Evo 8 ECU In 2g Works!!

[the_mork]

At the moment I just want some better global correction, I'm still trying to find my phantom knock and fix a pre-wideband exhaust leak before I worry about real tuning.

This explains why I was reaching as rich as 10.7:1 during wot occasionally.

 

[knochgoon24]

If I use the calculator for -^2(38/43) x 513 = 545.71

I think the calculator is right but after a full day of work I can't for the life of me understand why the simpler method isn't right.

That's the wrong way. It needs to be 43/38
That give you 580.5.

(43/38)=(x/513)
(43/38)*513 = (x/513)*513
(43/38)*513 = x
x = 580.5

And that -^2 out front isn't doing anything.

 

[the_mork]

I think that I have it the right way. Your post does show me why the different methods worked out differently, one is backward.

the -^2 is supposed to be the inverse of the square or a square root function but since I haven't had a math class in several years I've probably managed to remember how that works incorrectly. Now that I think about it that should simplify out to ^(2/2) = ^1 = and identity equation. It also should be 43/38 as you say.

You are correct, I typed the equation wrong, I did it correctly though. The answer should be 545.71. The difference between my answer and Delta's was based on .5psi difference but is irrelevant because we cannot alter the value to be either of our answers, just close, I think 542 is what I'm using for the moment and I haven't looked at the log yet to see if things are better or not.

 

[_Madman_]

You write SQRT like this -> (a+b)^-2.

It's like writing (a+b)^2.

 

[the_mork]

I just thought I should report that I've put race gas in my car and as I expected after reading that jrohner used e85 I still have the same knock issues. Next I think is trying an evo8 knock sensor but they are expensive. Has anyone seen anything from the evo1-5 guys that switch indicating that they have a similar problem?

 

[Ceddy]

For everyone running a EVO ECU, how do you like the Lean Spool feature?

Has anyone played with the "Boost Enhancement Enrich/Enlean Map" and the "Lean Spool" stuff?

Have you noticed faster spool up with the EVO ECU?



Asking because the 98/99 DSM has these features, but they are disabled in the periphery bits, wondering if it would be worthwhile to enable them.


Thanks.

 

[the_mork]

I'm not running it, it is generally considered a PITA because it can change the A/F ratio by up to 1 point and wreak havoc on you tune. I would suggest leaving it off.

 

[Ceddy]

Thanks for the info Mork.

Another question, are you guys able to get rid of the Rear O2 Sensor just by tweaking the periphery bits, or do you have to do a O2 simulator patch or something?

 

[the_mork]

There are two issues with the rear o2 on an evo. The first is a check for a reasonable output that confirms working cat status. This is easy and is controlled by a periphery bit. The second issue is a readiness monitor that throws an exception that lights the CEL if the o2 sensor heater circuit malfunctions (ie: the o2 sensor is removed). If you keep the o2 sensor attached to the harness you only need to flip the bit which is part of FCA, 56D2 is what we use IIRC but I'm not sure which part of that controls the o2, it is somewhere in this thread though. Once the bit is flipped you can use the input for anything that you want.

I think there is work on a patch for the o2 heater circuit but someone else will need to chime in on that because I haven't been interested in it since mine is just wired up along the wall of the sub-frame.

 

[mike96]

yes there is a patch for that. you can check evolutionM.net for the patch. look for the thread labeled "need evo9 to test rear o2 heater sim patch" or something along those lines. i would find it for you but i ran short on time this morning and need to get to work. may by in a bit.

 

[jrohner]

I'm running the o2 heater patch, and used periphery bits to turn off other stuff so I am CEL free, well unless I let it idle for a while cause I get a p0505 Idle code for some reason.

Yeah, the leanspool thing is terrible. You're too lean (or about right), and then all of a sudden way too rich.

I was showing knock even on the stock Evo8 timing maps, even revving it in neutral. I raised the low and mid rpm knock multipliers, and now all is good.

 

[the_mork]

Are you running 9653? If so can you post the xml for the knock control?

Also p0505 used to be a problem for me too until I turned it off in periphery bits. I posted it somewhere in this thread, p0505 isn't listed in the blobits descriptions but it is turned off by a bit that turns off a few p050x codes.


Edit:

I'm currently running with my low multiplier set at 18 and I'm seeing occasional 1-2 knock counts which I actually should see what I can do to take care of since I'm on a stock ignition map with 110 in the tank. I have to say though that it feels amazing to have part throttle response again and the whole pull feels consistent and powerful which is a big difference from before. Thanks to Jrohner for showing that this works and giving me the ambition to do it and write a little bit about it.

This is the entry for knock control on 96530006 I'm still interested in what you changed your values to though so that I have a sanity check for mine.

