Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/chip-makers-have-a-message-for-car-makers-your-turn-to-pay.16495/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Chip makers have a message for car makers: Your turn to pay

Can auto makers really change their behavior and start collaborating with semiconductor industry for a long term relationship? Or, are they still thinking they are the center of the universe?

Very good article and yes that is what happened. We did not have a wafer shortage problem, we had an allocation problem. Car makers cancelled orders then expected to get in front of the queue with their tiny little orders.

“In the past two years they call me and behave like my best friend,” he told a laughing crowd of TSMC partners and customers in Silicon Valley recently. One automaker called to urgently request 25 wafers, said Wei, who is used to fielding orders for 25,000 wafers. “No wonder you cannot get the support.” CC Wei, TSMC Technical Symposium 2022.

I was there, it was funny, typical CC Wei with his biting humor.

The foundries took advantage of this situation and as they say made a silk purse out of a sows ear. Would there have been a CHIPs Act without the shortage narrative? Would there have been price hikes and CAPEX splurges? I hate to say it but that false narrative was for the greater good of the semiconductor industry, absolutely.
 
Very good article and yes that is what happened. We did not have a wafer shortage problem, we had an allocation problem. Car makers cancelled orders then expected to get in front of the queue with their tiny little orders.

“In the past two years they call me and behave like my best friend,” he told a laughing crowd of TSMC partners and customers in Silicon Valley recently. One automaker called to urgently request 25 wafers, said Wei, who is used to fielding orders for 25,000 wafers. “No wonder you cannot get the support.” CC Wei, TSMC Technical Symposium 2022.

I was there, it was funny, typical CC Wei with his biting humor.

The foundries took advantage of this situation and as they say made a silk purse out of a sows ear. Would there have been a CHIPs Act without the shortage narrative? Would there have been price hikes and CAPEX splurges? I hate to say it but that false narrative was for the greater good of the semiconductor industry, absolutely.

A Toyota electronics engineer said that when they were tearing down Tesla, they were surprised how little electronics it had, and how "dumb" the overal functioning of key systems were.

Over first few generations Teslas had no active suspension elements whatsoever, comparable to economy cars.

Climate control had a single temperature sensor.

There was obviously no dashboards, and separate media system. They are notable microcontroller hogs in regular vehicles.

Lots of buttons were not sitting on buses, but were directly wired to MCUs in dashboards with long wires. So, no semblance of any kind of mesh architecture, or extensive use of CanBUS.

Window/hatch motors were dumb, with no sensors, only limit switches.

Everything outside of powertrain had surprisingly little electronics.

His take: it's an early nineties car with only difference of electric powertrain, and a smartphone for dashboard, which replaced tons of hardware with software.
 
A Toyota electronics engineer said that when they were tearing down Tesla, they were surprised how little electronics it had, and how "dumb" the overal functioning of key systems were.

Over first few generations Teslas had no active suspension elements whatsoever, comparable to economy cars.

Climate control had a single temperature sensor.

There was obviously no dashboards, and separate media system. They are notable microcontroller hogs in regular vehicles.

Lots of buttons were not sitting on buses, but were directly wired to MCUs in dashboards with long wires. So, no semblance of any kind of mesh architecture, or extensive use of CanBUS.

Window/hatch motors were dumb, with no sensors, only limit switches.

Everything outside of powertrain had surprisingly little electronics.

His take: it's an early nineties car with only difference of electric powertrain, and a smartphone for dashboard, which replaced tons of hardware with software.
This last item sounds smart to me.

For the results they get, most expensive cars, SUVs, and trucks are over-complicated, and indicative of "we've always done things with distributed microcontrollers" mentality. And they're a pain in the neck to diagnose and service. Especially German cars, which I've been addicted to for about 20 years. I'm looking forward to seeing what Apple is rumored to come up with entering the auto industry. If they do, it'll be the first car that's a computer first, and a mechanical device second. I wonder if they'll do to the auto industry what they did to the watch industry.
 
This last item sounds smart to me.

For the results they get, most expensive cars, SUVs, and trucks are over-complicated, and indicative of "we've always done things with distributed microcontrollers" mentality. And they're a pain in the neck to diagnose and service. Especially German cars, which I've been addicted to for about 20 years. I'm looking forward to seeing what Apple is rumored to come up with entering the auto industry. If they do, it'll be the first car that's a computer first, and a mechanical device second. I wonder if they'll do to the auto industry what they did to the watch industry.

I would say most of things you describe are a consequence of copy paste engineering, rather than inability to do engineering as such.

Cars became like chips with R&D costs. Almost no orginial engineering is happening today, and top brands only roll out a real new platform once in 6-10 years.

The "R&D" becoming a swear word to MBAs is the reason. Engineers are too expensive, and the professional manager people are too afraid to show their relative useless to higher ups as the most paid class of workers.

The above point sounds ridiculous, but I personally seen multiple times in my life that most stellar engineer gets ganged up on, and fired in a big co. environment. So it's 100% real for me.
 
I would say most of things you describe are a consequence of copy paste engineering, rather than inability to do engineering as such.

Cars became like chips with R&D costs. Almost no orginial engineering is happening today, and top brands only roll out a real new platform once in 6-10 years.

The "R&D" becoming a swear word to MBAs is the reason. Engineers are too expensive, and the professional manager people are too afraid to show their relative useless to higher ups as the most paid class of workers.

