Lancer Register Forum banner
1 - 19 of 19 Posts

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #1 ·
Hi all,

having issues testing a theory on mapping an ecutek on non standard internals.

Really need an aftremarket ecu to confirm what we think the standard ecu is doing.

Preferably a plug and play type module such as a power fc L jetro.

Will need to use it for probably a day at the most and on the dyno at scoobyclinic.

If anyone can help, please let me know.

Thanks in advance.

Adam
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #3 ·
Mark, thanks so much. Will be in touch today.

?Needed to contact you anyway, as if the test shows that the ecutek isn't up to much I will be needing to be source an autronic!

Can you post up the differences between the models?

Basically I want plug and play and would like to know:

a) does it run maf and/or map?
b) does it support closed loop wideband control, if so, which wideband types?
c) does it have idle speed control?
d) is the graphical psi3 display now working with all the autronic models?
e) does it require additional wiring or sensors?

For anyone interested in knowing the issue, the ecutek is pulling out timing so we can't get consistent runs. The current theory is that the forged pistons are "noisy" enough to trick the ecu into thinking there is det even though there is no audible or det can noise.
 

·
Registered
Joined
·
2,443 Posts
What knock sums is it running and is it reducing the octane number?

There are knock filter tables that can be adjusted, it is something I've not needed to play with using standard internals.

If you don't want to pioneer, might be best going standalone or trying to disable knock control entirely on the stock ECU.
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #5 ·
John, pat is doing all mapping, and ecutek say they don't fully understand knock control and are unable to disable it.

I think unbolting the knock sensor is the way forward.
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #6 ·
don't want to admit defeat on this one.

is the ecuflash software you are using, the same as the ecutek in terms of knock set up adjustability?
 

·
Registered
Joined
·
2,443 Posts
Unbolting the knock sensor I believe will throw a code and put it onto conservative maps. It looks for noise thresholds to be sure it is attached acoustically as well as electrically. Only makes sense really.

Knock setup adjustability in ECUflash is simply a case of defining the tables, you can define and adjust anything by editing the xml definitions as long as you know what to change ;) I did a definition from scratch for the basic maps on the JDM/UK 8 and 8 MR, but didn't look for or include knock related stuff because no one has needed it.

I will have a dig about and see what I can find, but this is obviously highly experimental, but perhaps worth developing?
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #8 ·
have just spoken to Mark Sheed from MAD on the phone and OMG if he isn't the nicest guy I have ever dealt with.

Have to sa big thanks to him and to Dave Gammon for being willing to bend over backwards to help me get this issue sorted.

Am going to try fitting an autronic so we can isolate the engine as not being the limiting factor.

If the engine is fine on autronic, then we know its the ecutek. Then its over to Sam and Pat to see if they can figure out why the ecutek isn't doing its thing.

If it stays unhappy then standalone it is.

Not sure if anyone has suffered from feature creep before, but this car went in for MR cams and ecutek.

MR cams turned out to be FQ300 cams, so hks cams were ordered from the states to arrive asap, so decided to do rods at the same time.

Rods without pistons turned out to be too difficult and slow, so pistons were supplied by Evo400 and we have a full forged engine.

Now ecutek isnt happy (maybe) and suddenly we are looking at stand alone ecus!

If that isn't the fastest decline from, no I only want small bolt on mods, then I don't know what is.
 

·
Registered
Joined
·
2,443 Posts
http://john824.fotopic.net/p37653004.html

Above is a knock sensor filter map (1 of 12) from the US Evo IX. It looks to me like they are likely to be noise thresholds by load and RPM, although the load scaling is different. It might be that maxxing all these values out would effectively disable knock control, I've not looked at the code that references these tables yet though. It would help to be able to log the innards of the knock control system, as I think everyone wants to understand it better, especially those with noisy multiplate clutches, stiff engine mounts, even stiff suspension or external wastegates etc that can trip a false knock retard. It really depends how pig headed you want to be to get the original ECU to behave :)
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #10 ·
john, are we talking ecutek or ecuflash here?

if the latter, do you know the ecutek to have the same tables and adjustability?

also pat mentioned that he saw octane dip as low as 0.95, but I have no idea what that means.
 

·
Registered
Joined
·
2,443 Posts
ECUflash, this is the only way the end user like me can do development which is why I like it, because I can see the unencrypted ROM, disassemble it, modify it etc.

I'll see what I can find as I want to look into this anyway to work out how the knock control finds its thresholds. It may well be very convoluted though.

Octane value of 0.95 or 95% or however you represent it means that the ECU is using 95% of the high octane maps and 5% of the low octane maps. It means the knock sums have been sustained at 6 or more to start changing over to the lower octane maps. Very much like the ignition advance multiplier on the Subaru.
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #12 ·
ahhh, I get you.

am just looking forward now to tomorrow to see what the autronic can hold.

Am not sure if I want it to be an engine fault or not.

Have you run the ecu with the knock sensor disconnected, or perhaps removed from the block.

Pat reckons it will expected a background noise reading in order to work without generating an engine fault code.

I still think that would be the easiest way to check what is going on.
 

·
Registered
Joined
·
2,443 Posts
No, the sensor is a pain to get to.

I don't see the equivalent to the USDM knock filter tables in the JDM ROMs I've looked at. Most tables have something very similar, not these it seems.

There will undoubtedly be other adjustments possible though.

I suspect you'll be buying the Autronic :)
 

·
Registered
Joined
·
3,271 Posts
johnbanks said:
Unbolting the knock sensor I believe will throw a code and put it onto conservative maps. It looks for noise thresholds to be sure it is attached acoustically as well as electrically. Only makes sense really.
Most of the USDM xmls have the knock filters identified, not seen any need to adjust them myself.

The raw knock signal would seem to be a 0 to 5v waveform. The ECU would seem to look of wave form paterms and not just voltage peaks. From comparing the raw signal to the knock sum it's pretty easy to see where the ECU sees knock and add to knock sum. It might however be possible to re-locate the knock sensor to a less noise part of engine.

MB
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #16 ·
burgers,

Makes sense what you are saying, I also have considered your last suggestion.

Mark Sheed's ecu is with Pat today, so time permitting (he has 2 other ecuteks to do today) it will be up and running by the end of the day.

Relocating the knock sensor might do the trick, but the speed of sound and lack of attenuation through solids makes me think that anywhere on the engine block is not going to make much difference.

At least its steel and not ally like the subaru, that would be even worse.
 

·
Registered
Joined
·
3,271 Posts
Hi

I'd be interested to find out what happens, when you take the timing back a couple of degres where you see the ECU pull timing. Does the knock sum still increase to the point the ECU pulls timing again?

What load cell, and what is the timing in that cell where the ECU see's the knock? Would be interesting to see the log.

MB
 

·
Registered
Joined
·
4,227 Posts
Discussion Starter · #18 ·
will have to get pat to reply since he is the one tinkering on the dyno.

Can you even post ecutek screen shots?

I know he is at autosport today and the car will not be back on the dyno until monday at the earliest since Dave and Sam have the dyno all day on saturday.
 

·
Registered
Joined
·
3,271 Posts
Adamantium said:
Can you even post ecutek screen shots?
Don't know about that, the logger is use outputs a CSV file, don't think there'd ba a problem with that sort of thing. Just a bit if text would be interesting, like

RPM=6000 Load cell = 220, advnc=11 knock sum= 11 octain number= 99.898 or something like that.

MB
 
1 - 19 of 19 Posts
Top