Hacker News Reader

Comment by userbinator | Hacker News Reader

Hyundai found out about this pretty quickly, and de-tuned engine output in later iterations to increase reliability.

Coincidentally, I was recently reading about early American cars and their heavily overbuilt construction is a huge contrast to this "value engineering" of modern engines; whereas those old engines could be tuned to make far more than stock power with little effect on reliability, it seems today the aim is to make the margin as narrow as possible --- just enough to get past the warranty period, with the associated effect that, due to how the statistics work, quite a few of them will fail before too. Samsung's "exploding" washing machines are a recent, non-automotive, example of this.

Today's cars may be more efficient and less polluting, but if they last a fraction of the time as the old ones and have to be replaced many times more often, are they really that much better environmentally overall?

View on HN

I'm not sure which era, or which detuning you are talking about, but the 500 cu in (8.2L !!!!) Cadillac engine in the El Dorado made 400 hp when it was introduced in 1970, and 190 hp in 1976, after several rounds of detuning for smog controls, among other reasons.

This chart is enough to make you cry: https://en.wikipedia.org/wiki/Cadillac_V8_engine#500

Anyway, eventually engineers learned how to make engines that are highly efficient and have very low emissions, basically by making the things as complicated as sin.

For a good 10 years or so, all of these goodies they added (like Mass Airflow Sensors, etc) would fail in mysterious ways, and be very expensive to replace.

Now most of that stuff is pretty solid too.

A lot of opinion about car tech is set by stuff that is 10-20 years old, which is both fair and not fair. It's fair, because it's hard to tell how things will wear when they are new. It's not fair, because engineers have already fixed a lot of things that went wrong in the previous generation.

Anyway, Cadillac now has a 3.0L Turbo [0] that makes as much hp as the 8.2 from above.

0. https://en.wikipedia.org/wiki/GM_High_Feature_engine#LGW

Disclaimer, I work for GM, but I'm on on any engine team.

thats not fair. you are comparing sae gross 1970 to sae net 1976. sae gross is naked engine on a stand without any accessories.

in the same wiki link you can find this for 1971.

365 brake horsepower (272 kW), or 235 horsepower (175 kW) in the new SAE net ratings

Fair enough. It's still a very depressing chart.

I was mainly referring to the 50s-60s, when engines could normally be tuned to produce significantly more power than stock without huge decreases in longevity, because their mechanical parts were far stronger than required. The same goes for other powertrain parts (e.g. THM400 transmission.)

after several rounds of detuning for smog controls, among other reasons.

I doubt any of those "other reasons" were because the engine couldn't survive 400hp.

'90s Japanese cars were much the same, you could generally increase the engine power by a factor of at least 1.5x - 2x with bolt-on mods and still keep things mostly reliable.

I just want to say: I owned a 1975 Cadillac the 500 ci engine. I wish I had never sold it, yet I am glad I did.

The biggest engine ever in a production vehicle, until the second-gen Viper came around.

I don't believe for a minute that modern cars last less long than old cars. Have you actually used a 60's or 70's car as a daily driver? The big difference is also maintenance. If you've ever had an old car you know about oil change intervals that's maybe 25% of modern cars, constant valve adjustments, distributor problems, carburetor adjustments, etc.

Modern cars as a rule go hundreds of thousands of miles without major problems while using a fraction of the gasoline per mile and producing a fraction of the pollution.

Yeah, sometimes they get it wrong and design something that doesn't last. But do you think that didn't happen before, too? Remember that the models we know and remember 50 years later are the winners. There's a strong survivorship bias at work here.

None of those 1970s cars broke because the factory used plastic parts in the engine. When you have experience with that and then find "Oh the plastic theromstat housing cracked" on your "modern" car, that appears to be a regression.

No they broke down far more frequently, for different things, and required a tonne more maintenance than modern cars.

Oh yes they did - and the plastics were not as advanced either so they were worse. For example: 1970 Lincoln Continental 460 v8 with plastic timing chain sprockets for "low noise".

I was aware of some screw ups of futuristic engineering from 1970s detroit but not this one!

The tendency I see is that old stuff requires more repairs while new stuff is much more likely to become unrepairable. Pick your poison.