<table name="Knock Load" category="Knock Control" address="421a" type="2D" 	level="2" scaling="Load8">
		<table name="X" address="6576" type="X Axis" elements="10" 		scaling="RPM"/>
	</table>

	<table name="Knock Multiplier - Low RPM" category="Knock Control" 	address="177a" type="1D" level="1" scaling="uint16"/>

	<table name="Knock Multiplier - Mid RPM" category="Knock Control" 	address="24aa" type="1D" level="1" scaling="uint16"/>

	<table name="Knock Multiplier - High RPM" category="Knock Control" 	address="21de" type="1D" level="1" scaling="uint16"/>

	<table name="Knock Adder - Single Gain" category="Knock Control" 	address="1780" type="1D" level="1" scaling="uint16"/>

	<table name="Knock Adder - Triple Gain" category="Knock Control" 	address="177c" type="1D" level="1" scaling="uint16"/>

Additionally I think this post is pertinent and leads me to believe that we still may need to switch to an evo knock sensor to have refined knock control on our cars.

Marcel, I've not traced the knock circuit, but from looking at the evolution of Mitsubishi knock control, they have brought more into the ECU in the Evo 7-9 than earlier models, eg the average knock sensor noise is now calculated from the knock sensor ADC input. I think it will have a bandpass filter, and is then gated through interrupts. Earlier models used an external chip that did the lot and fed in a fairly well processed knock signal. No evidence of any DSP activity.

Matt, there are some differences between the JDM 9 for knock control and other models. I think sometimes I underuse empirical testing which would speed things up by allowing us to log certain things to show there are areas of disassembly we don't need to worry about. On the other hand, from disassembly you can quickly become pretty definite about what variables are involved in what table. I do believe if your knock sums were always below 6 you would always stay on 100% octane and have no issue with octane ever again, so finding/sorting the noise control stuff is I think the way forward for your and every other false knocker. With high knock sums and their decay it will be very difficult to make much sense of how octane behaves as it is never stable as the update interval is much slower than it is for knock sum. I can be certain that the tables at 4a40 and 4a50 in your ROM are the load values for knock and octane adjustment, although with the stock settings in your ROM I think you'll still have knock control below the load value for knock but just less sensitive. However, that sensitivity is adjustable and able to be clipped by different amounts below and above the load thresholds.

I've been breaking down the disassembly of the noise routines into elements, and trying to identify the constants and unknown variables so I can see a typical path the code would take during normal (non-error) operation. I can see the outcome of the noise detection routine that sometimes sets the knock sum to zero. Following that there are complex routines that establish the baseline noise level. I think the key to the false knocking cars lies in one or both of these areas. I suspect that an engine that just likes to be noisy will have a sudden increase in noise at certain operating points that makes the stock calibration see knock. I don't think that once the ECU registers severe knock it will let go of the idea of it quickly due to two features - 1. On previous Mitsubishi ECUs the base noise estimation is not updated when the ECU believes there is knock and 2. the knock decay is slow. So it is a potentially clever self adapting system, but it needs some work on the constants and assumptions it starts with to suit the noisier engines yet still detect knock.

Knock control - load vs RPM table found - evolutionm.net


I'm adding more to this post again, this is another useful post for us. For anyone that doesn't understand what this is there are two settings that control knock sum. The first is a RPM/load map that either shuts down the knock system or reduces its sensitivity by half depending on the rom and the accuracy of testing. We know it makes the issue go away but it is a heavy handed approach. If you needed to you can set any rpm range to ignore knock in that area up to 159.6 load or any number lower than that. The second method is to adjust the RPM multiplier for the low, mid, or high range. The quote will explain what each range is so no need to go over that again but raising the RPM multiplier directly affects the knock noise threshold and is the more gentle approach that allows us to maintain safe knock control and still get results. Hopefully I saved some people some reading but you should really read the thread anyway.

I wanted to figure out where the transitions occured for the Low, Mid, and High RPM tables on the knock Multipliers. So I set them to 5, 25, and 50 respectively. Below is the log from a simple rev up to 5.4K at idle. Looking at the steps in the knock_base you can see the transition points.

Looks like:
LOW = 0- 3800 RPM
MID = 3800 - 4800 RPM
HIGH = 4800 to redline.

I've got things adjusted and have been doing a little light cruising with NO Knock flashes from the CEL. Where my first day out with the car I got all sorts of knock CEL flashing on just easy driving - THIS STUFF IS AWESOME! Since I'm not really getting into high load cells and haven't been tuning on pulls I'll save the details until I'm fully tuned.

But I think that most engines with false knock should be able to raise the LOW multiplier and get rid of the 3000-3500 RPM problems. And maybe bump the MID a little to make a smooth transition.

You must be logged in to view this image or video.

 

[mike96]

just wanted to let everyone know if they are considering upgrading to an EVO knock sensor, i found a source that isn't a bad price. the EVO sensor is only $10 more than a stock 2G. i would be picking up one mt self but i wont have the spare $70 plus shipping for a little while. that and i need to get my car running so i can get it smoged.

