What uC to use in Automotive application

[nctnico]

In many cases the diagnostic error code of a car tells absolutely nothing about the real problem!

A few weeks I ago my car wouldn't start. The diagnostics said that the fuel injection timing valve needed to be replaced. In reality the timing belt was broken. If the ECU had more horsepower it could be made smart enough to see the injection pump wasn't rotating but the camshaft was so it could say something about the timing belt. That would have been a meaningfull error.

[yyrkoon]

ARM is consumer/high volume stuff. I'm not saying it's bad. Just not designed specifically for automotive.

http://www.ti.com/lsds/ti/arm/hercules_arm_cortex_r_safety_microcontrollers/overview.page

No hands on or personal experience with it. But they are being touted as safety MCU's. Have some experience with the Stellaris Cortex M3 line. Which is buggy. Though the more I look at the C2000 line, the more I like. A piccolo seems like it could handle the job easily. Would it last though ? No idea.

[NiHaoMike]

In many cases the diagnostic error code of a car tells absolutely nothing about the real problem!

A few weeks I ago my car wouldn't start. The diagnostics said that the fuel injection timing valve needed to be replaced. In reality the timing belt was broken. If the ECU had more horsepower it could be made smart enough to see the injection pump wasn't rotating but the camshaft was so it could say something about the timing belt. That would have been a meaningfull error.

Wouldn't a broken timing belt prevent the camshaft from turning?

[poorchava]

Timing belts most of the times function well and then suddenly they tear break causing massive damage to valves and pistons, because everything loses synchronisation and pistons hit valves when they are open. Some engines are engineered in such way that this is not possible (eg. GM\Opel\Vauxhall\Holden C20NE).

But such failures that the belt doesn't break completly, but rather stretch are also found. it causes marginal timing change, but i guess if it was en extreme case it could affect vehicle performance.

Timing belt can also get damaged if one pulley or another breaks and belt can touch some static surface (eg waterpump or engine block). It becomes thinner and thinner and in the end will break prematurely.

such things happen, but a tech with 3 month of experience probably won't know that and will only rely on what diagnostic interface says.

[T4P]

Always trust your eyes! Not your diagnostics ...

[SeanB]

Mine said Idle adaptation was outside normal range. Solution? Change the fuel filter, as it was restricted.

[nctnico]

In many cases the diagnostic error code of a car tells absolutely nothing about the real problem!

A few weeks I ago my car wouldn't start. The diagnostics said that the fuel injection timing valve needed to be replaced. In reality the timing belt was broken. If the ECU had more horsepower it could be made smart enough to see the injection pump wasn't rotating but the camshaft was so it could say something about the timing belt. That would have been a meaningfull error.

Wouldn't a broken timing belt prevent the camshaft from turning?

Yes (ofcourse  ). But my engine has a rotation sensor on the injection pump (which is also driven by the timing belt) and the crankshaft. The engine even runs fine if only one of these two sensors is working. It would have been a small effort to check whether the sensors are connected and point into the direction of the timing belt and not a very expensive part (€ 400) which almost never goes bad. Anyway, the engine in my car has been engineered to sacrifice the rocker in case the timing belt breaks. Replacing the belt and a rocker fixed it without having to remove the cylinder head.