RomRaider

I have been using "primary open loop map enrichment" in my data logs for a long time, but recently added "final fueling base" out of curiosity. (Partly because with the new fast polling option, extra parameters are free!)

To my surprise, it appears that what I/we have been calling the "fmic rich dip" might actually be caused by ECU logic. I've attached an image from a fragment of a log that included a shift from 2nd to 3rd gear, just as I got back to full throttle.

Notice that the fuel table is calling for low 11s, but the final fueling base parameter dips down to almost 10:1. The measured AFR tracks the final fueling base fairly closely, with the exception of one data point that I'm assuming is just noise.

I've been pondering the fueling code for the last few days, and tried logging about 10 different experimental parameters that seemed like they might shed some light on this, but those parameters produce nothing but zeros in my logs. I am going to continue looking into this, though, because I really, really want to make this problem go away.

I was just about to get a blow-through charge pipe welded up to try to fix the rich dip, but now I'm not sure that's going to help.

I've also been wondering about this - running a TMIC - but since I was doing CAN logging, I had both open loop enrichment and final fueling base logged, and very often they did not correlate - by quite a margin. I never analysed the data to see whether there's a pattern - but I'm curious now.

I was going to try a hybrid-BOV and send only 50% or so of the contained volume of air back to the pre-turbo intake. I've been thinking about a blow-through setup to fix this if a cheaper hybrid BOV didn't do anything to help.

I always chalked it up to the lag between the pre-turbo MAF, the injectors, tip-in, and the O2 sensor. The ECU just isn't aware of the extra volume of the FMIC and piping nor are the factory-set tables that are used by the ECU to 'predict' what will happen.

My thought is like this, when going from cruise into WOT:
You press the throttle, ECU adds tip in, but since there is a bunch of extra piping the air doesn't get there fast enough (starts rich dip), but when it does get there (just a tad late) the MAF is just starting to read the piping filling up so that pushes it a further into rich, but after a few engine cycles, things sort themselves out.

When shifting:
You are WOT, you let up to shift and your recirculating BOV releases the air into the pre-turbo tract. Now you have more volume of compressed air (compared to stock) that is being dumped into the pre-turbo intake pipe that has no where to go. The turbo is in recirculation at that point, yes, but there is more volume of air at the non-boost pressure than there was at the on-boost pressure. A bit of that extra volume blows back through the MAF causing its reading to go bonkers. When you get back on the throttle it takes a few cycles to get things normalized again. Then you add on to that the tip-in that happened during all that, and the extra air it takes to 'fill' the piping to boost again...

You are compressing a much larger volume with air with a FMIC. Much like going from a tiny 2-5 gal compressor to a 60 gal compressor. Lets say the tank is at 135psi on both units. Now you disconnect the power to both compressors. You can still use the 60 gal for quite a while longer than the 2 gal one, even though they were both at 135psi.

Similar principle with the TMIC vs FMIC except that extra air has only got one way to leave the system (with a full recirculating BOV), the same way it came in (its not getting used by the engine with the throttle closed).

Then when the FMIC is at a lower than max psi, you have more volume to refill to operating pressure (which from the MAF's point of view, means more air is going into the engine, when a tiny part of it is actually going to fill the extra volume of the FMIC).

If you don't have a recirculating BOV, you are seeing the MAF sensor read the turbo pulling air through it without that air going to the engine. Which also causes rich spikes.



I do wish there is something other than tip-in that is undiscovered. Because I've been tweaking tip-in till the cows come home and it just doesn't cover a varied enough range of driving for me.

I used to believe it was due to pressurizing the intercooler as well... but after seeing Final Fueling Base dip into the low 10s, I'm certain now that the ECU is a big part of the problem. I'm determined to figure it out.

I still think there is merit to the "MAF lag" theory, since the measured AFR sometimes dips richer than the Final Fueling Base. But, I am also very sure that the problem will get much better if I can get the ECU to quit contributing to the rich dip.

Maybe there is something tied to boost error? It would be interesting to see MRP and boost error plotted on your graph. I know my fueling isn't near to being on till my boost error is under control. But I always chalked that up to MAF lag and not something the ECU was doing. I spent quite a while with the boost tables getting them closer to reality and that helped with transitional fueling issues.

I don't think it's boost error, because that goes from negative to positive while Final Fueling Base stays extra-rich. See below.

I just realized why all of my experimental logger parameters came up zero - I had them set up with length="1" and it really needs to be length="4". I copy-pasted that mistake into all of my new parameters, so all of them came out wrong. I will log a couple of pulls later tonight and see if any of these new parameters shed any light on the rich dip. I'm optimistic...

The good news is, I found an enrichment parameter that explains the difference between the fuel table and the "final fueling base" parameter. So now I need to try to find the tables, or logic, that governs this parameter.

