Seijuro

Well, I finally got the Evo8 ECU installed in my 91.

Fired up on the first try, no funny noises or nasty shakes.

Thing is, it's running steady at 2000 rpms, and I'm not really sure why.

I'm greener than green when it comes to playing with these roms, though the time to start my homework has begun.

Any thoughts for you guys who have come before me?

Thanks!

0xDEAD

iTrader: (3)

Is it warmed up and idling that high? If you have the engine fully up to temp and the cooling fans have kicked on at least once and it is still idling that high then you probably have to set the BISS screw. Connect evoscan, click the button for the BISS actuator, and set the RPM to the RPM you have set in the rom for the car to idle at, I think stock is 750rpm's. Unactuate the BISS in evoscan and it should idle fine.

0xDEAD

iTrader: (3)

Also if you are still running a rom that uses the MAF you will have to port over the MAF settings for the 1g into the evo rom. I have no idea what the settings are because I run speed density and I don't have a MAF to worry about.

Seijuro

Forgive me if this is a dumb question, but do I do this while the car is running? Or do I warm the car up, turn it off, turn it back on, and then set the BISS with Key On Engine Off?

0xDEAD

iTrader: (3)

You set the BISS with the car running. If you search the Internet or maybe this site you should be able to find the instructions.

E-SPEC INDUSTRIES

iTrader: (10)

I wanted to take them time to thank all of those who contributed to this thread.

We have a 1g up and running now on with a 6262... Should be fun!!

0xDEAD

iTrader: (3)

Its amazing when it first fires up huh? You look at that ecu and your like, no way. But sure as **** it fires up and runs better than a stock DSM ecu.

Seijuro

I see, I thought you meant somehow the RPM was set using evoscan.

Well, the BISS isn't the issue. It's actually adjusted almost all the way in as it is. Adjusting it out only causes the RPMS to climb even further.

I noticed however, that as the ISC steps decrease, my RPMS increase.

I think that's essentially backwards of what's expected. Looking at the chart in ECUscan, the ISC steps are decreasing as the Engine Temp increases.

Sure enough, as my engine temp increases, so therefore the ISC steps decrease, and my RPM increases.

So, I inverted the chart. With -32 degrees at 45 steps, and 82 degrees at something high like 190 steps. I just roughly mirrored the graph with inverted values by eye.

Sure enough, I'm idling at about 750 RPMS at 190 degrees F (I can't figure out how to set evoscan to use Celcius, and there's zero documentation for this program ).

Solved right? Well not really. The computer will decide sometimes to start decreasing the ISC steps, regardless of the fact it's in contradiction to the charts, and I'm not sure why.

The other thing is, nobody else is having this issue? That's baffling to me. I double checked the wiring, and it's all spot on. The thought has crossed my mind to swap the wires at the harness for the A and B side coils for the ISC, but I'm not sure what other impacts that may have, and if I that would really fix my problem, everybody else should have this problem too.

0xDEAD

iTrader: (3)

Yes, that does not make sense. Also the ecu does make on the fly adjustments that will affect the idle motor steps. I would switch the A and B wires for the ISC motor and see what happens. I don't run an ISC and not many people have gotten one of these up and running, so maybe it is a legitimate problem and needs documented. Maybe you can be that guy!

Edit:
You may want to also try swapping the pairs of A and B wires. Looking at how the ISC coil works, I don't believe you will cause any damage to anything. If you are worried put half amp fast blow fuses in line with the wires temporarily just to be on the safe side.

Seijuro

Good point, I was told this is bleeding edge stuff, but I don't think I took that as literally as it was meant. I was reading back through this thread and saw that on the first page you mentioned that you do not have the ISC hooked up. FAIK Jack doesn't either.

I think I'm going to give that a try and see what happens. I basically came the some conclusion you did on the ISC coil. Reversing polarity should just drive it backwards, I don't see how anything could be harmed.

Also, I think I may not be adjusting the ISC steps right in openecu anyway. I have the older 1.3 (not 1.3U) cable, so I'm relegated to using evoscan 1.29a.

I've since also installed 1.42 and I see that there's more stuff in here. There's an "Idle Stepper Look-up Table" which I did not mess with but it's got "load" on the Y axis, which is still very arbitrary to me.
What I was inverting, is the "Desired ISCV inital step position..." values, which is worded differently in 1.29a also. I think this is basically just setting a starting point. The initial settings as well as the lookup table should probably corroborate, and inverting one without the other may be why I was getting such odd behavior.

Seijuro

Yeah I'm catching hell figuring this one out. There's some charts in this thread for the 1G MAF, but the Titles of those charts do not correlate with anything in my 1.29a or 1.42 ecuflash versions.

Also there's reference to an "Maf adder", which does not exist (not in plain sight anyway) in either of my verions.

Seijuro

Well, this morning I decided that I wanted to give modifying the tables another try, before I do any sort of electrical trickery.

The ISC steps appear to be a 0 - 255 decimal value. Following my theory that the operation of the ISC on my 91 DSM is inverse of the EVO, I modified my charts with the following logic.

In the Evo 0 = closed, and 255 = wide open. Therefore in the DSM 255 = closed and 0 = wide open. In the Evo, steps are counted up from 0 to indicate how open the ISC is. Therefore in the DSM, steps need to be counted down from 255 to indiciate how open the ISC is.

Here is an example of the original ISC settings.



I took the value in Steps, which is counted up from 0, and subtracted it from 255 to reach the equivalent inverted value that is counted down from 255.

Here are the finished results.



And it works! I now have a fast idle when cold, that tapers down to an idle ~ 900rpms when warmed up.

It's only been a few hours and one brief test drive since the change, but the results so far are very promising.

0xDEAD

iTrader: (3)

Ok, now the trick is going to be properly changing the wiring to work with the stock evo rom. The reason being is that not every single table that alters the ISC is guaranteed to have been found. Thus you want to make the rom actuate the coils properly instead of trying to make the rom, of which you can't find all the tables, do the work for you. Good find. You should be able to just adjust your harness real quick, its only 4 wires.

Seijuro

This is my intention.

I first want to give my current "soft" fix some time to reveal any bugs or weird behavior. Your right, we don't know that all the tables have been found. One thing that this soft fix does not resolve that I'm not sure how to change, is the additional ISC steps when the AC is turned on. I'm pretty sure it's a modifier function, and as such it can only add to, not subtract from. Well adding to actually reduces RPMS in the DSM. I would need to alter it's function to a subtracting modifier for it to work for me.

Additionally, I want to be as sure as reasonably possible that changing those 4 wires is not going to damage my ECU or ISC, as neither are cheaply replaced. Although I am fairly confident no harm will come, I want to have some additional knowledgeable people put there eyes on it before I make the change.

keydiver

iTrader: (2)

True, it doesn't exist in the xml definition, but I found the adder in my hex editor at 1513h, and changed it to 64h, to match the 1G MAS compensation. The adder, scaler, and MAS compensation in that ROM are exactly what are in a 1G EPROM. The only thing I didn't know what to do with was the MAS smoothing, or whatever they call it. The 1G ECU's didn't use that.
The other thing that doesn't exist in that EVO definition is the fueltrim switch points. I'm sure they are there somewhere, or maybe they call them something different. The 1G and 2G chips each have a Hz where the fueltrims will switch, with hysteresis.