RockAuto Parts Catalog

 

[delta448]

Do we know what the difference is in the two knock detection systems yet? I would imagine it has something to do with how the signal is processed, since the knock sensor itself should be fairly alike for monitoring the same sound frequency range on basically the same engine block, y/n? Is there an additional knock sensor signal processor outside of the ecu for either cars?

I'm not sure I want to skew the strength of the signal with an edited table before knowing exactly what the real difference is and what my stock knock detection would have said if it were in place. Maybe I just don't understand what I'd be editing well enough.

On a side note, I finally got the 9653 ROM working, -with Tephra mods, and without any CELs for the time being. I may sell my Evo MAF and go SD eventually, now that the legwork has been done on a ROM I have successfully ran the car on.

 

[jrohner]

I just adjusted the low & mid rpm knock multipliers until revving in neutral didn't blink my knock cel. Once I did that, I can now run the same timing map I had in my EPROM ECU without hitting knock, where before I got knock even with the stock Evo 8 timing map. I've never even seen 1 count of knock since.

 

[the_mork]

The knock system on a DSM does have an external processing unit which feeds the ecu a nice clean signal. The evo ecu pretty much just takes the signal raw.

The knock multiplier just raises the threshold of what the computer sees as knock. It is just like turning down the knob on an amplifier, things become a bit quieter but real knock should still be way louder than the threshold.

 

[steve]

The knock system on a DSM does have an external processing unit which feeds the ecu a nice clean signal. The evo ecu pretty much just takes the signal raw.

That's news to me. Just where is this external processing unit?

jcbanks who you quoted before doesn't seem to know the difference between the ECU, which is everything inside the box, and the CPU chip.

On a 1G the MH6111 CPU doesn't have the processing power to do the signal processing of the analog knock sensor so the ECU has an custom chip to process the signal. On the 2G and other ECUs the CPU can do it without additional support.
Perhaps that's what your referring to?

 

[the_mork]

I imagine it is, I was passing on what I understood to be the case from jcsbanks, thanks for clearing that up. Thanks for clearing that up.

 

[steve]

jcbanks who you quoted before doesn't seem to know the difference between the ECU, which is everything inside the box, and the CPU chip.

I'm sure he knows the difference but he needs to be more precise so that others, who may not, don't get confused.

 

[the_mork]

So then I have a question, can you shed any light on the differences between the 2g and evo knock sensors? I know it might be out of your field of expertise but I hoped that you might know.

 

[steve]

Sorry I don't have any hands on with the EVO knock sensor.

 

[jrohner]

The only issue I'm currently having is the idle. I have an ISC that works, yet my idle is just stupid. When I start it cold, I have to hold the throttle for a couple seconds or it will just die. Then it idles kinda slow until warmed up. If I turn my headlights on, it's almost a guaranteed die until warmed up pretty good. I swapped the idle maps in my 96530006 tephramod rom to maps from the 91470015 rom, and I havent yet decided whether it's better or worse. Once warmed up, it idles around 950-1000 with the lights on (where I want it) but 11-1200 with the lights off. My 1G always idles great, and that's without a working ISC.

I actually had a better, more consistent idle before even though I always left the ISC unplugged, and I never had to touch the throttle to start it even when it was 0 degrees or less out this winter.

 

[the_mork]

FWIW I just ran around the block with my MAF unplugged and it worked quite well.

I just used the base VE curve for SD and I haven't tuned it at all and I wont be tuning it until probably Tuesday because it is my second anniversary with my girlfriend this weekend and we are heading out of town. I ran fairly rich at idle and mild throttle put me leaner than the stock map but still very safe at around 11.1:1. I think that I wont be using any data from my current logs and just playing with VE until I can get a reasonable setup and then tune it with the fuel map. I just don't feel like I should be simulating a MAF that isn't what the computer was originally built for even though I know the feeling that I shouldn't do that is kind of silly.

The car idled great other than being slightly rich and I think that at least for the ranges that I hit I could probably leave the VE curve alone and manage to dial it in anyway but I know I can get a better baseline so I'm going to. I was nervous about getting into boost but I built about 10psi before letting off (23psi is what boost is set to) with my eyes glued to the wideband.

I also just got my exhaust in the mail so I might put that on before I worry about tuning, I went with the Tanabe Medallion Hyper Spec exhaust to reduce noise and still flow a little better.

Maybe someone else will post more progress before I'm home, I'm interested in everyone else's experience.

 

[tom04841]

I was super lean on first startup with 9055 sd rom. I would use the 9563 v7 sd rom, but I can't get it to start with it. Actually no 9563 rom will allow me to start the car except the v5.10 you sent me.....the immob is disabled as well.......IDFK!! If you have a working 9563V7 sd rom or nonsd rom, wanna send it my way so I can try it out?

Anyway, I started fiddling with it, and got it a little less lean, but ran out of time. I'm gonna play with it a little more this weekend.

 

[the_mork]

Are you running Tephra patched roms? If so try disabling valet mode.