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
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.
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.
The real secret finding what specific platforms/engines/transmission packages have earned a track record for being reliable.
Most Toyotas sold are their reliable models so they get a good track record, but Toyota has put out some poor engines even in cars otherwise known as being reliable, like the Camry
BMW has had some extreme stinkers like their early hot V V8, but there's also BMWs that package the B58 and ZF8, which make such a reliable combo that they're in a currently sold Toyota.
I recommend anyone buying a car and worried about reliability search their specific engine and transmission to find issues, and avoid first generations with no track record because a very common story is a refresh fixing a design flaw in an first revision
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”.
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
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.
I'm reminded of how the tv repairman disappeared.
TVs too integrated to fix easily, too complicated to fix economically.
https://www.thedrive.com/news/scout-wants-to-build-evs-you-c...
I applaud the sentiment, but this has about as much impact as a random blog.
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.
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.
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.
Exception may be some European cars which have a reputation for very expensive dealer maintenance after the warranty period.
It is an inevitable consequence of modern rabid capitalism. The 99.99% shall own nothing in our name, while the 0.001% enjoy their life on the rent every single one of us has to pay for the privilege of existing.
More and more of basic life necessities are gated away behind subscriptions (and yes, I am counting a basic computer with an office program suite as such), or they are rapidly depreciating assets, and what remains gets bought out by the rich in fire sales during economic crises (remember the "for sale / foreclosure" signs 2008ff?) - and the frequency of the latter has only increased.
Of course this all has an end game because eventually people will have no money left to even lease these things, but by then those who are in power and enjoying their life will be long gone.
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.
> 30 years, all that complexity, and we gained 1mpg.
They’re incredibly safer and yet the cost hasn’t gone crazy.
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).
I have very little confidence I'd actually be able to keep a Grenadier on the road forever. I suspect it's at most a 20yr car. We'll see maybe in 30 years I'll get a used one and prove myself wrong.
this x1000.
80% of everything done since '01 has been the result of crap makework bullshit jobs.
Creating an end product that is elegant with intelligently integrated subsystems is not something we care about so it doesn't get done.
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.
I am not a mechanic but modular systems are great. Transmission and engine. That's all I want in computers to be involved in.
Ads. And tracking code to serve you ads. And AI - that collects your prompts to serve you more ads.
\s
It's very similar to the horse/buggy and car arguments of years past.
OP is advocating for developing efficient systems that are also easy to maintain. It is a valid concern.
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.
The Land Cruiser requires a lot more of my time working on it, but it’s a dream. I can fix something in a couple hours. The diesel, it’s a nightmare. Everything sucks working in it. Access is horrible, I end up having to jack into the CAN bus all the time, I spent a while with an oscilloscope plugged into it a month ago, and I’ve had to write my own software to interact with it.
Modern cars are more computer than car, and they are pushing more and more towards being fixed like them. I’d rather work on an electric car… what wears out? The cooling system? A bearing? Simple.