Volvo recalls 59,000 cars over software fault | Hacker News Reader

Volvo recalls 59,000 cars over software fault

bbc.com

64 pointstemp65 comments

Why don't any of the luxury car brands (other than Lexus) focus on reliability? It seems so strange that the market is willing to pay ~$40-$80k for a car totally indifferent to its engineering quality. If I were to own a luxury car, I'd want to know its engineering is rock-solid first and foremost. I'd want to be proud of it as a machine not just a symbol. Maybe that's just a techie thing, though.

Volvo has been notorious for extremely difficult-to-diagnose and expensive-to-repair electrical system problems for at least a decade.

I've always thought, if I was rich, I would hire one or more developers who's job was to educate me on software (empowering me), and write and run software I was interested in. Wonder why I haven't heard of any rich people doing that?

Stephen Wolfram does that.

Your comment reminds me of the question: The Greeks invented the steam engine, why didn't they have an industrial revolution?

One answer is that slave labor made a steam engine unneeded: any work the engine could do was already being done by slaves. Similarly, Cicero had no need for a printing press, any time he wanted to print something (election pamphlets, for example), he had an army of slaves copy it over and over by hand.

So maybe rich people have the same thing: they don't need someone to code them a train schedule app (or whatever to empower them), they just tell someone to figure out their schedule for them.

I think the simpler answer is they didn't invent a practical steam engine.

Thx. great idea.

Where is Volvo considered a luxury car brand? Here in the UK I don't know anyone who would call it that.

In North America Volvo has definitely evolved in to a luxury car brand, if just by the prices of their cars.

AFAIK the cheapest American Volvo (S60) is about the same price as a BMW 3-Series.

In many countries around the world it is definitely a luxury brand.

Toyota had it's fair share of bad software design issues (Lexus is a division of Toyota)...

> I'd want to be proud of it as a machine not just a symbol.

As a machine and as a symbol aren't necessarily independent of each other. Being less reliable and more expensive to fix can make it more prestigious in some cases:

https://en.wikipedia.org/wiki/The_Theory_of_the_Leisure_Clas...

Putting this kind of safety-related issue aside, the first purchaser probably won't keep the car long enough to have any real problems (at this price level it's probably a lease). Even if they do, the dealer will give them a loaner (and may even pick up the car) so problems aren't much of an inconvenience.

It's much more of an issue for those of us who like to buy several year old luxury cars -- they have become much harder and more expensive to repair in the last 10 years or so.

Can you explain "it's usually a lease?" Doesn't it defeat the purpose of signaling wealth if you don't actually own it? How common actually is leasing at this price level?

How can you tell who leased and who bought his vehicle as you cruise down the freeway?

It's simply that no one can tell whether or not you own the car, just by seeing you drive by in it, or seeing it in your driveway...

The true "equity" in a great number of cars on the road--those financed, or leased, is held by a bank, institution or business....

Remember the old false-fronted buildings in the frontier towns portrayed in numerous westerns...?

The wherewithal of those business was likely much smaller than the storefront indicated...

In other words, some consumers will go to great lengths to project "success" when in actuality they haven't quite "made it"...leasing, even renting, a luxury car for that reason is quite common around the world...

Not sure where they got their data from, but see https://www.cartelligent.com/blog/which-car-brands-do-people... -- claims 67% leasing for MB, and 70% for BMW.

It's common because people can't really afford them :-), and don't want to finance the full amount if they're not going to keep the vehicle past 3 years anyway.

I don't follow the line of reasoning that leasing is cheaper than buying outright. Via a lease you still pay interest on the full value of the asset, plus a negotiated cost having a relationship to depreciation. If leases were bad deals for the auto manufacturers they would probably stop writing them.

One nice thing that a lease offers is a defined cost. If your luxury sedan turns out to be a basket case, the warranty covers the problems during the lease, and you can simply wash your hands of it at the end of the term. Trying to sell a depreciated-to-$40K car with problems is not easy.

> If leases were bad deals for the auto manufacturers they would probably stop writing them.

Like 0% financing, it's possible that they're irrational from the perspective of a financial institution but overall good for the manufacturer because more cars get sold.

Certainly agree that leasing won't be cheaper in the long run in most cases. But in the short run it will get you a lower monthly payment, and that's all many care about.

Also agree that leasing helps fix the cost. Besides reliability issues, you win if depreciation is higher then was estimated up front (the one time I leased, the car ended up worth several thousand less then the final balloon payment so I simply turned it back in).

Leasing is cheaper if you have to pay VAT (10% of the sale price in Washington) - you only pay it your lease payments, as opposed to the full price of the car.

If you want to drive a new car every three years, and you buy & resell, VAT will cost you a lot more then a lease.

