There very well could be another fuel table for this purpose (the starting point) but it doesnt make sense to do that..

 

So the closed loop is like a gun and the open loop is like setting the scope up? Aim for 13.3 AFR(open loop) so you hit 14.7 AFR ( closed loop) ?

 

on a side note I set all my low load open loop AFR's to 13.2 and drove for about 25 miles of mixed driving with the O2 sensor disconected. Seems like I am hitting an average of 13 on my wide band 02. Since I was running lean ( positive fuel trims) I would have expected 13.2 in my map to be higher on the wideband. By this I mean 14.7 in my map caused the ecu to add fuel so therefore without trims it would be a higher AFR (leaner) without the trim helping, am I right here?

 

more observations: When I disconect my O2 sensor Ive noticed my fuel trim go to +25%. So I am assuming that is a built in fail safe of the ECU. So now I need to find a better way to force the ECU into open loop. I think the min coolant temp for closed loop is the oute I will take. Anyone know of any drawbacks to using this method?

 

Jack,

I understand what you are saying, I think ....I'm just saying that I don't agree that this is how the fuel trims work.

What I am saying is that it doesn't matter what your open loop maps are set at. Closed loop fueling cares about one thing, the narrowband o2 sensor and adjusting the fuel until the O2 sensor cylces around .5V.

You can see this, too for example, if you reset your ECU by disconnecting your battery. The fuel trims will reset until the ECU re-learns what they should be. It doesn't use your open loop map as a base map. It simply uses O2 feedback to re-learn the correct trims.

The fuel trims go out of whack not because of changes to open loop fuel maps, but because of upgraded injectors that aren't scaled properly, aftermarket airflow sensors, aftermarket intakes/filters that alter aiflow readings, etc. On a stock system, fuel trims will drift when injectors start to clog, fuel filter starts to clog, etc.

That range of adjustment is so that the ECU can maintain the cycle around .5V from the O2 sensor due to these varying conditions and age of the car. Obviously our modifications are more extreme than what was designed, but the trims are only to maintain a stoich burn at closed loop.

I know that we think we may have found a 'closed-loop' map, but I don't think we have verified this, but even if it is, it's just going to be a map with an offset for the O2 sensor. For example, instead of having the O2 sensor cycle around .5V, maybe have it cylce around .52 V or something similar. But, since narrowband sensors are gossly innacurate the further away from .5V you get, you will only have a very small range of adjustment anyway.

Everything that I have stated above is from the years of experience with the DSM Ecu, since I am new to the Evo ECU, like most of us are. But, when the Evo ECU is fully disassembled, we will know for sure.


Eric

 

Your missing the entire point.. Closed loop tries to maintain .5v, but you have to give it a place to start.. The open loop maps SHOULD be adjusted to be close to 14.7 so it requires less time for the trims to stabilize at 14.7 (or whatever is in the closed loop map)

I think your completely confusing what I'm saying.. The only value the open loop maps have with relationship to closed loop fueling, is to give the ECU a starting point so the trims don't have to spend as much time to find the desired AFR.. There **HAS** to be a starting point that the trims are "Zero", if the closed loop system has to adjust to add fuel TO that value to reach the closed loop target, then the trim adds fuel, and vice versa..

Your completely hung up on the open loop maps being unrelated.. And its possible the specific maps are unrelated, but the fuel trims operate by adding or removing a percentage of fuel from a starting point.. Where the heck do you think the starting point comes from? It makes more sense that the open loop fuel maps give the ECU a starting point for closed loop trims to operate from..

On the other hand, the entire system can be based on injector scale, MAF size, and the calibration tables.. but its **STILL** a starting point..

If you had nowhere to start from, then you'd have no way to give the car a start-up point that it can adjust from, it makes the most logical sense that it come from the open loop maps.

My car has 0% fuel trim at idle, and its -7% at cruise.. When I disconnect my O2 sensor and drive the car around in open loop.. My idle is right around 14.7 (give or take a bit) and at cruise, it runs rich at about 13.9-14.1... I don't know about you, but that seems to correlate directly with the amount the trims need to adjust to get to 14.7...

 

FWIW If it works any differently, we'll find out soon enough, most of this code is in the process of being documented.. I am going from empirical observation from my own car and a bit of logical thought, since I do have control over quite a few parameters that might otherwise not be controllable I have seen how some of this stuff reacts.

It does make sense what your saying, that in order for the car to maintain stoich, it has to keep the fuel trims where the sensor is at around .5v.... All that was ever said is there needs to be a starting point, and the fuel maps makes the most sense..

On my car, I was able to use my open loop fuel maps at idle to get my fuel trims back in check, My car was running pretty lean and it had to add fuel.. so I set the open loop map to 14.0 at idle.. This did nothing to change the closed loop AFR, but the fuel trims are now 0%.. In my case, when the car was first started, it would idle real lean until the car went to closed loop.. but altering the table fixed both...

I think that was enough data for me to be comfortable that it works.

 

Eh.. You know what.. It really isn't important as long as the tools you have to work with, work for what you need to do..

 

I think there is a MUT request that forces open loop mode.. There are a bunch of requestID's and bitmapped stuff that should correlate to diagnostic functions..

 

Yes, this is what I am saying the starting point is....the injector scale and calibration tables. Sorry I didn't make that clear in my previous post.

Then, when starting the car, it starts in open loop anyway, giving the O2 sensor sufficient time to warm up and give an accurate reading. The trims then adjust to the voltage of the sensor.

Yes, that does seem to corelate, but what cells were you landing on during your cruise and what values were in your maps? If every cell that you hit during both idle and cruise were identical, then I would agree that is very interesting.

I appreciate your input and data. That was really my original questions...where are you getting this from?

