learn the systems. There must be a huge learning curve.
In the end it really is the final result that matters. Is common off the shelf architecture better or is a greenfield design with no legacy constrain better? Depends on how many units you need and how much R&D you want to pay for (as well as how much risk you wants to tolerate).
Typically, greenfield design would have bugs, lots of bugs, that takes years to find out and fix. As you can see from Tesla's flash memory fiasco you end up with recalls and unhappy owners who cannot just ignore it and keep driving. Legacy multiple pieces design may have some problem but if 1/2 of them are proven in existing platform they do not need to replace them all at once with new designs that have teething problems. Need an EV? Just swap out maybe 1/3 of them, and leave the rest of the design the same and share them with the 80% of your existing product line.
Risk of failure is also lower if you only need to recall and replace 1 out of 20 units that cost you $200 instead of a monster PCB that cost you $7k.
Gradually you can integrate some of them and consolidate some of the modules into one units, after they are mature in existing design.
Typically, greenfield design would have bugs, lots of bugs, that takes years to find out and fix. As you can see from Tesla's flash memory fiasco you end up with recalls and unhappy owners who cannot just ignore it and keep driving. Legacy multiple pieces design may have some problem but if 1/2 of them are proven in existing platform they do not need to replace them all at once with new designs that have teething problems. Need an EV? Just swap out maybe 1/3 of them, and leave the rest of the design the same and share them with the 80% of your existing product line.
Risk of failure is also lower if you only need to recall and replace 1 out of 20 units that cost you $200 instead of a monster PCB that cost you $7k.
Gradually you can integrate some of them and consolidate some of the modules into one units, after they are mature in existing design.
Not sure I agree. A ground up approach vs a compromise approach?
I can tell you with 100% certainty that fixing and retrofitting old code is far harder and dangerous than writing from scratch with a single purpose.
I do this for a living. Low level code, firmware, can be the worst...
To a certain extend. In the long run it is going to be how much risk and how much resource you want to spend on. Teething problem on new clean design vs proven design that you already have maturity. It is like asking whether an old Toyota with 4 speed auto is more reliable or not.
Disagree. Ford is compromising by using old architecture.
As you posted in #6, what matters is the result. Teslas are more efficient than Ford or VW and probably more dependable.
And definitely more advanced as shown by their OTA superiority. Tesla's technology is years ahead of the pack.
Fewer parts is better just about any and every way you look at it....reliability, cost/profit, weight, ease speed of assembly, diagnostics, updating, fastening, repair....
compound this with the giant casting machine removing parts and assembly time from hundreds of other separate or welded pieces.
the Y is massively different - other 2 are actually fairly close from an electrical standpoint.
compound this with the giant casting machine removing parts and assembly time from hundreds of other separate or welded pieces.
the Y is massively different - other 2 are actually fairly close from an electrical standpoint.
Last edited:
But as soon as someone gets a bill of multiple $$ thousands for one of those large integrated modules they'll be moaning about the cost.
Roughly half of CAN-controlled peripherals will have nothing to do with the car being an EV and others replace legacy ICE items such as transmission and engine controllers. Some items are located at physically remote parts of the car and must have some sort of appropriate ruggedised comms.
I see these differences simply as alternative engineering choices and no technical group is going to pick an architecture that they think is prone to failure, or is unnecessarily expensive to manufacture. I don't recall the panel of experts had any specific conclusions between these cars when I watched this when it first came out. Sandy himself is way out of his depth in this subject and can't hide his dislike for the VW because he couldn't figure out dash menus that any teenager could understand.
Roughly half of CAN-controlled peripherals will have nothing to do with the car being an EV and others replace legacy ICE items such as transmission and engine controllers. Some items are located at physically remote parts of the car and must have some sort of appropriate ruggedised comms.
I see these differences simply as alternative engineering choices and no technical group is going to pick an architecture that they think is prone to failure, or is unnecessarily expensive to manufacture. I don't recall the panel of experts had any specific conclusions between these cars when I watched this when it first came out. Sandy himself is way out of his depth in this subject and can't hide his dislike for the VW because he couldn't figure out dash menus that any teenager could understand.
Implementation matters.