Apparently there is a huge market for lending things to rich people. Art, fine wine, cars, houses. Being rich gives you vast access to credit.

Many luxury car shoppers are not that concerned with reliability since they plan on only owning the car for a few years. 3 year leases are common and even those who buy a vehicle outright may trade in long before reliability becomes a real concern.

> Why don't any of the luxury car brands (other than Lexus) focus on reliability

The Mercedes-Benz/Puch G-Class hasn't changed that much since 1979. It's a mid-size four-wheel drive luxury SUV, a real cross country vehicle. Deer hunters, various military, VIPs, presidents and the pope, many use the G-Class.

https://en.wikipedia.org/wiki/Mercedes-Benz_G-Class

I went and found a few images of this car and variations of it.

http://blog.caranddriver.com/wp-content/uploads/2014/10/2015...

http://www.benzworld.org/forums/attachments/g-class/209746d1...

http://olive-drab.com/images/id_mb_gclass_03_700.jpg

http://i.kinja-img.com/gawker-media/image/upload/s--qAO8Ma4i...

http://spintires-mods.lt/wp-content/uploads/2014/06/Mercedes... (CGI)

http://4.bp.blogspot.com/_DobFdTEMh3Y/TDoii7WEbZI/AAAAAAAABB...

http://cars-pics-db.com/pics/steyr-puch/steyr-puch-g-290/ste...

http://s1.cdn.autoevolution.com/images/testdrive/gallery/108...

http://car-from-uk.com/ebay/carphotos/full/ebay1416702369969...

https://s-media-cache-ak0.pinimg.com/236x/4a/bc/16/4abc163c1...

http://www.papercraftsquare.com/wp-content/uploads/2015/12/D... (Paper Model)

https://s-media-cache-ak0.pinimg.com/564x/3d/84/04/3d84040cc...

http://i.kinja-img.com/gawker-media/image/upload/s---M1k3rjO...

I have a theory that software is so hard, that internet startups win simply because they can actually do software, and thereby obtain its benefits.

It's not that the start ups have such disruptive business models; but simply that software is great, and incumbents can't do it.

Web startups are good at software because that's what they focus on, and are selected for. So we're in this strange situation where getting software to do it is more important than doing it. It's the flexibility.

I'm afraid even my anecdatum is scanty: taxi calling is terrible. Uber could only be better.

>> ">> The glitch had been reported by drivers of new Volvos who said the engine could cut out without warning, creating a brief absence of steering and braking."

Do they just mean the power steering cuts out and brakes lose hydraulics? Because there's a pretty huge difference between that and 'a brief absence of steering and braking'.

Still, losing those functions is an issue. When it rains, my old Town & Country's power steering stops working and turns become difficult. I would hate to be driving and suddenly have it cut out on me, say, when I'm making a turn, or even just switching a few lanes.

My power steering went out once. The car was drivable, but just barely. Likewise if you've ever pumped the brakes while the engine is off, you know that braking power drops to almost nothing.

If that happened suddenly on the road 'loss of steering and brakes' is exactly how I would describe it. Sure it might not be completely accurate, but it is definitely a dangerous situation whatever you call it.

Eh, my car has a finicky engine and I've learned that it's perfectly drivable without power steering or brakes. I've lost power while taking tight turns, and while it was a bit surprising the first time or two I had no trouble keeping it under control. Unless the driver has significantly less strength than the average person, or is unwilling to drop whatever might be in their off hand to get it too the wheel, I'd say it's more scary than dangerous.

That being said, I certainly wouldn't be happy if my new volvo did that; they should definitely be held to higher standards than my $500 beater.

Newer cars generally have a much lower steering ratio (forgot the actual term for it) so steering without the power assist is a lot more difficult since turning the wheel fully in one direction might be only 1 full rotation of the wheel on a newer car vs 3-4 on an older one. Newer cars are also generally heavier making it even more difficult to steer.

How new are you defining newer to be? My car is a '96 mazda 626. I have driven around a '68 volvo, with no power steering and that definitely was much easier than my car when the engine dies. Has typical steering setup changed that much since '96? (Legitimately asking the question here, not trying to be rhetorical) Cars have definitely gotten heavier as you say, and volvos have always been on the heavier side.

A 96 626 is on the older side as far as steering goes, and from what I recall driving that generation 626 it has a lower/looser steering ratio than most cars these days, though I can't seem to find any numbers with a quick search.

As far as how much steering systems have changed, many are still more or less the same variable ratio rack and pinion systems, but most have moved to electronic steering assist instead of hydraulic, which allows for a higher steering ratio since you can adjust the strength of the assist on the fly (to have more control/less of an assist at higher speeds and with the wheel centered but a stronger assist when slower and turning the wheel) so low ratios aren't needed to keep the wheel easily turnable at low speeds. On top of that some cars are going full on steer by wire now, though not many due to the understandable reliability issues.

