Hacker News Reader

Comment by agoodthrowaway | Hacker News Reader

I race motorcycless and so have some experience with high horsepower small engines. My first thought is I wonder how high the compression is to get this HP with only 3 cylinders? With high compression, everything wears much faster and components like pistons, connecting rods, and bearings need to be replaced at regular intervals for the engine to remain reliable. Additionally frequent oil changes become necessary as the oil breaks down more quickly under these conditions and metal shavings from wear build up in the oil. Things like connecting rods become stressed and need to be replaced at regular intervals for the engine to remain reliable.

I’d imagine that the Konigsegg buyer probably doesn’t care about maintenance costs but they might be irritated at the service intervals.

I wonder how much maintenance that will be?

View on HN

Mean Piston Speed [1] is a good indicator of engine longevity.

~16 m/s for automobile engines

~25 m/s for Formula one engines

~26.5 m/s for Koenigsegg’s 2.0-Liter

[1] https://en.wikipedia.org/wiki/Mean_piston_speed

And Formula 1 engines are only meant to last hours (yes really, most engines don't even last one season), albeit at ridiculously high stress levels.

If we extrapolate from this, where high performance drag cars typically last minutes (20 years ago they only lasted seconds), that would mean this engine might only be good for a couple of hours of driving around the track. Assuming this is true (I am not saying it is), this engine would be pretty worthless for anything other than being a collector's item or being used for 1 or 2 races before it had to be retired.

> yes really, most engines don't even last one season

"Even" one season? If they last more than one race it means they didn't push it hard enough so it makes sense that the engine last just marginally more than the race.

The new rules set the limit at 3 engines per season, which is 21 races plus testing. So it's a balancing act, but you definitely need to reuse the engine for more than 1 race.

And for those not in the know, a F1 race is ~305km, and they have do two days of practice plus qualifying in a race weekend using the engines they have (same engine for qualifying as for racing). There's some more detail in this[1] article, where they point out the Mercedes F1 engine did over 3000 miles (~4900km) during pre-season testing without issues (most in race-like conditions).

That said, from my impression it is usually the turbo or the hybrid systems that break down, it's rare for the actual engine block to be the issue barring specific production issues.

[1]: https://autoweek.com/article/formula-one/mercedes-f1-engine-...

F1 regulates the maximum number of engines a season (to 3 currently). So they have to last ~7 races.

Edit: old numbers updated

and F1 tires only last a few laps. it’s all designed in. the F1 engines don’t expire in a few races because they can’t build them more robust, they expire in a few races because the rules require them to last that long. they could last all season (yes, with same performance) if they were required to do so.

I'd be surprised that they could build tires to go on for 22 GPs with the same performances, but who knows. The goal was raw speed when there were multiple manufacturers. The only year with a rule to forbid tyre changes during a race was 2005. Maybe you remember that Indianapolis GP with only 6 cars racing because thr banking destroyed the tires of the other manufacturer (which won all the other GPs.)

Indeed. Back in the days they used to weld the cylinder heads to the engine block before qualifying, so they could run it that bit harder to get that extra bit of performance. Obviously not something that increases the lifespan of the engine...

Similarly in current F1, they know quite well how much life they have of the engine, and how much life a quali lap takes from the engine compared to a calm outlap.

If the regulations mandated a single engine per season they could do it, though they'd mostly just turn everything down.

7 races per engine including Saturday practice and qualifying. It's about 5 hours per weekend times 7. 35 hours, which a commuter car does in about 10 days.

Interesting that they use the mean of the absolute value instead of root-mean-square as in other sinusoidal applications (63.7% of the peak value vs. 70.7% for RMS).

RMS has all sorts of interesting properties, being directly proportional to effects that result from the square of the quantity being measured such as force on the connecting rods or acceleration of the piston, but mean piston speed is easier to calculate from familiar quantities to an automotive engineer like stroke and RPM. I wonder if engine longevity is actually proportional to mean piston speed or RPM, it would be easy to mistake the 7% difference given all the confounding factors...

If they're only different by a constant factor, then both have the same interesting properties and neither is much more difficult to calculate than the other -- at least for sinusoids.

If something is proportional to one, it's naturally proportional to the other.

Almost anyone on this website could answer better than me for this, was always weak in math, but I believe they differ by a constant factor for a sine, but for a more complex waveform they will not (well, the amount they vary by would be different for each waveform).

It's true that the factor between RMS vs peak-to-peak is different for e.g. a sine vs a sawtooth wave, but for other waveforms its still a constant (just a different one), and for this engine it should be just about a sine wave anyway.

Circular motion about a crank produces a precise sinusoidal waveform.

Koenigsegg Gemera use their direct drive system, with only one gear. So the engine will only see max revs when you're traveling at top speed, which will probably be quite rare since it's 400 km/h (249 mph).

Cars in the class of Koenigseggs are actually rarely driven, things like wear and tear are often not a concern at all, what's more important is exclusivity and exotic-ness. Maintenance is something a buyer in this class doesn't even consider in my experience.

