Over engineered cars are pushing technicians away

[SQL2219]

An interesting video from an auto tech and below text was in the comments that I thought some engineers might appreciate.

It wasn't engine work, but I worked on a friends Hyundai Elantra that had the bights for the head lights stop working. The car had less than 10k miles on it. Come to find out, all new Elantras use a lens on a servo to adjust the focal point of the light to simulate just having an extra bulb in the head lamp assembly. The servo hooks onto a gear that is made out of ABS with no fiberglass reinforcement so the gear melts half the time after prolonged usage if you commute a long way on back roads at night. Oh by the way, this is one of the only parts not covered on their warranties. I replaced the entire assembly twice for her ($400) before just giving up. I ended up drilling a hole in to the assembly, gluing the lens in place and adding a new fuse and wire lead. Then I ended up having to 3d print an attaching assembly to hold a new light that would serve for brights. I had to figure out how the heck to rewire the servo circuit to trigger a relay instead which was an entire other rabbit hole. The lights have never had a problem minus the occasional bulb replacement since and its been 60k+ miles now. But seriously, why do modern engineers try to reinvent the wheel for everything?? I don't even work on cars for a living. I work in software engineering and I see the same thing happening. The same programs need more RAM, more CPU resources just to do the same thing that it did 10 years ago. What does windows 7 do that windows 10 doesn't? Why does the same web page need 60MBs to load when it only need 1-3MB 10 years ago. All this bad engineering is going to catch up to us at some point. It really worries me to be honest.

https://www.youtube.com/watch?v=Op1D7zWwQA8

[bri3d]

I guess I kind of agree with the thesis, although in turn auto technician is gradually becoming an increasingly skilled job and pay is (probably not quickly enough) generally increasing as well. Of course, this is frustrating to consumers as dealer repair costs become astronomical. In some ways the market is working as intended as especially used cars from manufacturers with perceived easier repair and maintenance (Honda, Toyota) now command an enormous premium, at least in the US.

The specific comment is terrible, though:

> Come to find out, all new Elantras use a lens on a servo to adjust the focal point of the light to simulate just having an extra bulb in the head lamp assembly.

It's unlikely this adjusts the focal point, it's more likely it's just a shutter, although this is neither here nor there.

Regardless, this is a normal way to construct high beams with HID bulbs and there's a real engineering reason for this: HID lights shouldn't be short-cycled as they need to warm up before they reach brightness, plus cycling rapidly wears the bulb and ignitors out. So, having a separate bulb is not plausible for HID high beams which need to flip on and off quickly.

Some cars with LED headlights _are_ often switching back to simpler housings without shutters and adjustable lenses, since they're cheaper and easier to build.

This is a case of engineers engineering solutions, not engineers making things "hard" for no reason.

[franktankbank]

Most profit is nabbed by shop owners, by an insane degree (unless you are owner AND operator).

[HPsquared]

Optimizing for cost, weight & performance often comes at the cost of repairability. Buyers compare cars on measurable statistics so they'll buy the one that is cheaper, faster etc.

Watch a Sandy Munro video and see the design details that get a lot of praise. The gigacasting is a nice example: great for performance and manufacturing, but bad for repair. Or, the car not having a separate floorpan and mounting the seats etc directly to the battery casing. Or the octovalve. These are great for everything except repairability.

[kridsdale1]

I guess I break the mold. I buy Toyota/Lexus that are inferior on stats today, but will be fine in 15 years while the other competitors build this year will have been in a scrap yard for 10 already.

[matt_s]

The reason a lot of things appear to be re-invented is that its Resume Driven Development but applied to auto head lights. Or blame product people for needing to come up with something even though the product is probably “done” from a consumer standpoint of fulfilling the needs. Smart refrigerators come to mind.

I keep watching stuff on youtube where people find really old cars, like 50-60 years old, and they walk thru what needs fixing to get it running. Sometimes they add some gas to a 50yo engine in a barn find and replace spark plugs and it fires up, its really amazing.