Critical firmware must be open source and public by law. Especially as it continues to take over more control of things like cars.

With cars essentially becoming computers that hold your life in there hands, its going to be interesting seeing how software development strategies evolve to build software rapidly but without the fall back of knowing you can simply send out a large firmware update to fix an issue. Cars can't afford to be stuck parked because someone made a small coding issue that got pushed before it was caught.

https://boingboing.net/2012/08/23/civilwar.html

And so far car manufacturers have done a piss poor job.

Hell, they are even insisting on using dbus for IPC as they replace QNX with Linux.

The space industry has been doing this for decades. NASA published 'The Power of Ten' with some guidelines.

http://pixelscommander.com/wp-content/uploads/2014/12/P10.pd...

This was published _20_ years ago. http://www.fastcompany.com/28121/they-write-right-stuff

Jeep had a very similar recall. I have a 2007 Jeep Wrangler, the first year of the major redesign. This was the first Wrangler with modern electronics, including stability control. In the first year, there were three recalls for software updates.

The worst one was a tendency for the system to crash while going uphill at slow speed in hot weather. The crash caused the engine to quit and restart, the dashboard would blank, light up for self-test, and reset, the transmission would downshift, and the CD player would stop. However, everything did recover within one second. That sounds very similar to Volvo's problem.

I'd double check the emissions readings before and after this "bug fix".

I've rented a Volvo V70 for the weekend that sounds like it's just the model that they're recalling. So far, no issues, and given the ice patches on the road today I'd hope it stays that way. It's a pretty cool car otherwise, lots of nifty features to make driving easier and safer, like adaptive cruise control, lane warnings, dead angle warnings, automatic break (at low speeds) when detecting obstacles, like a car suddenly stopping ahead. If this thing would steer by itself I wouldn't have to do anything – if this is the way of the future then sign me up!

Update: I'm still alive, and the car returned. Never had an issue, so I guess the model wasn't affected or its not easily reproduced in the kind of casual cross country driving I did.

Regardless – great car, 10/10 would rent again.

We had a similar issue about half a year ago while driving 2016 XC60 D4 (i.e. 4 cylinders) on a highway. I guess they'll have more cars to recall.

The D4 actually have 5 cylinders. All the DX motors on the XC60 has, the difference is in the cc.

I'm sure the dealers are excited as hell about this (and non-software-updatable-over-the-air cars in general)

At least one of the configurations of the 2012 Ford Focus/Fiesta transmission had issues. The mechanic at the dealer agreed it was a problem. Ford never acknowledged their fault as far as I know. A firmware update about a year later seemingly fixed the issue. Why wouldn't they acknowledge the issue and tell people they're working on it? Liability issue?

>Ford never acknowledged their fault as far as I know.

Yes they did, they sent letters out and and they extended the warranty on the transmission. Twice.

http://www.mcall.com/news/local/watchdog/mc-ford-fiesta-tran...

"Ford has acknowledged the potential for transmission problems in letters to owners and dealers. It has extended the warranty but has not recalled the vehicle."

I'm sure they are -- it's a nice little profit center for them (without even requiring any inventory).

Is there any coding standards like in aviation (DO-178) for cars?

Yes. Check out MISRA: https://en.wikipedia.org/wiki/Motor_Industry_Software_Reliab...

I don't know too much about vehicle side of things but my understanding is the following:

MISRA is not on the same level as DO-178. DO-178 covers a lot of ground, from testing, documentation, requirements, that sort of thing, and it's made by a government agency which enforces it during certification, where every single hardware and software component in the aircraft related to flight operations is certified individually. So you will actually have engineers from the FAA sit down and look at code coverage, how requirements are written, how they map to tests, see if the tests cover a lot of cases in your requirements, see if the code implements the requirements, that sort of thing.

MISRA on the other hand are treated like guidelines that car manufacturers generally want you to meet when you deliver them a vehicle component.

There is the Federal Motor Vehicle Safety Standards you have to meet on a vehicle as a whole, but I do not know if they have to prove that they meet the following during a formal certification process. http://www.nhtsa.gov/cars/rules/import/FMVSS/

So as far as I know, there's big approval process at the end for vehicles, while there's a formal certification process for all components in an aircraft, as well as a big approval process at the end for aircraft.

Someone please correct and fill in the blanks for the vehicle side of things.

The scope is not the same as DO-178 but the nearest would be the Function Safety norm ISO 26262.

Yes, MISRA, but it doesn't help if it isn't followed which seemed to be what was discovered when the Toyota code was examined.

And as there is no way to see the source or replicate the build process no one can easily tell if the guidelines have been followed.