News for hackers. Stuff that matters.
Re:The big automotive semiconductor problem by JKCalhoun on Thursday March 31, 2022 @11:35AM (#30868367)
> Automakers found that mechanical, hydraulic or pneumatic controls failed to achieve enough accuracy and consistency over each vehicle's usage life to meet these emissions tests. This was especially the case as the car aged.

Decades ago, I owned an old Alfa Romeo that incorporated a complicated cam system to control the oxygen ratio to the fuel injection system. More complex than your typical 1-dimensional function of a cam, this cam was "two-dimensional" in its output. The cam follower was able to slide along the axis of the cam based on something or another (RPM? temperature?) and the profile of the cam varied from one end to the other. Now there's a pretty wild "mechanical computer".

A mechanic though told me that over time the middle of the cam would wear — a kind of "saddling" — and engine performance (and probably emissions) would suffer.

Re:The big automotive semiconductor problem by saiya-jin on Thursday March 31, 2022 @11:46AM (#30868498)
Every car engine loses gradually power with wear. Top gear famously showed this multiple times with their older car episodes. Usually it was in 20-30% range.

Basically new car is a 100% system and from first kilometer / month of ageing and exposition to elements various items go down the hill. Each car type had its own set of reasons for that, but result was +-same.

Re:The big automotive semiconductor problem by throwaway0a5e on Thursday March 31, 2022 @11:43AM (#30868461)
>Decades ago, I owned an old Alfa Romeo that incorporated a complicated cam system to control the oxygen ratio to the fuel injection system. More complex than your typical 1-dimensional function of a cam, this cam was "two-dimensional" in its output. The cam follower was able to slide along the axis of the cam based on something or another (RPM? temperature?) and the profile of the cam varied from one end to the other.

V-tech, yo.

Re:The big automotive semiconductor problem by linksnapzz on Thursday March 31, 2022 @08:47PM (#30874001)
You laugh, but the 1980 Alfa Romeo Spider (with the Arese DOHC 4) was the first regular-production car with variable valve timing. This might actually be what the grandparent post is referring to-it had a mechanical variator attached to the intake camshaft to adjust the phase of valve opening....
Re:The big automotive semiconductor problem by dclowd9901 on Thursday March 31, 2022 @12:56PM (#30869514)
Sounds a bit like BMW’s VANOS or Honda’s Vtec. Variable cam timing is super cool tech and a huge reason newer cars are so much faster than old cars.
Re:The big automotive semiconductor problem by unregistereddev on Thursday March 31, 2022 @04:12PM (#30871709)
If described correctly, it is a different concept. Variable cam timing in the context of VANOS / Vtec / VVT refers to variable timing on the camshaft that controls valves.

This sounds different, and I think it refers to mechanical fuel injection. Several automakers experimented with mechanical fuel injection in the 80's and early 90's, but I think the concept died pretty quickly. It was rather complex, failure prone, and less efficient than EFI.

Re:The big automotive semiconductor problem by linksnapzz on Thursday March 31, 2022 @08:34PM (#30873913)
Much earlier than that; there were running mechanical fuel injection protoypes for gas engines before WW1 (diesel got FI quicker); mass-market mechanical fuel injection followed WWII-the Mercedes 300SL had FI in 1955.

Thing is, mechanical FI is almost as hard to tune for emissions and economy as carburetors, so electronic control is necessary.

Chrysler offered an EFI system based on the Bendix Electrojector in '58, but the tech was not fully baked, for one thing, the electronic controls were vacuum tubes in an enclosure in the dash; I think there may be one remaining operational '58 Imperial with that system still running....

Anyway, Bendix sold the patents for the Electrojector to Bosch, which a decade later had developed them into the J- and K- Jetronic transistorized throttle-body EFI systems so beloved by Porsche 914 enthusiasts.

Re:The big automotive semiconductor problem by dclowd9901 on Thursday March 31, 2022 @10:09PM (#30874492)
Oh interesting. The way I read it it sounded like it was manipulating the cams. I have heard of mechanical fuel injection — pretty cool stuff but yeah very finicky.