Believe me, I'm not arguing in any way with you. I'm just going off of my experiences too and how I know the Mitsubishi ECUs work. I could very well be wrong with the Evo ECU. I'm just trying to figure everything out like everyone else is. Maybe I am a little spoiled coming from DSMLink in the DSM ECU. We were able to log absolutely a ton of stuff, that we can't yet do today. Even little things like when the ECU is in open loop/closed loop. And the DSMLink creators disassembled that entire ECU and even wrote their own custom code in RAM, sort of like what is going on with us and the Evo ECU right now. Most of my knowledge comes from those guys and that forum. I just share what I know.

I hope you never take any of my questions or comments in a bad way. I'm here to learn and help other people learn what I know. I never say that I am absolutlely correct with regards to the Evo ECU. And if someone does say something that contradicts what I know to be true, I simply question them to find out more. That's all I did here. I hope you understand where I am coming from...just trying to possibly learn something that you know and I don't or vice versa.

Thanks for sharing your data....it definitely is interesting, at least without seeing your maps.

Eric

 

Understood.. Keep in mind that the DSM ECU is 64k and the Evo 8 ECU is 256k, very little of it is "Unused"

I basically went through and adjusted all of my open loop maps with the O2 sensor disabled at idle through 6000rpm, got them around 14.7 (give or take) and I now have my trims real close to 0% (under 5% where I was as high as 15% at times)

Unfortunately, I intentionally was holding off on doing this because I am replacing my blowthrough setup with a new setup, and was going to have to do this all over again, but I really needed to see how off it was.


The ECU has a bunch of bitmap requestID's that tell you if its running open loop, closed loop, the status of many different solenoids and relays that are controlled by the ECU.. its just we haven't really figured out everything yet.

The Evo8 ECU has somewhere along the lines of 50 or more tables representing configurable data.. I think only about 20-30 are defined and understood at the moment.. That doesnt include single byte or single word configuration parameters that are part of the ECU's subroutines and aren't as easy to find, there can be dozens of those..

One thing I do know.. there are portions of the code that are similar or identical to the DSM, and other portions that are completely rewritten from scratch.

Being that there are stringent OBD-II guidelines, and emissions laws, both in the United States and the world in general.. Things like this and the ability to have a level of control over them like we are finding, are very real and important necessities.

Engineering is a very logical process, including software engineering (for the ECU) and being a former engineer myself, if I was working on this stuff, I would have done EXACTLY what I have been describing, which is baselining the tune with open loop maps, then allowing the fuel trims, octane trim, etc, all to adjust from that as a starting point. If I was limited in space, and didn't have room to add the flexibility I needed, then I would make compromises, such as a simple closed loop table consisting of AFR voltages targets by RPM, and TPS crossover thresholds.. But every day I see more and more clever things in the ECU that make it better than any standalone, but think about the learning curve, if you were given a standalone, and some basic information, but only a serial connection and some rough information on the communication protocol, someone figured out how to read its rom, and write it back, but you have nothing else, PLUS you have to write the code to log data, and your not even sure what data you have to work with, you can see what we're up against..

I'm almost 100% certain that there will be a completely user written rom to give you all the standalone functions that anyone could ever need.. But that takes time..

 

What a coincidence. I have an Engineering degree myself from Purdue, so I am no stranger to engineering or logical thought.

It's pretty funny, though....I only used my degree for a short while...now I work in IT for IBM. Go figure. My engineering degree is in Materials Science Engineering. I got to work on failure analyses and the sort. It's definitely an interesting field.

On to the ECU....

I guess we will find out for sure when the Evo ECU is completely disassembled. I just don't completely agree with your logic about using the open loop maps as a starting point. I do agree that it is one way to approach the situation. I think using injector scale and latency, along with Maf calibration tables are a much better starting point. I mean if you used the open loop maps as a starting point, or 0 point, then you would never be able to get your trims to 0 if your open loop tune wasn't a stoich tune at a particular RPM/load cell.

But, as you mentioned, your data seems to point that the open-loop fueling maps may be the starting point. I'm eager to see the code in the ECU. I don't care either way....I just want to know for sure. If the open loop maps are the starting point, maybe I will buy you a beer.


Eric

 

Small world, I worked for IBM for awhile.. offtopic but I had a problem with their intellectual property policy with regard to developments you make OUTSIDE of the company while you work for them.

Yep, we'll find out soon enough..

But if you look at the maps on a stock ECU, the closed loop regions are more or less tuned to 14.7 in the open loop maps.. The only exceptions are over 5500rpm and under 500rpm, both regions seem to drop out of closed loop control when their reached..

 

I remember reading that, too, and got a laugh out of it. I don't really do development work like you, so I new, at least in my case, it would have little bearing on my situation. Actually, though, it wasn't even my choice to work for IBM...it was part of an outsourcing from my original company about a year ago.


I was thinking the exact same thing and this is a very good point. It definitely lends itself towards your logic as using the open loop maps as a starting point. But we would first need enough data to see how far closed loop can actually extend into the load/rpm cells. There are switchover points based on load/throttle, etc, but since we currently can't log when the ECU switches from closed to open loop, we aren't sure. If you we could log it, we could see how far the closed loop conditions can actually extend out.

I would just think that the Mitsu engineers would have kept the open loops completely separate so that, if needed, they can be adjusted without messing with the fuel trims.


Eric

 

Yeah thats been my entire point for the most part..
Small world again, I was part of an outsourcing of another company.. I did dispute the contract and had my contract amended to exclude developments since I wasn't hired for that purpose.. At the end of our outsourcing, I ended up staying with the company we were outsourced from.. I really hate corporate life and the BS I deal with.. I'm trying to get myself more deeply into doing this stuff professionally.