Mazda car radios getting stuck on NPR

[NiHaoMike]

https://arstechnica.com/cars/2022/02/radio-station-snafu-in-seattle-bricks-some-mazda-infotainment-systems/

[Kim Christensen]

Well, that's just crap coding at it's finest. 

[Ed.Kloonk]

And what happens if something like this happens again after the car warranty runs out?

edit. I mean much later on. When nobody at mazda remembers how this stuff works.

[MrMobodies]

From there, the infotainment systems became trapped in a rebooting loop, never successfully completing the task. When afflicted owners took their cars to be checked at local Mazda dealers, they were told that the "connectivity master unit" was dead and needed to be replaced.

The snag? A new CMU costs $1,500—if you can find one, which you can't, because of supply chain problems

Vipre77 Ars Praefectus
jump to post

Dr Gitlin wrote:
show nested quotes


The screen on which the backup camera is trying to project images is broken. So it's obviously true.

I’m not so certain…. I had a 2016 Mazda CX-3 and my wife still has a 2016 CX-5. Both of our cars had failures exhibiting identical symptoms as described in the article. In fact, it happened twice on my CX-3, but the failure the second time seemed to a little different. There is a button combination you can press to force a reboot. Doesn’t help the problem. I tried disconnecting the battery for an hour and that didn’t work either. In all cases, in my experience, the rear view camera continued to work even while the computer was constantly rebooting. I believe the display itself may have two inputs, one for the camera and one for the head unit, and it switches between them when you put the car in reverse. Either that or the camera system is somehow operating a lower level than the rest of the infotainment system and runs Independently of it.

The second failure I had on my CX-3 was odd. FM radio continued to operate at all times, even during the reboot process, but I was unable to change stations, inputs, or even the volume (mute still worked). The screen would cycle through the usual animations during the boot process. Eventually, the FM radio stopped worked, and later still, the display quit showing the reboot animations. You’d still see a little flicker of the back light to tell when it was rebooting. Even in this state the rear-view camera display kept working as intended.

Both cars were still under original factory warranty at the time and all three CMU replacements were covered.

Someone asked earlier and I’ve read on the Mazda forums that Johnson Controls developed the Infotainment system in these. If true, then odds are good that they did both the hardware and software work here. The problem can’t simply be flashed away with a firmware update. My dealer tried for several hours and the system reboots before you can even access the menu to start the flash. There’s no safe mode from which you can run the flash process, nor will it boot from a USB or anything. You can only run the updates after the system successfully boots into the OS and you can access the normal config menus. From what I can see, you’d either need to access a JTAG header like they’d use in the factory or a way to physically replace the cache memory with the offending file to have any hope. Even reflashing via JTAG might not do it if it doesn’t wipe the cache.

How can they allow radio stations and broadcasts to interrupt their system at a firmware level.

I don't remember this ever happening with car radios or consoles whatever they are called.

[AVGresponding]

And what happens if something like this happens again after the car warranty runs out?

edit. I mean much later on. When nobody at mazda remembers how this stuff works.

I doubt anyone at Mazda knows how this stuff works, it'll be designed and manufactured by some third party like Matsushita or Samsung, or as it's North American market maybe Motorola or someone?

[sleemanj]

I hate how cars have gone to all these fully integrated systems, my $20 Aliexpress MP3 playing DIN slot headunit does me just fine.

[BrokenYugo]

And what happens if something like this happens again after the car warranty runs out?

edit. I mean much later on. When nobody at mazda remembers how this stuff works.

I doubt anyone at Mazda knows how this stuff works, it'll be designed and manufactured by some third party like Matsushita or Samsung, or as it's North American market maybe Motorola or someone?

Doesn't seem to matter who or where it comes from, automotive is just sloppy like that. I'm reminded of the Toyota "unintended acceleration" fiasco, when it came out that the ECU was full of spaghetti code and single random bit flips could make it pin the throttle, among other horrors. If Toyota was screwing up that badly I'd assume their generally lower quality (IMO) competition is worse.