They're actually marketing this freevalve tech for consumer cars. They already have a tech demonstration partner (a Chinese car maker).

I'm curious if it will be adopted. I assume cost is a big issue. Hyundai considered it, but decided to roll their own (of course they were already pretty invested in their own technology, but Hyundai is known for reliability and Koenigsegg is not).

https://www.roadandtrack.com/new-cars/car-technology/a312246...

Konigsegg is part of the echelon with bugatti owners that have warehouses of cars

the continuously variable valves is the key here, you can have the engine make good power at very high rpm, AND very low rpm, AND in between.

Exactly. VVT (variable valve timing) was promising for decades, starting back from the early 1900, then in the 1980's VTEC ( https://en.wikipedia.org/wiki/VTEC ) was much more promising, and now this. The delay between each major step is approximately half the previous one, let's plan for a new one in ~20 years.

1.0 litre 3 cyl petrol engine in my car makes 120hp and gets 50mpg extra urban. It has dynamic servicing and first service is currently shown as 550 days away. Obviously the 2.0 konigsegg makes 5 times the HP, but maybe double the cylinder volume and half the service time and you can ramp up the compression enough to multiply the hp by 5?

The article says 600 hp on high test fuel, 500 hp on conventional, but that’s still 167 hp per cylinder which is nuts.

That's relatively low hp for the displacement, so it's harder to extrapolate.

For another comparison the 3 cylinder 765cc engine from Triumph (street version) does about 120 or 125 with a red line of 12.5k if I recall correctly. In racing form (i.e. the moto2 engine version) it pushes about 140 (I think mostly via tuning and a bit higher red line). This is naturally aspirated but probably a good rough guestimate for bounds of what you can do on regular fuel and air. This also shows you why just ramping up the compression won't get you there, you need to change the air pressure too. For the street version of the triumph engine the service interval is something like 10k miles, valves every 2nd one.

If you scale that linearly you still "only" get close to 350 , so you have an idea of how much stress is on this design to push 600 on 2l.

By comparison the inline 4s in motogp make 250+ from 1 liter, so that's getting closer. They do probably represent something close to what is possible without induction though.

> If you scale that linearly you still "only" get close to 350 , so you have an idea of how much stress is on this design to push 600 on 2l.

-Removed, I misunderstood the original post, still you can design an engine with more power if torque requirements are low, and they claim 280hp at 2l which is a lot but doable if you don't intent to run it like a roadcar and have infinite budget like with those super/hypercars-

updated:

Agree low torque requirements help, which has a lot to do with the rest of the drivetrain, i.e. how you want to actually deliver power and at what speeds.

That's impressive. My car has a 4 liter V6 gasoline engine making 264hp, but fuel consumption is more like 17mpg.

It’s also a high rpm motor, which will still require lower tolerances and better materials. I don’t recall but don’t higher rpms have a lower rate of wear than higher compression?

Wouldn’t it require higher tolerances?

It’s kind of counter-intuitive, you’d expect lower tolerance to mean the tolerance for deviations is lower, or higher precision

In practice, “high tolerance” is used to mean “tight tolerance”. “Tight” is more commonly heard, at least in the US.

Lower tolerance numbers, higher precision

Tolerances are like margins of error. As the error range goes down, the exactness of the specification goes up. Or as the others said, the precision.

If it makes you feel any better, I have to pause for a beat any time I try to put an adjective in front of 'tolerances' to make sure I don't sound like a dope.

The cylinders are especially large, but the displacement is still only 2ℓ. I think the unusually high horsepower for a 2ℓ engine is not just because of high compression but also because ⓐ they normally run it at especially high revs, like your motorcycles; ⓑ they run it on a two-stroke cycle at low speeds, up to 3krpm; and ⓒ they optimized it to run on alcohols, which as you know have lower energy density but are better at keeping the engine cool.

The engine does not run as a 2 stroke below 3k, the article says it theoretically could but they haven’t tested that.

Also, 8,500 RPM isn’t ‘especially high revs’ these days...

And finally, it still produces 500HP on pump gas - even that is outrageously more than any other 3cyl engine available.

8500 is still pretty high for a turbo. most performance cars with turbos don't have a redline that high.

Huh, I appreciate the corrections.

The service intervals aren't good on a lot of cars in this class. This is mitigated by a lot of them not being driven much.

These kinds of cars go bad by just being parked. Even if you don't move it at all in a year, you probably still need to have it serviced.

I don't mean never driving, just rarely. So often in practice that means once a year even if your interval is 7 or even 5k. Nothing onerous.

Besides, if you have one if these you have a bunch of other cars.

Car oil can get additives that motorcycles cannot get because the engine and transmission share oil. Friction modifier can extend effective oil lifespan. 600 hp also requires e85.

Not totally true, a honda CRF450 for example has separate crank case and transmission oil

Reply on news.ycombinator.com