This is a false equivalence. On average, cars last about twice as long today than they did in the 70's. Just since the mid-90's, the average U.S. car age has gone from 8.4 to 11.4 years.

https://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pu...

Doesn't that fit in line with what he said though? Newer cars last much longer than older cars did without repairs, but once the newer car breaks youre SOL

The average age of cars on the road is itself a good measure of when "you're SOL". Cars had a much shorter lifespan in past decades because they'd rust and break down much sooner, to the degree that it was more economical to buy a new car than to repair.

I wonder if it's hard for younger HN readers to appreciate the progress here? Having a car last 200,000 miles used to be a stroke of insane luck, like living to 90 years old - for most car models in the last 10 years, 200,000 is basically the least you could expect to get out of a car.

That's not really the point that's being argued though. Modern cars are undeniably better in terms of how long they last, but you can't go in and just repair all the pieces any more like you could with old cars. That is the point being discussed here

Modern car engines reach around 200k--depending on the make.

The transmissions are another beast. They just seem to die at 140k, plus, or minus how you drive.

Their wasen't constant valve adjustments, or carb adjustments , distributors were dead simple. 90% of the time it was a bad condenser, or the points contacts needed sanding, unless you happened to own a British Vehicle.

And yes Amercan vechicles made in the late 70's, and 80's were aweful.

Old was not necessarily better, but Shade Tree mechanics could work on them.

There's a reason wealthy men buy old classic cars. A few years ago the average guy could buy a vice classic. Today--it's changed. Their are guys stockpiling older collectable vechicle.

Every wonder why the new cars offer just 7 year, or 100,000 miles of the Powertrain warranty?

The two eras have dramatically different duty cycles.

They had a million mile tundra not long ago, I don't think you'll find a million mile engine from the 50s-70s

Here's one with more than 3 million miles: https://www.facebook.com/irvgordon3millionmiles/posts/188158...

Picking a world record is not a fair representation.

How many typical new cars would it take to get to that? How many typical old cars?

This is a rather rose-tinted view — old cars (c. >25 years) broke down very frequently (by modern standards), rusted easily, and very rarely lasted to 100,000 miles. By a substantial margin, new cars have never been so reliable or required so little maintenance.

Conversely, this doesn't mean that cars were made poorly in the past, it's just that the technologies and methods used in their manufacture have improved over time.

> This is a rather rose-tinted view — old cars (c. >25 years) broke down very frequently (by modern standards)

25 years ago is 1992, this is already way into "modern car" territory and million-kilometer executive cars.

Most early 90's cars were garbage. Some would live to high mileage but fall apart in the process. Rust, bad paint, plastics that faded and cracked, terrible suspension, etc.

What the other guy said, but that's also why I said "circa" instead of an exact date. The reason Camcords and Corollics became ubiquitous was their reputation for consistently lasting, on no more than the manufacturer-recommended maintenance schedule and a warranty repair or two, past the 100k mile marker without throwing a head gasket or rusting away.

Nowadays that's the minimum we expect (even for bargain-basement Korean subcompacts,* exotic Porsche sports cars or bleeding-edge high-technology computer-controlled Teslas), but >25 years ago it was virtually unheard of outside the Japanese brands.

*The Hyundai Excel and Pony had such poor reputations for reliability that their brand image still remains somewhat tarnished in the US, while the Daewoo brand had to be retired in North America/Europe entirely (I believe some of their models are now sold as Chevys).

Thanks to the 'heartbleed' mess I was passed several collections of design documents (for then-current and then-future production) from an european carmaker after some less than scrupulous colleagues got into their vpn.

Specification: Engine design lifetime = 2000 running hours.

CAD and FEA are great, but the above spec ruins it. You get parts made of light alloys and plastic, made of the minimum amount of material possible. They happen to be mostly fine when everything is perfect but any unintended load or crack in the plastic due to aging results in a catastrophe.

At least in aerospace, the 2,000 hr spec comes up often as a reasonable service life. We could make everything last forever, but that usually means beefier, heavier parts. How much more in fuel per flight hour are you willing to pay for a product that lasts longer? What about decreased range? Also, how much of a unit cost increase will you accept to pay for larger chunks of aerospace aluminum? Tradeoffs are necessarily driven by consumer demands.

