Check my original post just above - I have edited it to include the complete definitions. Let me know what you think of the 3 maps.

 

Well after looking at the locations, first off they arent the same in the USDM ROM's and second the tables arent the same size.

In the US rom the fuel table is 20x15 and the spark is 21x20 (as are these maps). There are only 2 fuel maps in the US as well. Maybe JC or MJ have more data but to me it looks like they are ignition maps of some sort. Can you post a screen shot of one of these definitions?

 

JohnBradley - I believe there are 3 locations for the USDM evo IXs: 3AA9/5883/5D93...20x15 (need to validate)

 

This is the closed loop map. 1, 2, 3 all have the same data. It is scaled using "AFR".



And this is the "Ignition Map 4" map that I located. It is scaled using "Timing". It may well not even be an ignition map, but I'm not sure. However, it is very, very similar to the other ignition maps apart from the massive hole at low rpm/load.

 

Table one looks like the enlean table in the US VIII's to me. Table 2 is definitely very interesting. Ours have 38* of timing in the 0-10 load cell 2500+ whereas that has 45. That timing hole almost coincides with the antilag tables I have seen. The upper rpm/high load areas of the map are aggressive compared to other factory (especially US-13) but then tapers from aggressive to mild with the 15* and 12* areas. Interesting. If I get some time to work on this tomorrow I will check ST's addresses and see if I find something similar in the US ROMs or that at least matches what you have Brendan. Unless someone has already done this and I missed the thread/post.

 

Call it whatever you want but it directly affects closed loop afr.

 

That looks a lot like I would expect a timing map with antilag to look like.
I would of expected the low timing numbers to extend up the load scale a bit more though, and maybe a little higher on the rpm side too, maybe the axis are not correct?? but the general shape of the table looks right to me.

 

Okay so its not a question of if it does or not, but on what level? Does it have all the discrepancies between target AFR and actual AFR like the open loop tables? Can you elaborate on your experiences with this table?

 

Basically I took the 80 and 90 cells to 14.7 and slightly lowered the 70 load cells anything more than .1 or .2 at 70 or below you will be running a 10.5:1 afr when you stomp the gas at low rpm. Like 2500rpm 50% throttle in closed loop it's not enough throttle to bring in open loop at that low of an rpm I had to tweak this a few times to get it right.

 

any updates on this? I'm struggling with a part throttle lean condition when the car is above the minimum coolant temp for closed loop but not fully warmed up. When totally warm, it goes away. Sounds like others are experiencing the same thing.

 

Have to adjusted the actual open loop maps (Low Octane map) for this?

There is some weirdness when driving the car when its not totally warmed up, sometimes the turbo is "Sluggish" from the cold oil which will slow airflow into the engine, causing stumbling, sometimes its high engine vacuum allowing the DV to vent back into the intake, other times its just the fuel maps are a bit lean in a few areas due to modifications.

I think there was some speculation that the low octane maps are where to make adjustments for cold start open loop (which very few people actually adjust) which makes it hard to find the culprit. So I would take a look there.

 

FWIW, I can't imagine how there wouldn't be discrepancies. The only point a narrow band sensor can find well is stoich (approx 14.7:1). For emissions control purposes, this is the point that the ECU wants to be. Two of the three things a cat removes are low, and the third has not yet peaked. Also, the gas is hot which the cat requires, and since stoich is a singularity for the sensor the ecu tends to oscillate back and forth (just rich, just lean) around it. This gives the cat a chance to collect O2.

I could see how there might be an closed loop fudge factor. But, with only narrow band sensors it seems like it would still have to be a best guess - like open loop.

-jjf

 

I've made no adjustments to the low octane fuel map. That's an interesting angle, the first I've heard of the low octane map in control of cold start conditions.

But my issue is in closed loop, before you reach operating temperature. I can force the car to run in open loop by adjusting "mininum temperature before closed loop" up to say 70C, and adjusting those low load cells a bit richer, and it runs great when cold (when I say cold, it's still about 45-50F) while using the open loop maps. But then it throws the PO125 code. I hate to make changes to the open loop map like this, because the car runs perfect once warm, so 99% of the time, those low load fueling cells will be too rich. When cold, below a coolant temp of say 60C, I see a/f's under light load between all over the place and as high as 17:1 to 20:1. But once you reach operating tempetures, everything lines out and controls well at 14.7:1.

It's like there needs to be some sort of coolant temperature dependent enrichment map. That would be ideal. It's very frustrating.

 

Once its warm, its in closed loop, so you should always be relying on the AFR target map (which we're still not sure where it is BTW) So the only real purpose of those map cells that would normally be hidden by closed loop, would be for open loop operation, either due to a problem forcing open loop, or when the ECU has not gone into closed loop.

I would definitely look at making adjustments to the maps as once the car is in closed loop, the maps are usually just an "offset" for which the closed loop targets work from.

Its really just a balance, the rich code is pretty common when its running richer than necessary according to the ECU's fuel target adjustment range, if it falls outside of the ability to adjust, you get that code, of course again, we're not 100% sure where/what the closed loop fuel target map is (We thought we knew, but there is too much uncertainty as my results were encouraging, among a few others, and several others had no affect)

 

Oh and I think the closed loop target values are in .10v increments.. As noted in an earlier post, there's very little point to try to fudge it outside of stoich as it will have very little effect due to the nature of how the narrowband sensor works, and I'm sure its based on voltage as thats a concrete value that the ECU may try to hover around.. I don't know how much control you ultimately end up with anyway even if we did find a usable target table, as your best option would be to use a configurable narrowband output from a wideband setup to fool the ECU instead..

What would be cool is to use a wideband signal to replace the narrowband and rewrite some of the code for this stuff, but that won't happen until the ECU's rom is very well documented, and widely distributed efforts start happening.