The above point sounds ridiculous, but I personally seen multiple times in my life that most stellar engineer gets ganged up on, and fired in a big co. environment. So it's 100% real for me.
Many managers provide a very high level of value to their superiors. It's called "sucking up" and being "trusted". There is a pattern to it. Managers surround themselves with an inner circle of people, many of whom may be competent engineers, and the entire operation from top to bottom is a hierarchy of quid pro quos. Not every engineering team works like that, but a lot of them do. And, in my experience, a lot of engineers like this structure, because they know the managers need them and they'll be rewarded for good outcomes and good behavior (in other words, making the manager successful). Some engineers are purists and can't or won't play the game. They have to hunt around for the managers/leaders who add leadership value and don't play the game. It can be a long hunt.
 
Dumb question - does Porsche actually sell a car with 8,000 pieces of separate silicon? This seems insane.

"One example: the battery-powered Porsche Taycan has over 8,000 chips. That will double or triple by the end of the decade, according to Volkswagen."

Kudos to Tesla for waking these OEMs up..
 
I don't believe it.
I was trying to think of the most liberal interpretation of this possible.. but even in the case of batteries - Porsche uses pouch cells, so they don't even have 1,000 cells with "electronics on each".. lol
 
I was trying to think of the most liberal interpretation of this possible.. but even in the case of batteries - Porsche uses pouch cells, so they don't even have 1,000 cells with "electronics on each".. lol

Not likely. I think you would be hard pressed to find any car with that many chips. Too many points of failure. I'm a long time Porsche owner and can tell you that they are the best engineered cars I have ever driven. I'm talking about the 911s not the other ones.

I looked at the Tesla Model 3 teardown. It is a very simple car. I don't remember how many chips but not many. It was definitely not a semiconductor showcase.
 
Not likely. I think you would be hard pressed to find any car with that many chips. Too many points of failure. I'm a long time Porsche owner and can tell you that they are the best engineered cars I have ever driven. I'm talking about the 911s not the other ones.

I looked at the Tesla Model 3 teardown. It is a very simple car. I don't remember how many chips but not many. It was definitely not a semiconductor showcase.
Given Porsche is "old school" (OT - to the point they recently fired their forward looking CEO), I would expect them to have 'a lot of separate silicon functions' - sensors, switches, and tons of wires like most cars from legacy OEMs, but I figured maybe 500 unique functions requiring ICs at most. (Modern gas engines are extremely complex - it's why an EV only needs 40% as many parts to operate). And yes - the 911 is very well balanced design. Now I'm curious if Porsche contracts out to less OEMs on average than other makers, including mothership VW.

Ironically, I own a Model 3 myself, and it is refreshing to see them using modern nodes for their chips. Samsung 14nm is used for the FSD v3.0 computer, and TSMC N7 is used for (Ryzen) infotainment on recent builds. Tesla FSD v4.0 is due soon and rumored to use Samsung 7nm or TSMC N5. I'd guess other automaker are using nothing newer than 28nm today, other than maybe the radio chips..
 
Ironically, I own a Model 3 myself, and it is refreshing to see them using modern nodes for their chips. Samsung 14nm is used for the FSD v3.0 computer, and TSMC N7 is used for (Ryzen) infotainment on recent builds. Tesla FSD v4.0 is due soon and rumored to use Samsung 7nm or TSMC N5. I'd guess other automaker are using nothing newer than 28nm today, other than maybe the radio chips..

TSMC automotive grade processes used to lag the industry but TSMC is really pushing the FinFET nodes. N5 is fully automotive qualified and you can expect N3 to follow. We will know more at the TSMC OIP in October.

Tesla was the first car company to frequent SemiWiki. Today we have 48 registered members from Tesla. Most other car companies are now SemiWiki regulars which suggests that they are all following Tesla into the chip business. The automotive semiconductor shortage certainly moved that along. EDA is also spending a lot of time with car companies.

My next car may be a Tesla when full autopilot is more mature. I bought my current Porsche during a business trip to Las Vegas unbeknownst to my wife. She was a little surprised when I drove home instead of flying. :ROFLMAO:

Dan Porsche Las Vega.jpg
 
TSMC automotive grade processes used to lag the industry but TSMC is really pushing the FinFET nodes. N5 is fully automotive qualified and you can expect N3 to follow. We will know more at the TSMC OIP in October.

Tesla was the first car company to frequent SemiWiki. Today we have 48 registered members from Tesla. Most other car companies are now SemiWiki regulars which suggests that they are all following Tesla into the chip business. The automotive semiconductor shortage certainly moved that along. EDA is also spending a lot of time with car companies.

My next car may be a Tesla when full autopilot is more mature. I bought my current Porsche during a business trip to Las Vegas unbeknownst to my wife. She was a little surprised when I drove home instead of flying. :ROFLMAO:

View attachment 856
Real men buy Porsche during business trips.
 
TSMC automotive grade processes used to lag the industry but TSMC is really pushing the FinFET nodes. N5 is fully automotive qualified and you can expect N3 to follow. We will know more at the TSMC OIP in October.

Tesla was the first car company to frequent SemiWiki. Today we have 48 registered members from Tesla. Most other car companies are now SemiWiki regulars which suggests that they are all following Tesla into the chip business. The automotive semiconductor shortage certainly moved that along. EDA is also spending a lot of time with car companies.

My next car may be a Tesla when full autopilot is more mature. I bought my current Porsche during a business trip to Las Vegas unbeknownst to my wife. She was a little surprised when I drove home instead of flying. :ROFLMAO:
That's a beautiful color choice - Thanks for sharing! There is very little that beats a convertible. I'm also glad to hear you're still married :).

For the record, TSMC 16nm silicon is used in the (dirty car) picture below.

1659836445051.gif
 
Back
Top