[MrMobodies]

Doesn't seem to matter who or where it comes from, automotive is just sloppy like that. I'm reminded of the Toyota "unintended acceleration" fiasco, when it came out that the ECU was full of spaghetti code and single random bit flips could make it pin the throttle, among other horrors. If Toyota was screwing up that badly I'd assume their generally lower quality (IMO) competition is worse.

I remember that, they blamed driver error, floor mats and I read the authorities started to take more notice and investigate when some highway officer and his family got killed in one of them.
I last read they got fined for it.

https://www.latimes.com/archives/la-xpm-2010-mar-03-la-fi-toyota-saylor3-2010mar03-story.html

MARCH 3, 2010 12 AM PT
  Off-duty CHP Officer John Saylor, his wife, daughter and brother-in-law died Aug. 28 when their 2009 Lexus ES350 took off at full throttle, reaching an estimated 120 mph in rush-hour traffic in suburban San Diego, before crashing and killing all four on board.
...
The brother of Saylor’s wife placed a chilling 911 call from the vehicle. He told an emergency dispatcher that the accelerator pedal was stuck and that Saylor was unable to slow the car down.
“We’re in trouble. . . . There’s no brakes. End freeway half-mile,” Christopher Lastrella said during the recorded conversation.

[Haenk]

The brother of Saylor’s wife placed a chilling 911 call from the vehicle. He told an emergency dispatcher that the accelerator pedal was stuck and that Saylor was unable to slow the car down.
“We’re in trouble. . . . There’s no brakes. End freeway half-mile,” Christopher Lastrella said during the recorded conversation.

"pedal stuck" is not a software thing, still that should never ever be possible

[Cerebus]

The brother of Saylor’s wife placed a chilling 911 call from the vehicle. He told an emergency dispatcher that the accelerator pedal was stuck and that Saylor was unable to slow the car down.
“We’re in trouble. . . . There’s no brakes. End freeway half-mile,” Christopher Lastrella said during the recorded conversation.

"pedal stuck" is not a software thing, still that should never ever be possible

I think the only people who have the sang-froid to be precise and say "The accelerator system has become unresponsive to control inputs" when in a vehicle heading at top speed down a road that ends in half a mile also normally have Martin-Baker ejector seats to get them out of trouble and a brand new necktie in their near future. For the rest of us I suspect screaming "The ******** pedal's stuck!" is the nearest we're going to get to terminological exactitude under those circumstances. So there's no reason to believe that the description does not fit a "software induced failure in the acceleration system" any less well than it describes a simple mechanical failure.

[AVGresponding]

If you've time to make a phone call (even as a passenger), I'm surprised there wasn't time to try things like force it into a lower gear, or reverse (you might break the gearbox, but at least you're no longer accelerating), or even just turn the engine off (without completely removing the keys obviously, don't know if that model is keyless though).

[SeanB]

Lexus has been keyless for a while, top end model range almost always are these days. Probably also locked out shifting down a gear to protect the gearbox, and thus the shift does nothing, as it is also just a set of logic inputs to the drivetrain controller, with no actual linkages any more. Might have reset if you held in start stop button for 30 seconds, but kind of hard to read the manual while being in a semi guided missile.

[voltsandjolts]

Click it into neutral?
I guess panic had set in.

[WatchfulEye]

Certainly at the time Toyota was under fire for unintended acceleration, there were significant usability issues with the keyless system.

I had a lexus at the time and given the fuss, I thought it would be sensible to check functionality so that I had a plan. So I found an empty road and drove at 10mph while I tested mitigation strategies.

Pressing the start/stop button was my first thought. This did nothing at all.
Repeatedly mashing the button also did nothing.
I had to hold the button for what was a very disconcerting period of time, even at 10mph, although it was probably only 5 seconds.
Pushing the transmission lever into neutral worked perfectly and this was the obvious best choice out of the solutions tested.

It was fortunate that I had done this, as about 1 year later the car suddenly accelerated uncontrollably as I pulled out of a junction. I popped it onto neutral and coasted to a stop with the engine bouncing off the limiter.

The cause in this case was easily identified. The pedal had become Mechanically caught by the floor mat, when I had floored it to make a quick getaway from the junction.

