Camera recall …any news [update: recall issued]

[Scott Asheville]

Trust me on this. As a former embedded software engineer. It's all about testing. "Joe-doesn't-know-squat-about-developing-software-but-has-an-opinion" describes 99.999% of members on MTC. Just like I know squat about a gazillion topics but I have an opinion.

It's all about testing it to death. That's because you have a collection of controllers, each with multiple versions of third party software and hardware, out in the field. And once you go off the map, "there be dragons". After a vehicle has been in the field a few years, it can become an "n-factorial" kind of problem to fully test it. In other words, impossible.

Really, it's an unsolvable task across a wide variety of configurations and situations (expensive and time consuming). You cannot deterministically test and validate the typical legacy auto hardware-software build the way you would a critical avionics system (unless you want your Maverick to cost $20 million per unit). The QA engineers - and yes they are engineers specializing in QA, will bust their butts on this. Then they'll give a thumbs up, cross their fingers, say a few Hail Mary prayers, and make sure their resumes are up-to-date.

There is a solution. The Rivian and Tesla solution. Reduce the controller count to single digits. Write the vast majority of the software yourself. Enable OTA updates. Now you have at least an insulated ice cube in Hell's chance of properly testing it.

BTW - here's a shout out to all you QA engineers out there. Software engineers take pride in delivering bug free software. But the software gods created QA engineers to keep us humble. QA engineers are ingenious devils in human form. They know the weaknesses of software engineers and take diabolical delight in proving that your "I tested it myself and I promise it is bug free" software is nothing of the sort. I spent 30 years thanking the software gods for creating QA engineers.

Sponsored

 

[Glen Baker LLC]

Trust me on this. As a former embedded software engineer. It's all about testing. "Joe-doesn't-know-squat-about-developing-software-but-has-an-opinion" describes 99.999% of members on MTC. Just like I know squat about a gazillion topics but I have an opinion.

It's all about testing it to death. That's because you have a collection of controllers, each with multiple versions of third party software and hardware, out in the field. And once you go off the map, "there be dragons". After a vehicle has been in the field a few years, it can become an "n-factorial" kind of problem to fully test it. In other words, impossible.

Really, it's an unsolvable task across a wide variety of configurations and situations (expensive and time consuming). You cannot deterministically test and validate the typical legacy auto hardware-software build the way you would a critical avionics system (unless you want your Maverick to cost $20 million per unit). The QA engineers - and yes they are engineers specializing in QA, will bust their butts on this. Then they'll give a thumbs up, cross their fingers, say a few Hail Mary prayers, and make sure their resumes are up-to-date.

There is a solution. The Rivian and Tesla solution. Reduce the controller count to single digits. Write the vast majority of the software yourself. Enable OTA updates. Now you have at least an insulated ice cube in Hell's chance of properly testing it.

The 22-24 Lariat's don't seem have the problem.
Just saying

 

[Charles T.]

Trust me on this. As a former embedded software engineer. It's all about testing. "Joe-doesn't-know-squat-about-developing-software-but-has-an-opinion" describes 99.999% of members on MTC. Just like I know squat about a gazillion topics but I have an opinion.

It's all about testing it to death. That's because you have a collection of controllers, each with multiple versions of third party software and hardware, out in the field. And once you go off the map, "there be dragons". After a vehicle has been in the field a few years, it can become an "n-factorial" kind of problem to fully test it. In other words, impossible.

Really, it's an unsolvable task across a wide variety of configurations and situations (expensive and time consuming). You cannot deterministically test and validate the typical legacy auto hardware-software build the way you would a critical avionics system (unless you want your Maverick to cost $20 million per unit). The QA engineers - and yes they are engineers specializing in QA, will bust their butts on this. Then they'll give a thumbs up, cross their fingers, say a few Hail Mary prayers, and make sure their resumes are up-to-date.

There is a solution. The Rivian and Tesla solution. Reduce the controller count to single digits. Write the vast majority of the software yourself. Enable OTA updates. Now you have at least an insulated ice cube in Hell's chance of properly testing it.

What the hell are you talking about?

 

[Scott Asheville]

The 22-24 Lariat's don't seem have the problem.
Just saying

Don't see the relevance. Literally tens of millions of vehicles don't have the problem. The point isn't preventing problems. The point is the difficulty diagnosing and solving them when they appear. And that's all about software and hardware architecture.

 

[Packer Bill]

The 22-24 Lariat's don't seem have the problem.

Lariats have SYNC 3 and the XLs and XLTs do not (22-24 Mavericks).

 

[Glen Baker LLC]

Lariats have SYNC 3 and the XLs and XLTs do not (22-24 Mavericks).

It was a rhetorical.
I know the Lariat's have sync and the XL's and XLT's have Connected Touch Radio.
My point was. Ford already has a viable system that works.

 

[bckrupps]

They have to work around the atari chip

 

[billbillw]

The 22-24 Lariat's don't seem have the problem.
Just saying

Mine does...Remember, only the Lariat's with the add on Lux Package for MY 22-23 had the Sync 3. I don't know the numbers, but I'm sure there were quite a few that were sold like mine, ie: Lariat without the Lux package.

 

[CurtisB]

Spoke with my dealer today. He told me Ford still saying a S/W fix. He feels it won't be ready til March 31. Meanwhile, they start and move my Mav every two weeks.

So based on this, it appears that the APIM module does have to be replaced.

 

[CurtisB]

Damn young punks, no respect for us elders!

I once saw a Star Trek The next generation show where the computer system went kaput and everyone panicked...................except the old fart who was trained on an older star ship and could run the darn thing.................without the computer.............

 

[georgeb]

lot of posts for essentially “we don’t know what’s going on”.

also Long McArthur should change their name to Long McDong.

 

[oysterville]

lot of posts for essentially “we don’t know what’s going on”.

Until we do, we don't.

 

[billbillw]

So based on this, it appears that the APIM module does have to be replaced.

None of the affected vehicles have an APIM module. The non Sync 3 models only have an ACM to control everything. Adding an APIM would require a different wiring harness, a different ACM, a mounting bracket, and a bunch of module coding/programming. That probably adds up to 6 hours of shop time. I don't see Ford doing that, unless they absolutely can't get things fixed by cleaning up the embedded software.

 

[torbar]

They have to work around the atari chip

hey, give them some credit, it's a Pentium II chip at least!

 

[Fcnrwy]

I once saw a Star Trek The next generation show where the computer system went kaput and everyone panicked...................except the old fart who was trained on an older star ship and could run the darn thing.................without the computer.............

YEA!!! Mr. Scott!!

Jerry

Sponsored