The bad news is, I don't think I'm going to have an opportunity to investigate further until after the weekend... but I'm pretty excited about this. The rich dip is the most annoying thing about my car, and all this time I thought I'd have to go blow-through to fix it. And I was just about to get another pipe welded up after the first one proved too turbulent. And it looks like it's a software problem after all!

As a bonus, this same parameter also appears to perform both tip-in and tip-OUT, in addition to the rich-dip thing. So maybe I'll be able to shed some light on those as well. The tip-in tables that Merchgod documented are related to "an additional and separate firing of the injectors" (according to his help-text) but the tip-in that comes from this parameter is just something get gets added to the regular injector pulses, for a duration of about 300 milliseconds, after you stomp on the throttle. And then a few seconds later it kicks in again, causing the big rich dip.

If anyone running A2WC522N wants to look into this, add this to your logger.xml:

<ecuparam id="E906" name="NSFW Enrichment Comp 2" desc="TBD.">
<ecu id="2F12785606">
<address length="4">0xFF5614</address>
</ecu>
<conversions>
<conversion units="Lambda enrichment" storagetype="float" expr="x" format="0.00" />
</conversions>
</ecuparam>

If anyone else wants to look this up in additional ROMs, look for the function that reads these already-documented parameters...

AF Learning
AF Correction
Primary Enrichment Final
Final Fueling Base

...and does a bunch of adding (of undocumented enrichment values, including this one), and then repeats some additional math four times (setting values for each injector, I suspect). I've attached the A2WC522N assembly code that I'm talking about. NsfwEnrichmentComp2 is the parameter in question.

It will probably take some investigation via experimental logging parameters to figure out which of the undocumented enrichment values is "the one" in other ROMs, but hopefully once I find the tables/logic that calculate this value, that will make this easier.

Last Calculated PW is used to lookup per injector comp values in the routine at 0x30128 before getting to the function you have defined where these compensations are added and stored on the stack for further manipulation.
0xFFFF5990 = Per_Injector_Primary_Fuel_Offset_Compensation_A
0xFFFF5994 = Per_Injector_Primary_Fuel_Offset_Compensation_B
0xFFFF5998 = Per_Injector_Primary_Fuel_Offset_Compensation_C
0xFFFF599C = Per_Injector_Primary_Fuel_Offset_Compensation_D

Your Comp2 value is also converted to a 2Byte value in the routine 0x24D88 that makes all the engineering 2B/1B extended parameters.

AWESOME! Great work guys!

viewtopic.php?f=15&t=1735&start=15
Bottom of the page Would this be relevant to what you are trying to 'fix' from a tuning point of view?

I am not seeing any FMIC Rich Dip after playing with my tip-in enrichment settings, it involved a lot of rescaling to the point where the throttle can be either squeezed down gently or cracked to full throttle with 0.15 AFR difference between the 2.

In the past I have completely disabled tip-in, and the rich dip was unchanged. The tip-in table (at least for my car) only affects AFR for about 300 milliseconds after an abrupt throttle change. The rich dip happens about 1500 milliseconds after.

What year/model/market of car do you have? And what kind of intercooler? Can you post a log showing you stomping on the throttle at 2500 RPM, and another at 4000 RPM? Log these parameters:

* RPM
* Throttle
* Primary Open Loop Enrichment
* Final Fueling Base
* Wideband AFR
* Boost
* Load

I had some time to study the ROM last night. I found some tables which, if filled with 0, will probably make the "enrichment comp 2" parameter stay at zero. However I don't know what the side-effects would be, so I am still investigating...

Tangential discovery:

<ecuparam id="E920" name="NSFW Timer Ticks" desc="Timer ticks, approx 135/second, stops at 65535.">
<ecu id="2F12785606">
<address length="2">0xFF52E4</address>
</ecu>
<conversions>
<conversion units="raw" expr="x" format="0" />
</conversions>
</ecuparam>

This counter is zero when you start the engine, counts upward at about 135 ticks per second, and stops at 65535 (since it's a 16-bit unsigned integer). There are a couple of checks of this timer, and the "enrichment compensation 2" logic is skipped if the tick count is less than 625 (4.6 seconds).

Not particularly useful, but there it is anyway. There are a ton of places in the code that reference the timer value, probably for warmup-related stuff I guess.

Will log it next weekend when the cars are a little closer to me

Have seen it on both an 02 STi and an 06 WRX, both AUDM with generic cheap FMIC kits

My other consideration is that this could be an emissions related fuel additive?

Thanks. I only have logs of this kind of thing from my own car, so I'm curious how other turbo Subarus behave under similar conditions and with other peoples' tuning tweaks.

I figure that most of the non-obvious things the ECU does fall into one of four categories:

* less emissions
* less fuel consumption
* better driveability
* more power

I don't see how dumping in lots of fuel during spoolup would help with any of those, but I don't think it does anything for driveability, and it's definitely not helping with mileage or power.