Theory: Heat sensors in cylenders

[Nim]

I was just having a think before about a purely theoretical situation. We all know that detonation is bad, and that once you start boosting an engine you can run into making things too hot in there.

Now I was thinking, why not have a temperature sensor in each cylinder that reports back to the ECU? If things start getting a bit warm in there pre-spark, the ECU can start pumping extra fuel into just that cylinder to reduce the temperature. It could also measure the detonation heat, and if it starts getting up to 'melt your pistons' level, it could add more fuel again. It'd mean you could run much closer to stoic right up to redline, and just let the heat sensors tell the ECU to add more fuel when it starts getting too hot.

So, two question headers:

1. Is it better? Assuming it can be done, would this be a good way to make an engine run more efficiently? Do you think it'd be better than current methods?

2. Can it be done? What would be the issues with doing something like this? Getting sensors into the cylinders? Getting sensors that will remain accurate enough? They have laser heat sensors, so could you could make the lens of the laser deal with combustion heat and pressure?

Cheers,
Nim

[Norbie]

I think what you're looking for is pressure sensors, not heat sensors, to detect detonation. Any sudden and large spike in cylinder pressure would indicate unstable combustion. I remember reading an article in Autospeed which covered something along these lines.

[Nim]

Well the idea is to pre-empt detonation. When the sensor picks up too much heat in the cylender, the next time around it'll start putting in more petrol to stop the heat from becoming a detonation issue. And as I said, it could also be good for racing applications, to tell you when the pistons are getting so hot they're about to melt.

EDIT: It could also detect detonation by sensing the heat spike from the detonation before the ECU has sent a spark. ECU will look at see "oh shit, it's already detonating." for those times when you get small flashpoints in the piston which are localised and may not be picked up by the heat sensor.

[YLD-16L]

When the sensor picks up too much heat in the cylinder, the next time around it'll start putting in more petrol to stop the heat from becoming a detonation issue.

Isn't this what an EGT (Exhaust Gas Temperature) sensor tells you?

EGT goes too high you know you have timing/air:fuel issues.

[Nim]

Yeah but you wont know in which cylender. This way you can target each piston.

I was thinking along the lines of the old 7M-GTE and 1JZ-GTE running rich up high. With a temperature sensor right in the cylender, you wouldn't really need to run it richer untill it was about to become an issue.

I dunno, it's just an idea.

[TurboRA28]

I have always been tempted to get a EGT probe in each exhaust port on the manifold itself. Easy enough to tap into each one especially if its a custom steam pipe or stainless manifold.

But not really wanting to have 4 gauges to read each port. Maybe some electronic guage that can read the 4 sensors?

[ViPeR_NiPPleX]

u want a microcontroller to monitor the sensors, so if one is abnormally high it will point it out instead of trying to read 4 or more guages all at one

[river]

Hi,

You can get EGT and CHT sensors quite easily. The small Rotex and similar brands of ultralight aircraft engines have them. There are aviation suppliers that sell this stuff.

As previously mentioned the best way to monitor a stack of these sensors would be to feed them into a microcontroller, which could run a cyclic display and also alert you when an abnormal reading has been detected.

seeyuzz
river

[TurboRA28]

I think the WB02 meter can read up to 8 inputs.. Probably could use this - http://wbo2.com/3b1/default.htm

Added bonus of being a wideband 02 meter also

[NeoNasty]

Blue tooth?

it sounds like a good idea, but I think in practice trying to make a reliable and accurate sensor that can resist 8000RPM for instance would be hard.

I think the seperate sensor on each cylinder is prob a better idea, with a micro to watch over it all. I wonder in reality, how much of a 'better' tune you can actually get by doing this. What does F1 do?

[Grega]

IIRC the guys who setup the MORBST celica did this.
they ran individual temp sendors on each combustion chamber on the 1G
i don't know if this was to aid/abet overly high combustion chamber temp OR something to do with cooling (the article did seem to mention something about the 1G boiling water in the head/developing steam pockets - leading to cracking ect)

[Nim]

IIRC the guys who setup the MORBST celica did this.
they ran individual temp sendors on each combustion chamber on the 1G
i don't know if this was to aid/abet overly high combustion chamber temp OR something to do with cooling (the article did seem to mention something about the 1G boiling water in the head/developing steam pockets - leading to cracking ect)

How did they manage this? Drill an extra hole in the head / block and run a laser heat sensor? Or was it on each exhaust port?

[kingmick]

Nim we do it with every engine! one for each pot wether 4cyl or 8cyl. you weld lugs into headers a few inches from the head! works well on either atmo or turbo!
its import to have and you always do it on an engine dyno! my 4ag will have the same thing, when i get around to it.

[Nim]

Hrm, I see. But not into the block itself? So it can just detect exhaust charge temperature? Is that accurate enough? Would there be any advantages to having a sensor in the cylender itself?

[kingmick]

pressure yes, heat is fine in the exhaust and accurate!
as has been said, they use them on aircraft on a lot of diffrent engine types!