The worst you want is either under funding for a new architecture, or underfunding for an existing architecture. When you have that you have half butt implementation that aren't tested, and let your customers be the one holding the bag. When in doubt, you can throw in more processing power, memory, keep things simple but sandbag more resources, etc. I have seen success in both types of design philosophies and I have seen both types screw up because they try to push something too far.
Risk assessment and mitigation matters. When in doubt pick the devil you know or something already proven.
The worst you want is either under funding for a new architecture, or underfunding for an existing architecture. When you have that you have half butt implementation that aren't tested, and let your customers be the one holding the bag. When in doubt, you can throw in more processing power, memory, keep things simple but sandbag more resources, etc. I have seen success in both types of design philosophies and I have seen both types screw up because they try to push something too far.
Risk assessment and mitigation matters. When in doubt pick the devil you know or something already proven.
You understand that car manufacturers pay Munro's company, Munro and Associates, to reengineer their products for better manufacturing and improvements, right?
I have made this point before, and continue to believe it. Most, but not all of new EV's from major, legacy automakers are going to be electrified modifications of their existing vehicles using many common parts with their ICE models. I get it. They simply lack the capital to build EV's from the ground up in most cases.
They all see the writing on the wall and is much as good intentions count for when they want to offer EV's to the car buying public, it is going to be extremely difficult to play catch up with Tesla unless they can come up with the capital. The automakers are not like the current Washington administration who can simply print money. And even if the many many billions of dollars somehow became available it would be years before they could introduce an all new model and not some kludge of an existing model.
Tesla will dominate for a long time.
They all see the writing on the wall and is much as good intentions count for when they want to offer EV's to the car buying public, it is going to be extremely difficult to play catch up with Tesla unless they can come up with the capital. The automakers are not like the current Washington administration who can simply print money. And even if the many many billions of dollars somehow became available it would be years before they could introduce an all new model and not some kludge of an existing model.
Tesla will dominate for a long time.
Last edited by a moderator:
The likelihood of failure is always lower with fewer parts. Do you have any idea what the replacement modules cost or are you guessing?
The guys on the panel were pretty clear that the lower part count and design of the Y was the way to go. The question kept coming up as to why VW and Ford would do it the way they did, and parts bin engineering and lack of a clean sheet design preventing tighter integration was usually the answer.
Sandy brings in people where he is weak like all good experts - even on his own team he has a guy for this and that.
If he can figure out a Tesla and the Ford the VW should be equally easy right?
The VW has been buggier than most from day 1 and search for ID4 software bugs/ problems fills the screen.
I am sure they have a lot of outside vendors helping their staff. I didn't know that about Munro. This is the second time I am watching this same video guys around a table talking about placing smaller modules around versus a bigger one somewhere. I also watched his test drive of the Mach E. For someone involved in the engineering of it, he said they did a good job with this and was surprised. I thought he was like a car dealer or reviewer, thanks for info. You guys own some TSLA I get it, or maybe the car. Promote Tesla in any way. Turns out F bought awhile back would have been sweet too. It was at 3, wasn't it? I am surprised how many people on this site make money trading stocks, or are trying to. Where is GM on the chart? People do realize the Bolt was made from the ground up, wasn't it?
My understanding is the Bolt was designed by Daewoo (GM Korea). I believe it was built from the ground up which is costly but has resulted in a great EV. They were introduced in CA in Dec 2016. They sell like crazy; my next door neighbor is on her 2nd lease; she likes it better than our Model 3.
I wonder what she likes more. I almost bought a Bolt because of the discounts. It is a bit small and narrow, reviews say the ride is choppy, and it has no heat pump. They were selling to staff at the dealer the price was so low. That was before the battery fire thing came out. I liked it though, but don’t want a pure ev and an old pickup truck as my only choices to drive. I do want two cars as through Costco insurance I insure two cars for about $100 less per year than one. I am not against Tesla btw it is another American asset. We need those.
I see Tesla is doing another recall. This time for brake caliper bolts not being tightened correcctly on some Model S and Model Y of certain years.
I think at least Ford knows how to do this correctly by now.
I think at least Ford knows how to do this correctly by now.