Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why Carmakers Can’t Transition to Newer Chips

jalopnik.com

Post reply on HN
Print view View original

221–230 of 335 posts

Re: Why Carmakers Can’t Transition to Newer Chips

#221

Post by kortex »

baybal2 wrote: post #30
TaylorPhebillo wrote: post #27

Earlier quoted context omitted.

I think the Cortex-R chips are more targeted here- I think the V8-R chips are designed specifically for automotive use, so I assume they have appropriate ratings.

There are very lazy car companies which simply specify "automotive certified" car chips for everything, including MCUs running window switches. So, don't be surprised $20 ARM-R based MCUs being used to do something trivial like driving window motor. I seen a few dashboard where there is an STM32 sitting connected to a single button, whose only duty is to register a key press, and burp something on the CanBus in respo…

A) I doubt you need a $20 component to register a key press and do some canbus IO

B) If you did have a $20 microcontroller for Canbus control of a window motor driver but it saved $19 in extra wiring/labor (driver controls usually route to all windows), provides reliable debounce, automated one-push-to-open, and allows things like holding keyless lock to close all windows, it's totally worth it.

Re: Why Carmakers Can’t Transition to Newer Chips

#222

Post by perl4ever »

optimiz3 wrote: post #150
csours wrote: post #121

Earlier quoted context omitted.

Small company risk is failure to grow. Big company risk is every other failure. Tesla is valued as a growth/tech company. As long as people believe in their growth, they get more cash. I don't personally understand how Tesla hasn't had more problems with their apparently uncontrolled engineering changes. At least part of it is customer enthusiasm for the product papering over any drawbacks. It's not that GM can't mak…

Tesla is past the cash raising phase though - they are massively profitable and actively retiring debt early. https://www.sec.gov/ix?doc=/Archives/edgar/data/1318605/0000... "On July 16, 2021, we issued a notice of redemption to the holders of the 2025 Notes informing the holders that we will redeem the notes in full in August 2021 at a redemption price equal to 102.65% of outstanding principal amount, plus accrued a…

>Tesla is past the cash raising phase though

It appears to me that they are steadily issuing stock at the rate of about 20% of their market cap per year, or roughly at the rate of $10 billion per month.

It looks like the number of (diluted) shares outstanding increased by over 20% between 2019 and 2020.

What did that consist of? Their annual report says mainly:

  - Issuance of common stock for equity incentive awards
  - Issuance of common stock in public offerings
e.g. "On February 19, 2020, we completed a public offering of our common stock and issued a total of 15.2 million shares (as adjusted to give effect to the Stock Split, as described in the paragraph below), for total cash proceeds of $2.31 billion, net of underwriting discounts and offering costs of $28 million."

"On September 1, 2020, we entered into an Equity Distribution Agreement with certain sales agents to sell $5.00 billion in shares of our common stock from time to time through an “at-the-market” offering program. Such sales were completed by September 4, 2020 and settled by September 9, 2020, with the sale of 11,141,562 shares of common stock resulting in gross proceeds of $5.00 billion and net proceeds of $4.97 billion, net of sales agents’ commissions of $25 million and other offering costs of $1 million."

"On December 8, 2020, we entered into a separate Equity Distribution Agreement with certain sales agents to sell $5.00 billion in shares of our common stock from time to time through an “at-the-market” offering program. Such sales were completed by December 9, 2020 and settled by December 11, 2020, with the sale of 7,915,589 shares of common stock resulting in gross proceeds of $5.00 billion and net proceeds of $4.99 billion, net of sales agents’ commissions of $13 million and other offering costs of $1 million."

Also, as of their 2020 annual report, roughly a billion shares were authorized to issue, which is on the order of another 100%, or double the current outstanding.

https://www.sec.gov/ix?doc=/Archives/edgar/data/1318605/0001...

You might dismiss this as "way back in 2020", but it does seem to be their latest annual report.

Their latest quarterly report is as of June 30, 2021, and guess what? Shares outstanding are up about 4.6%. Compounded over a year, that's going to be just about 20% again.

10-Q: https://www.sec.gov/ix?doc=/Archives/edgar/data/131860

"Stock-based awards" seem to have been 93 million in the last three months. TSLA is around $775/share.