At an average speed of 50mph, 2000 running hours comes out at 100000 miles. If the spec aims for all engines to last this long then you'd expect a good majority of them to last a lot longer. This doesnt seem too bad a goal to me.

Why on earth would you imagine 100,000 is an ok goal? My 2004 Toyota Carolla has around 150k miles on its original engine and supposing you keep maintaining it 250k is not unreasonable.

It seems odd that you would accept a design lifetime that is 40% of a car made 13 years ago.

I'm also not clear why you would expect the majority to perform above spec? Wouldn't a significant number fail under spec? Isn't a care which may become a total loss in 60-80k miles basically tissue paper quality?

Manufacturers often use a term MTBF or "mean time between failure". I'm not sure if the 2000 hours spec is an MTBF or not, but essentially that is an average life with some standard deviation.

Personally, I think a pretty big swath of engines make it well past 100,000 miles now - there are usually other issues with cars at about the 10 to 15 year mark with most makes and models that have to do with rust.

Designing engines that last significantly longer on average (say 250,000 miles) would require either a beefier design, which will add weight and reduce efficiency, or exotic materials that will significantly increase the cost.

Automobiles are pretty amazing for what they are. I agree with many others that they have gotten ridiculously complex, but this is mostly because of the emissions requirements - and that said I do enjoy the fact that smog is not much of an issue (at least in the US) these days.

He said "if the spec aims for all engines to last this long", implying enough safety margins to make the average engine last longer.

My main point was that 2000 hrs sounds quite short, at least to me, but 100K miles sounds considerably longer. I suppose that if a manufacturer's goal is that all their vehicles last at least that long, and they can reach that goal then most will last much longer, especially with some TLC.

I think your average speed of 50mph is incredibly unrealistic, unless the car is driven almost exclusively on the highway.

What makes you think that modern cars have to be replaced many times more often? It seems like replacement intervals have actually been creeping up.

The value engineering he just described. Read his comment again

How many people do you know with 20 year old 2000-2017 cars?

None, as that year range is less than 20 years.

Honestly, the issue there is that the car market pushed people further towards buying a new car on a loan and then trading it in every few years, or leases. Nobody expects to be driving the same car for the next 20 years, and loans are easy enough to get that few people who're doing reasonably ok with money are "forced" to get a used car.

I did not think of this until reading this thread, but the practice of leasing may be driving the longevity of modern cars, which is in stark contrast to what is happening with domestic appliances. Somewhat paradoxically, the leasing model seems to benefit from a robust used-car market.

Without leases, I'm not sure that there would actually be any supply for the used car market. There are some small percentage of people that buy and sell new cars every two years, but most people I've seen buy cars and drive them into the ground until they hit a major, expensive mechanical repair or couldn't get them to pass inspection any longer.

Most used cars I've looked at in dealerships (as opposed to the Craigslist/Uncle Henry's as-is-where-is private sales) were overwhelmingly former lease vehicles.

Precisely - it is a problem for the industry if it cannot easily sell formerly-leased cars at prices well above what a clunker would fetch.

None, since that range would make that impossible, but I have ~170k miles on one of mine at the moment with few issues.

It replaced a car that had 190k miles from around 2000. Its worst problems were a bad A/C compressor (replaced at ~100k miles) and a slipping clutch (it was the original clutch). The plugs and wires were replaced once during its service.

Modern cars are amazingly reliable and amazingly safe compared to their predecessors.

You seem to be looking at the past through rose-colored glasses. I owned and drove cars made in the '60s and '70s. Compared to modern cars the quality was terrible in every way. The only advantage over today's cars is they were simpler and thus easier to fix. And that was a good thing, too, since you were going to be fixing them.

In those days you were lucky to get 80k miles on an engine before it needed a rebuild, while it's not hard to find post-2000 cars with over 250k miles on the original, un-rebuilt engine.

To be fair no one had access such vast amounts of CAD/FEM tools and the computing resources required to optimize them. This probably meant that the choices boiled down to over-engineering or bust.

Reply on news.ycombinator.com