jcsbanks

As I just posted on www.lancerregister.com ...

A gentleman who had his car mapped by WRC was not happy with the performance and smoothness. When coming onto boost (using AVC-R) it was very hesitant, and would continue to be through the range. Notably at the same time as it became hesitant the exhaust note changed character. Throughout the boost as controlled with the AVC-R was stable.

Spec of the car: Evo IX GT, aftermarket pistons and rods, Ecutek map, fuel pump, exhaust, panel filter. Everything else including engine mounts, clutch, turbo are stock. There is no piston slap. The tappets rattle a bit. A knocking downpipe on the cross member has been sorted.

He approached me regarding the OpenECU material I had posted and subsequently got a cable and Evoscan. I suspected from the description that the car was retarding the timing because of noise being picked up from the internals. I've seen this before on Subarus with noisy internals. We don't know if WRC plugged in an OBD-II scanner to see what timing the car was running. This would have clinched the problem, without knowing this you're mapping blind. He did his first logs tonight....

Usually a standard car or a well mapped car will run knock sums of 0-3 and an octane value of 100%. This means that at most the ECU is pulling up to 1 degree of timing (if the knock sum is 3) and it is reading exclusively from the high octane fuel and timing maps. If the knock sum reaches 6 or above the ECU will pull down the octane value gradually, and if the knock sum doesn't come below 6 it will keep pulling it to 0% when it is running off the low octane fuel and timing maps. At 50% octane value it will be reading an average of the two and so on in between.

Below are some charts, if you require explanation just ask. It appears to me that the ECU is picking up false knock. The suspicion is the tappets or the pistons assuming the knock sensor is not faulty.

Being able to datalog the ECU opens up the knowledge of what the ECU is doing. On the chart view it shows -10 degrees timing as the ECU pulls out everything it can. I don't like the idea of the car running this, not only will it be flat, but the ECU would also want to pull back the boost, but with the AVC-R on it can't. There is a very strong chance of excessively high EGTs that could damage the exhaust valves, exhaust manifold and turbo.

Knock against RPM
http://john824.fotopic.net/p33769223.html

So much knock the ECU thinks for so long it drops the octane value to zero, so reading from the low octane maps
http://john824.fotopic.net/p33769221.html

Knock against manifold pressure
http://john824.fotopic.net/p33769222.html

Timing drops to -10
http://john824.fotopic.net/p33769219.html

View of some of the data, at 2800 RPM just over 1 bar of boost, -10 degrees of timing!
http://john824.fotopic.net/p33769220.html

mrfred

iTrader: (50)

Looks like the knock sum going away above 5000 rpm. Any idea why its doing that?

jcsbanks

He simply didn't dare take the car to higher revs under decent load because it ran so badly with this retarding.

The knock sum has minimum load boundaries to work. Octane value ditto.

This is the most extreme example of knock sums and octane posted so far.

dan l

iTrader: (3)

First, where is the octane value in the usdm ecu?

All I can say is that we don't have many issues in the USA with built motors and knock sensors. Sometimes the machinist bores incorrectly for the expected piston expansion rate and you run into issues. You can't hear these motors "slap" but they do cause knock issues. Also noisey Hydrolic Lash Adjusters (HLA's, not to be confused with a pushrod motors lifters) are of a huge concern. However they usually cause weird tip in knock and as oil pressure climbs they become quiter. However on new motors sometimes debris can clog them and cause a few to stick open and be noisey. I'd be replacing my HLA's if I were you.

donour

iTrader: (6)

Search.

https://www.evolutionm.net/forums/sh...=octane+number

This isn't jcbanks' car.

d

C6C6CH3vo

iTrader: (9)

Forged noise

TTP Engineering

iTrader: (465)

I have seen this on 2 evos at partial throttle 17% and up to 36 counts from 3000-4000 rpms at 80-100kpa.

< 0 psi...

I think I may have located the problem in one motor however the rod bearings are perfect so not certain if it was an out of balance vibration or what...

JohnBradley

iTrader: (30)

Nice bend

ST

iTrader: (8)

assuming it is really a false knock, as you earlier surmised, one experiment you could perform is to map ou the low octane ignition retard altogether and see if the ecu will pull timing anyways, as guestimated in another thread. this would be interesting to see if there is further "failsafe logic" in the ecu not tapped into yet. again, i would only caveat this with the presumption the knocksums are false.

mchuang

iTrader: (11)

Dam did he gouge some water into that cylinder with the bent rod?

TTP Engineering

iTrader: (465)

The main symptom besides the part throttle miss was the fact that coolant was dissapearing from the radiator every few days.

I looked at the HG and it does not appear damaged. Not sure if there is a leak in a coolant hose or a crack in the head.

Replacing all valves, guides, springs, stem seals, Ti retainers, seats with Supertech parts in the rebuild. Should be up and running by Monday. Guy is obviously also going with rods and pistons with new bearings.

-=SPECTRE=-

iTrader: (4)

Holy shiite militia, one of those things is not like the others.

jcsbanks

I can see in the logs where the octane number reaches zero it still pulls timing using knock sum. However, it appears to "only" retard to -10 and then hit a limit - probably a software/hardware combination limit.

TTP Engineering

iTrader: (465)

Here is a part throttle log of what appears to be similar. I have so many damn logs I pulled 35 just to find this one. I do have others that pull timing to -10.

dan l

iTrader: (3)

Yep thats a classic case of part throttle knock above. Does the car have thin oil or noisey HLA's? Vary rarley is part throttle knock like that real.