[SilverSolder]

Certainly at the time Toyota was under fire for unintended acceleration, there were significant usability issues with the keyless system.

I had a lexus at the time and given the fuss, I thought it would be sensible to check functionality so that I had a plan. So I found an empty road and drove at 10mph while I tested mitigation strategies.

Pressing the start/stop button was my first thought. This did nothing at all.
Repeatedly mashing the button also did nothing.
I had to hold the button for what was a very disconcerting period of time, even at 10mph, although it was probably only 5 seconds.
Pushing the transmission lever into neutral worked perfectly and this was the obvious best choice out of the solutions tested.

It was fortunate that I had done this, as about 1 year later the car suddenly accelerated uncontrollably as I pulled out of a junction. I popped it onto neutral and coasted to a stop with the engine bouncing off the limiter.

The cause in this case was easily identified. The pedal had become Mechanically caught by the floor mat, when I had floored it to make a quick getaway from the junction.

The floor mat in some cars (I've seen it in Fords) has a small hole that a hook engages (from the floor), to prevent the mat sliding forward.  Should really be mandatory...

[AVGresponding]

Lexus has been keyless for a while, top end model range almost always are these days. Probably also locked out shifting down a gear to protect the gearbox, and thus the shift does nothing, as it is also just a set of logic inputs to the drivetrain controller, with no actual linkages any more. Might have reset if you held in start stop button for 30 seconds, but kind of hard to read the manual while being in a semi guided missile.

It's quite troubling to realise that, especially with the ubiquity of electronic controls over physical ones nowadays, manufacturers seem to be happy to ignore the concept of fail+safe.

[BrokenYugo]

If you've time to make a phone call (even as a passenger), I'm surprised there wasn't time to try things like force it into a lower gear, or reverse (you might break the gearbox, but at least you're no longer accelerating), or even just turn the engine off (without completely removing the keys obviously, don't know if that model is keyless though).

For the record, the correct action is to stand on the brake pedal until the car stops. Assuming the brakes aren't half functioning worn out junk, this will stop the car.

[Cerebus]

For the record, the correct action is to stand on the brake pedal until the car stops brakes overheat, fade and catch fire.

TFTFY

[SilverSolder]

It matters what kind of car the throttle sticks on...  if it is a Toyota Corolla, the brakes can probably win over the engine.   If it happens in your supercharged hot-rod, you may have a problem!

[Monkeh]

Brakes are sized according to vehicle mass more than anything else. They are not, as I understand them, intended to stop a vehicle at wide open throttle, especially if you already have significant momentum.

If you find yourself on an empty stretch of road suitable for trying it, please, feel free. Myself, I value my brake components and would prefer not to ruin them testing this, when I can simply take the vehicle out of gear.

The cause in this case was easily identified. The pedal had become Mechanically caught by the floor mat, when I had floored it to make a quick getaway from the junction.

A classic case of 'driving like a twat'..

[Cerebus]

The cause in this case was easily identified. The pedal had become Mechanically caught by the floor mat, when I had floored it to make a quick getaway from the junction.

A classic case of 'driving like a twat'..

As my mother would say: "Takes one to know one!".

[Someone]

It matters what kind of car the throttle sticks on...  if it is a Toyota Corolla, the brakes can probably win over the engine.   If it happens in your supercharged hot-rod, you may have a problem!

Brakes are going to overcome the engine of pretty much any road going (combustion engine) car. Stalling a runaway diesel in 1st gear? no problem, and thats close to corner case worst possible situation.

[Cerebus]

It matters what kind of car the throttle sticks on...  if it is a Toyota Corolla, the brakes can probably win over the engine.   If it happens in your supercharged hot-rod, you may have a problem!

Brakes are going to overcome the engine of pretty much any road going (combustion engine) car. Stalling a runaway diesel in 1st gear? no problem, and thats close to corner case worst possible situation.

Try it. You're in Aus, and if there is anywhere in the world you can find a long, straight, unoccupied road it's in Aus. You might be surprised. Full throttle remember, not idle.