Re: Why Carmakers Can’t Transition to Newer Chips

#223

Post by winrid »

mcguire wrote: post #214
formerly_proven wrote: post #175

Earlier quoted context omitted.

Siblings covered why 12 V is annoying, let's talk about why 12 V was chosen in the first place. 6 and 12 V electric systems in vehicles came about because lead-acid batteries made sense in this application, and cetpar. a higher voltage lead-acid battery is overall costlier to manufacture, because it contains more individual cells. Another big thing is that there are many switches and relays in a car, and those often…

I seem to recall, several years ago (when stopping the engine at traffic lights and restarting as soon as the driver released the brake), at least one company was exploring moving to 24V. However, the effort failed precisely because the lifetime of 24V switches was significantly shorter. Naturally, I can't find any links to the info...

FWIW Almost no switches in a modern car (turn signals, wipers, door switches) carry a significant amount of current as they are controlled via confuzers.

Just built a car without a single relay. I used PMUs (basically boxes of mosfets and current sensors AFAIK).

Re: Why Carmakers Can’t Transition to Newer Chips

#224

Post by sbierwagen »

baybal2 wrote: post #30
TaylorPhebillo wrote: post #27

Earlier quoted context omitted.

I think the Cortex-R chips are more targeted here- I think the V8-R chips are designed specifically for automotive use, so I assume they have appropriate ratings.

There are very lazy car companies which simply specify "automotive certified" car chips for everything, including MCUs running window switches. So, don't be surprised $20 ARM-R based MCUs being used to do something trivial like driving window motor. I seen a few dashboard where there is an STM32 sitting connected to a single button, whose only duty is to register a key press, and burp something on the CanBus in respo…

A big chunk of this crisis was caused by car manufacturers being insanely focused on cost cutting. Doing obviously dumb things with BOM isn't exactly a calling card of the automotive industry.

Re: Why Carmakers Can’t Transition to Newer Chips

#225

Post by GeorgeTirebiter »

schiffern wrote: post #7

> “We were able to substitute alternative chips, and then write the firmware in a matter of weeks,” Musk said. “It’s not just a matter of swapping out a chip; you also have to rewrite the software.” https://www.theverge.com/2021/7/26/22595060/tesla-chip-short...

So much speculation here. Let me set the record straight(er). (I worked in Tesla FW, left 5 years ago.) Tesla does use industry best practices. All code must pass MISRA checkers, code is modular, and reused when possible. We used all sorts of chips (that were auto qual) of many different architectures. We had extensive tools for e.g. stack monitoring, diagnostic readouts, and much more. If you move from, say, one ST ARM chip to another one in a similar family, the peripherals may be a little different, but generally work the same way (with maybe a few more or fewer features). So reworking the I/O drivers is work, but given the excellent layering of Tesla "body control" FW, it's really quite straightforward to pick a different processor from a given family. It's for sure true that if you went from ST ARM to Freescale/NXP ARM the peripherals are diff, so yes, more time would be needed to write the I/O drivers. But the effort ongoing when I left was exactly to make the appropriate SW generalizations so it would be possible to do exactly this.

Re: Why Carmakers Can’t Transition to Newer Chips

#226

Post by perl4ever »

formerly_proven wrote: post #220
perl4ever wrote: post #215

Earlier quoted context omitted.

>Cars can generally stop safely. "Brake-by-wire is used in all common hybrid and electric vehicles produced since 1998 including all Toyota, Ford, and General Motors Electric and hybrid models." "Three main types of redundancy usually exist in a brake-by-wire system: Redundant sensors in safety critical components such as the brake pedal. Redundant copies of some signals that are of particular safety importance such…

Generally speaking those are not purely brake by wire. The master cylinder is still mechanically connected to the pedal. However, in normal conditions the brake actuation will be performed over wire.

This is a question of semantics, but I don't understand your usage of "purely".

How can it not be "purely" brake by wire, if, when depressing the brake pedal, the friction brakes are not always triggered?

If electronics decide not to apply hydraulic force when everything is going fine, that there must be a potential failure mode where they ignore the pedal inappropriately.

Can you explain further?

