Hacker News Reader

Comment by agurk | Hacker News Reader

On all the fuel injected engines I have owned there is a physical cable that controls the position of the throttle plate. There is an airflow mass sensor the other side of the plate to measure the amount of air and therefore how much fuel needs to be injected. So interestingly in these sort of engines you're really just controlling airflow to the engine rather than fuel/air mix like on one with a carburettor.

More modern engines have electronically controlled throttle plates, and this is definitely somewhere you could do something clever like you suggest - cutting fuel flow but also maximising airflow when there is zero throttle input.

I assume engine braking is generally considered a beneficial thing by manufacturers, but it could be fun to be able to customise the amount. Or do something like have the braking come on gently at first then harder. Maybe even try and have a linear or flat response curve vs. engine rpm.

View on HN

> cutting fuel flow but also maximising airflow when there is zero throttle input

You don't want to do this. Much of the engine braking effect is from pulling the intake air charge past the mostly closed throttle plate. On a car with a wide open throttle plate [even with no fuel], the engine is acting more like a spring than a damper. On the intake stroke, it will pull an intake air charge past the small restriction of the open intake valve(s), then compress it on the compression stroke, then release that compressed energy on the "power" stroke, then exhaust it past the small restriction of the open exhaust valves. Pushing air past the valves will cost energy, but it's not much.

This is why diesel trucks' engine braking works differently. (Diesels don't have a throttle plate.) They can open the exhaust valves to prevent the energy recovery in the "power" stroke to create a higher net braking force. Jake Brake: https://en.wikipedia.org/wiki/Compression_release_engine_bra...

> Or do something like have the braking come on gently at first then harder.

You can do this by letting go of gas pedal slowly. I have "current amount of fuel used" info in my car (liters/100km), it shows pretty clearly, that when going fast and slowly letting go of gas, amount of fuels slowly goes to 0. If I let go of gas fast, the engine is intelligent enough to not close throttle as fast as possible, still probably takes 1 second.

> More modern engines have electronically controlled throttle plates, and this is definitely somewhere you could do something clever like you suggest - cutting fuel flow but also maximising airflow when there is zero throttle input.

They cut fuel flow and close throttle plate almost completely but still allow some small amount of air, in order to actually do engine braking. If you need to coast, you can apply clutch in manual. Don't know that much about automatic, but from what I've driven, they use "lift gas" as a "engine braking" signal, so probably they can't really coast that good.

I've got both at the moment on my two Hondas. Both manual, one with a throttle cable and one with throttle-by-wire. There's quite a bit of difference in how they handle off throttle. The cable will just slam the throttle shut if you just jump off the pedal (obviously) and it jerks pretty hard. The throttle-by-wire car hangs the throttle a little when letting off and doesn't just immediately start decelerating. Then it's much smoother once it does start slowing down. The ECU definitely doing something to smooth it out.

Funny because the cars build dates are only 2 years apart, 2005 and 2007, and they're both K20 engines but the engines handle so different.

The ECU is doing that because rapid changes in state are bad for emissions. Letting you just slam the throttle closed could result in a tiny, but measurable at OEM scale, amount of extra fuel going half burnt out the tailpipe. Slamming it open can cause too lean combustion and oxide byproducts which.

The OEMs try real hard to prevent this because the amounts of emissions byproducts that aren't water or C02 they're allowed to produce are on the order of single digit grams per multiple miles (you can mentally file it as "about the baseline air quality in urban areas" though the rules are hugely more complex than that) so these edge cases matter.

An idling I4 has about 10 injection cycles per second. And the ECU clocks injection time corrections at least at that rate, more likely at double that rate or more. So I think that the smoothing is mostly there for the owner's wellbeing, not emissions.

> it could be fun to be able to customise the amount

Some recentish motorbikes have an option to customise the amount of engine brake, I suppose cars could have something similar, too.

Reply on news.ycombinator.com