Alternatively you can work it out mathematically. My car produces 174 Nm at the engine shaft, 559 Nm at the rear shaft in 5th gear, 2314 Nm in 1st. That's a lot of torque to overcome. The brake force calculations versus engine torque are left as an exercise for the student but I've just done them on the back of an envelope and it's surprising. The engine wins.

Also the 143 BHP (107 kW) my car's engine produces is a lot of heat to get rid of from the brake discs. A brake disc is cast iron and weighs 4.5kg or thereabouts each, they have a specific heat capacity of 460 j/kgK and a melting point of around 1150 C, so need 9.35 MJ to raise the temperature of 4 x 4.5 kg = 18 kg discs to melting point from 20 C, which is just 87 seconds at 107kW ignoring any cooling (which wouldn't be very much). Anyone who has done any racing will tell you how quickly brakes fade once they get too hot.

[jmelson]

If you've time to make a phone call (even as a passenger), I'm surprised there wasn't time to try things like force it into a lower gear, or reverse (you might break the gearbox, but at least you're no longer accelerating), or even just turn the engine off (without completely removing the keys obviously, don't know if that model is keyless though).

The Toyota Prius does not have a traditional transmission, it has a fixed planetary gear set between 2 motors and the IC engine.  Reverse is accomplished by running the electric motors backward and the IC engine stopped.
Their drivetrain control had something like 46 tasks, and the LOWEST priority task was monitoring the BRAKE pedal!  YIKES!  When overloaded, the real time manager would drop lower priority tasks.  Supposedly, holding the "power" button for 4 seconds provides a hardware shutdown of all electronics that can't be overridden by software, and would have caused the car to coast to a stop.  The gear shift on the dash is obviously just a set of switches that are read by the drivetrain computer.  The Prius has no "ignition key/ignition switch" it just has a "power" button.  The key fob has to be somewhere near the dash for the immobilizer to detect it.
Jon

[Someone]

It matters what kind of car the throttle sticks on...  if it is a Toyota Corolla, the brakes can probably win over the engine.   If it happens in your supercharged hot-rod, you may have a problem!

Brakes are going to overcome the engine of pretty much any road going (combustion engine) car. Stalling a runaway diesel in 1st gear? no problem, and thats close to corner case worst possible situation.

Try it. You're in Aus, and if there is anywhere in the world you can find a long, straight, unoccupied road it's in Aus. You might be surprised. Full throttle remember, not idle.

Alternatively you can work it out mathematically. My car produces 174 Nm at the engine shaft, 559 Nm at the rear shaft in 5th gear, 2314 Nm in 1st. That's a lot of torque to overcome. The brake force calculations versus engine torque are left as an exercise for the student but I've just done them on the back of an envelope and it's surprising. The engine wins.

Also the 143 BHP (107 kW) my car's engine produces is a lot of heat to get rid of from the brake discs. A brake disc is cast iron and weighs 4.5kg or thereabouts each, they have a specific heat capacity of 460 j/kgK and a melting point of around 1150 C, so need 9.35 MJ to raise the temperature of 4 x 4.5 kg = 18 kg discs to melting point from 20 C, which is just 87 seconds at 107kW ignoring any cooling (which wouldn't be very much). Anyone who has done any racing will tell you how quickly brakes fade once they get too hot.

I said runaway, not idle. So something in the order of 4000Nm at the wheels, in first gear, engine stopped no problem. You jump to maths to try and justify your imagined truth but there are some basics which immediately point to it being wrong:
Combustion engine cars decelerate faster than they can accelerate, even if you take a car that can freely spin the wheels under power, when stopping hard ABS pulls down peak stopping force on the brakes to maintain traction.
Cars dont have brake fade from a single pull to stop from cruising speed, its something thats actually physically tested for Australian design rules compliance (15 back to back partial declarations).

So these things have been tested, it might be that the brakes on my vehicles are wildly high performance but being stock from the factory I dont think they are anything special (they do have highly non-linear servo assist for emergency force, but that should be fairly commonplace, again its something tested by the ADR compliance). Depending on region in Australia you are looking at 0.4 to 0.6g as the minimum braking force to meet roadworthy testing, which is more than enough to arrest regular engines, [sa]oddly enough high power vehicles have higher performing braking systems...[/sa]