Man I sound like my grandfather lamenting about “they don’t build them like they used to”.

[jcgrillo]

It's very true, though. There's something to be said about technology you can build without a huge capital investment. With maybe $50k worth of tools and materials you could probably build a vehicle from ~1980. You can't manufacture them at scale, and if you factor in your time it won't pay out, but you could in principle get it done with accurate drawings.

Proprietary software, though, is a huge moat. And the complexity that comes with it. There's no reason we shouldn't be able to build more modern cars as one-offs, except that the systems are so locked down

[daemonologist]

I love my bicycle for this reason - I can do 100% of the maintenance and repairs myself (short of welding a cracked frame, but I could take everything off and swap it to a new frame). Almost all parts are standardized and available from third parties, even in case the original manufacturer goes under. And of course the whole thing cost less than that headlight assembly.

Anyways, I think the reason software follows this path is cost - rather than pay an engineer for optimization you have them work on new fancy features, and just expect the user to buy a faster machine. As computers get faster/cheaper this applies more and more.

[m463]

I think cars will become either disposable or large assemblies will need to be replaced in full.

I'm reminded of how the tv repairman disappeared.

TVs too integrated to fix easily, too complicated to fix economically.

[hibikir]

The integration also leads ro low prices though: A bug trinitron used to be $1500-$2000. Today you can get a 4k TV that is twice as big, much cheaper to run, and beats the 1080i that the largest tubes could support 20 years ago. The lower upfront coat just wins, especially when technology keeps improving so fast you'd rather get the new thing than repair the old one at, say, half the price

[ge96]

I get how mass manufacturing works but it's crazy too seeing people walk on the side of a massive parking lot where cars are just sitting there not used (not sold yet).

[wakawaka28]

Unless cars become extremely cheap (they have done the opposite), it will never be acceptable for them to be disposable. Materials and processes are better than ever yet cars are becoming less repairable. Manufacturers deliberately make things difficult to repair to encourage people to buy new cars, but that only goes as far as people will tolerate it. We aren't talking about pocket change. A new car is usually in the top two largest purchases anyone makes, and they really should be built to last at least 20 years. There's no reason why they can't be built to last 50 or 100 years, or indefinitely (with repairs).

[recursivedoubts]

I have some faint hope that scout will go the other way:

https://www.thedrive.com/news/scout-wants-to-build-evs-you-c...

[pbronez]

Aptera too https://aptera.us/giving-you-the-right-to-repair/

[tonymet]

Transportation has a higher demand for reliability and maintainability than our entertainment devices.

Not every "innovation" is positive. A 2% efficiency gain with 300% maintenance cost and 25% lifespan is a major loss.

Let's act like responsible engineers here and remember what our job is. We are supposed to be creating tools with longevity & utility that improve welfare. We are not here to create toys and infotainment for people.

The tools are in our hands people. Quit placing blame on marketing, customers. Take responsibility for the control that you wield.

[jcgrillo]

> We are supposed to be creating tools with longevity & utility that improve welfare. We are not here to create toys and infotainment for people.

The same can be said for b2b SaaS (which is where I work), and yet...

For some reason we've lost the core concept of what makes a tool good. I think there are a few components (not an exhaustive list):

1. When you are using a tool it disappears. You don't know you're using it.

2. It grows with you and doesn't infantilize you. As you get more skilled it gets more useful, not more limiting.

3. It never changes.

Giving the computer nerds the ability to change and tweak tools while they're in the customer's hand has yielded disastrous consequences. Now everything is a subscription and users are treated like idiots.

[holtkam2]

Engineering is just means to an end. In most cases the end is: make money.

That typically involves minimizing labor (engineering / R&D) costs. That’s why you see solutions implemented in a quick / scrappy way even though it’s often obvious a better solution exists from the end user perspective… the chosen solution was the best solution for the actual objective: maximize profits.