Re: Why Carmakers Can’t Transition to Newer Chips

#227

Post by MichaelZuo »

hinkley wrote: post #218
csours wrote: post #51

Disclaimer up front: I work for GM. I don't work on chips or components for my day job. What follows is solely my own opinion. Some things to emphasize: OEMs (GM, Ford, Toyota, VW, etc) do not design components, and they do not want to. They design specifications for components, and then get suppliers to bid. This is great for efficiency in established ecosystems, not great for agility. To my knowledge, GM did not ca…

The suppliers are running in pretty thin margins aren’t they? Did they really have any choice but to cancel those orders or potentially go bankrupt? If you sit on someone’s shoulders and forget that they’re doing most of the work, you can be shocked all you want when they collapse out from under you but nobody on the ground is going to have any sympathy at all for you. I’ve worked for software subcontractors and it’s…

‘ sweetheart deal loaded with scope creep not spelled out properly in the contracts’ That sounds more like a problem to take up with your lawyers?

Re: Why Carmakers Can’t Transition to Newer Chips

#228

Post by baybal2 »

HeyLaughingBoy wrote: post #80
baybal2 wrote: post #34

Earlier quoted context omitted.

It is very troubling development when doing something so trivial even requires a computer. That's a matter readily solved by a PID controller made of a few dozen 74XXX parts. Possibly made n-times redundant. Even a more fancy LP solver probably wouldn't need a fully fledged computer.

Let's start by assuming that the engineers are at least basically competent, shall we? I mean, I could implement a PID controller without any 74xxx logic at all: a basic op-amp, a handful of resistors and capacitors and boom I'm done! But there's a reason that engineering doesn't do that these days and it's not that engineers don't know how. My PID control built from an opamp? In 2021 that's almost always a stupid id…

> The reality is that in 2021, often the simplest, most robust and cost-effective way to do something trivial is by using a computer to do it.

Very much not. Complicated MCUs are a very risky single vendor supply, and as we see now, people are now paying for an engineering overkill.

74xxx are available in every shape, and form, from dozens of suppliers.

Re: Why Carmakers Can’t Transition to Newer Chips

#229

Post by tuatoru »

perl4ever wrote: post #216
formerly_proven wrote: post #175

Earlier quoted context omitted.

Siblings covered why 12 V is annoying, let's talk about why 12 V was chosen in the first place. 6 and 12 V electric systems in vehicles came about because lead-acid batteries made sense in this application, and cetpar. a higher voltage lead-acid battery is overall costlier to manufacture, because it contains more individual cells. Another big thing is that there are many switches and relays in a car, and those often…

>All of those switches become much more expensive Wouldn't a more important question be how does the failure mode change ?

Above 30V - 50V DC, contact arcing becomes a huge problem. (It's a problem below that, but less huge.)

With AC, the arc will self-extinguish a half-cycle after the switch opens. With DC there is no cycle, and contacts can be completely vaporised in tens of milliseconds.

Commodity switching components are usually rated "30V DC, 250V AC" for this reason.

It is possible to design switches so that even DC arcs self-extinguish, but the result is expensive and not as reliable as one could wish.

If cars had been invented after 1980, they would probably use solid-state switches except for the very high current circuits such as the starter solenoid and headlights. (Transistors were around a lot earlier than that, but engineers are sensibly cautious about new technologies.)

Edit: To answer your question directly: no. Cost is prime.

Re: Why Carmakers Can’t Transition to Newer Chips

#230

Post by formerly_proven »

idiotsecant wrote: post #190
buescher wrote: post #182

Earlier quoted context omitted.

One of the big issues with 40-60V systems, which have been right around the corner for 20 years now, maybe more, has been that typical relay contacts arc over at about 28V DC. Yes, there are ways to adress this. No, none of them are as practical as relays yet.

This isn't really a fundamental problem. It just means using relays with contacts rated for higher voltages. It's probably the easiest part of the bom on a typical vehicle to move to a higher voltage range. I use 125VDC contact/coil rated relays all the time.

It is a fundamental problem (it's physics). The current rating for relays drops off sharply as the DC voltage increases; you might see 125 V DC rating, but the current will be a fraction of the nominal current at the rated AC voltage. If you're using an relay rated for up to 250 VAC/DC, then the breaking capacity with DC might be as low as 1 % compared to AC. Bad relay datasheets don't mention this at all, or maybe only for 30 V. Good datasheets have plots of capacity vs. voltage and current vs. life (rated capacity of relays is generally for around 100k cycles).
Post reply on HN