[RGamma]

People need to justify their role. As such they change things around or add complexity, even if user benefit is questionable. And why wouldn't you add some planned obsolescence for good measure?

[neuralRiot]

The “enshitification” applies to everything, as basic functionality is harder or almost impossible to improve features are added as a sales driver together with lower price points, as you can imagine corners needs to be cut as the metrics take in account sales AND profit. General public doesn’t take in account (they’ve driven to think that way) durability or repairability, they just pay a lease so cars have become disposable.

[germinalphrase]

This emphasis on leasing must be a short term fixation. Once companies move onto more mature generations of vehicles, the upgrade cycle will lengthen pushing more people toward ownership.

[jcgrillo]

I was seriously considering buying an Ineos Grenadier because I wanted a nice, modern, safe solid axle 4x4 with decent towing capacity that'll last 50 years or so, but the complexity of that B57/58 and the ZF 8spd, let alone all the other electronic trash like parking sensors, lane sensors, etc scared me off. Theoretically I might be able to keep it running on that timescale with standalone controllers for the transmission and engine but the complexity is just too much. Also, plastic intake manifold? Lmao.

So I bought an 80 Series Toyota. It only gets 1mpg less than the Grenadier. 30 years, all that complexity, and we gained 1mpg.

I'm working on a 1HZ-T swap. So I'll have a 1 wire engine with a nice simple aftermarket transmission controller, and an exhaust brake. I should be getting around 20mpg hwy when that's done and 100k+mi from a set of brake pads. I'm confident I'll be able to keep this running for 50yr.

[dotancohen]

  > 30 years, all that complexity, and we gained 1mpg.

We also got far lower tailpipe emissions, less expensive manufacturing processes, far better rust prevention, and far better passenger and pedestrian safety features. Not to mention a more comfortable ride, almost zero water intrusion, lower cabin noise and vibration, less wind and tire noise, and longer lasting consumables such as plugs, oil, tires, and filters.

[nickff]

>"So I bought an 80 Series Toyota. It only gets 1mpg less than the Grenadier. 30 years, all that complexity, and we gained 1mpg."

The Grenadier complies with many emissions, collision, and other regulatory requirements than the Toyota, and it probably costs less than the series 80 did new (compensated for inflation).

[onetokeoverthe]

But seriously, why do modern engineers try to reinvent the wheel for everything??

this x1000.

80%￶ of everything done since '01 has been the result of crap makework bullshit jobs.

[potato3732842]

You get what you measure. We measure part counts, cost, materials and very specific performance objectives.

Creating an end product that is elegant with intelligently integrated subsystems is not something we care about so it doesn't get done.

[havblue]

So I'm not looking up the numbers for this but to devils advocate...

Acceleration is way up since then, gasoline cars included. Crash tests are better. Collision avoidance and rear views make us safer as well. Reliability probably peaked already though.

[submain]

>What does windows 7 do that windows 10 doesn't? Why does the same web page need 60MBs to load when it only need 1-3MB 10 years ago.

Ads. And tracking code to serve you ads. And AI - that collects your prompts to serve you more ads.

\s

[billy99k]

This is also why all of the auto union are against electric cars. Some person working in their garage will no longer be able to work on cars and get a job as a mechanic/technician and instead will need to have the experience of an engineer.

It's very similar to the horse/buggy and car arguments of years past.

[kogus]

Serious question: why would auto unions care whether people can work on their cars in their garage or not?

[tonymet]

are you advocating or critiquing? i can't tell.

OP is advocating for developing efficient systems that are also easy to maintain. It is a valid concern.

[rUsHeYaFuBu]

Have you seen the inside of a modern ICE vehicle?

You pretty much need trade school to become a qualified mechanic anyways.

Additional, with trade school one can increase the amount they can get paid.

They educate on all vehicle types too, not just ICE.