Fuel Map, Delta Compare and +172 -> -184

modulation · Post by **modulation** » Fri Jan 16, 2009 10:21 am

ka24de s13 ecu, stock injectors (~280CC I think.)

When I have a "raw value" of 127 and increase that by one, the whole "raw value" map view turns from green/red to brown..

When I do this in the delta view, the value goes from whatever the actual delta compare is to -184...

A few question.
#1. Why does it do this?

#2. I'm NA, targeting a specific A/F ratio of 12.9, on one portion of the map (where I would expect max VE,) the raw value for my fuel injector pulsewidth is 127, but my AFR's are still in the mid 13's, and if I throw alot of fuel at it, they barely go down. I am sure that all the cells around it are at a very high value to (so it's not just averaging out a hi/low value.)
Am I missing something here, or should I really be throwing that much fuel at a NA car?

Legionnaire · Post by **Legionnaire** » Fri Jan 16, 2009 11:49 am

I think I can explain. Inside ECU there is only 1 byte of info that stores each value in your fuel map. 1 byte = 8 bit = 8 x [0,1] = 2^8 = 256 different values, from 0 to 255. this range has (due to ecu algorithms) to include
- stoich fuel point
- richer side of the fuel scale
- leaner side of the fuel scale
- an indicator whether this exact point should be worked with in closed loop or open loop mode

The latter is the most important one. Basically, it means that this byte includes two different scales - one for closed loop and one for open loop and can be interpreted in both ways. High-order bit in this byte acts as this very indicator for closed loop mode. Each mode has its own base value that represents 14.7AFR. If HOB =0 (open loop), then the value representing 14.7AFR is 0raw; if HOB equals 1 (closed loop), then 14.7AFR matches the value, where next order bit (sixth) is also 1 and all others are 0, i.e. 11000000bin - in the middle of a new scale with 128 values of 128-255raw range 128 + (128/2) = 192dec raw value. If you are not really into this mathematical-computer stuff, just ignore this paragraph

So values behave like this:
raw values --- AF ratios
0-127 --- stoich to ~7.38 open loop
128-255 --- ~29.4 to ~9.85closed loop

In an open loop mode stoich base value (value where your AFR is stoichiometric) is 0raw. 0raw equals 14.7AFR and gets richer till 127raw value, that equals 7.38AFR. So this means that you can not run leaner than stoich in an open-loop mode. I guess this has been done that way to avoid possible engine damage - just one more line of defence against lean-state overheating etc.

In a closed-loop (128-255 raw values) mode stoich base value is exactly in the middle of aforementioned 128 value range - 192raw = 14.7AFR. 255raw = 9.85AFR and 128raw = 29.4AFR.

Note that this range of AFRs is wider than in open-loop mode, but both modes are represented by 128 different values. This means that altough open-loop mode is stoich or rich only, it has finer resolution than the closed-loop AFR scale, so you can adjust your AFRs more precisely under load.

Cell painting in nistune relies on these principles + on the difference between the most rich and the most lean spots. So by making one of the values 128raw you suddenly introduce a 29.4AFR cell, which obviously makes all the other cells look pig rich to nistune painting algorithm, hence a lot of brown color. Hope that answers your #1.
Don't know about delta compare, i don't use it, tbh.

Not sure that I understand the second question. You can not reach 12.9AFR? Have you tried values like 20-25 raw?

MichaS14a · Post by **MichaS14a** » Fri Jan 16, 2009 11:51 am

If the raw value is greater than 128 then the ECU uses the lambda-signal. You can see this if you tap "flags" in the fuel-map. Then the areas with lambda-control will be marked in a different color. No surprise: These entries are all over 128.

The formular to calculate the target lambda is:

if DATA > 128 then MULTIPLYER = (DATA - 64 / 128)
else
MULTIPLYER = (DATA + 128 / 128)

I have this informations from this thread:
http://forum.bikirom.com/viewtopic.php? ... &sk=t&sd=a

I have tested the formular with a few entries of a original fuel map from a SR20DET and it matches. Please try if the fromular is okay with your map and ECU too.

EDIT: @Legionnaire was faster and has explained it better.

modulation · Post by **modulation** » Sat Jan 17, 2009 9:54 am

Thanks guys. I forgot about the whole map cell values >128 mean that the ECU should run in closed loop. Those were two awesome explanations.

Legionnaire wrote: Not sure that I understand the second question. You can not reach 12.9AFR? Have you tried values like 20-25 raw?

I can't reach 12.9:1 AFR in a few specific cells (~3200-4000 rpms under ~75-100% load) on the map.
I'm NA, and started out with a stock tune. I have a LC-1 WBo2. I've slowly incremented fuel, and I now have 12.9:1 AFR's in almost all cells under load but those where I would expect my max VE (again ~3200-4000 rpms under ~75-100% load.) I've added more and more fuel at this and AFR's have gone down but only to ~13.5:1. I've now added so much fuel in those cells that I'm at 127 and if I increase it anymore it'll be in close loop (cell value >127 raw.) I use to think that I needed to mess with the accel inrich table but I don't think that is the case and I don't know what else I could be over looking that I'm throwing so much fuel at the engine but still getting higher AFR's then I'd like..

(I'm not going by the fuel map afr's, I'm going by what my wide-band reads, and the values populated into the wideband input tracer) I know the wideband works too since in other cells and while in close loop it reads fine. I'm also filtering out wideband input tracer values that aren't under WOT (4.3V TPS value.)

Legionnaire · Post by **Legionnaire** » Sat Jan 17, 2009 10:49 am

Hmm, that is really strange situation you've got there. The first thing that comes to mind when i read that your actual AFR ceases to follow your map value is that you have a mechanical problem somewhere. What is your actual injector duty cycle at those load/rpm cells? Have you tried higher fuel pressures? other fuel pump? Check for vacuum leaks. Oh, and what is the least value that you've tried to enter in your area of interest?

Sorry, that is not of much help, I know.

MichaS14a · Post by **MichaS14a** » Sat Jan 17, 2009 11:21 am

I have load the adress-file and the bin for the s13-ka24de.
The bin contains different values in the normal fuel maps and knock fuel map. Are you sure your engine runs on the normal fuel map rather than on the knock fuel map in the range you wrote (3200-4000 upm)?
If your engine runs on the knock map than you can change what you want on the main fuel map without getting any results.

modulation · Post by **modulation** » Sat Jan 17, 2009 1:22 pm

MichaS14a wrote:I have load the adress-file and the bin for the s13-ka24de.
The bin contains different values in the normal fuel maps and knock fuel map. Are you sure your engine runs on the normal fuel map rather than on the knock fuel map in the range you wrote (3200-4000 upm)?
If your engine runs on the knock map than you can change what you want on the main fuel map without getting any results.

Yep I'm always sure I click "knock copy" right before I save so my knock/regular maps are the same.

modulation · Post by **modulation** » Sat Jan 17, 2009 1:28 pm

Legionnaire wrote:Hmm, that is really strange situation you've got there. The first thing that comes to mind when i read that your actual AFR ceases to follow your map value is that you have a mechanical problem somewhere. What is your actual injector duty cycle at those load/rpm cells? Have you tried higher fuel pressures? other fuel pump? Check for vacuum leaks. Oh, and what is the least value that you've tried to enter in your area of interest?

Sorry, that is not of much help, I know.

Since I'm NA I doubt I'm hitting the limits of the fuel pump. Especially since I'm not running lean at higher rpms. My injector duty cycles aren't over 75% anywhere in the area I'm having the problem on the map.

The lowest values I've had in there was the stock map, I increased fuel in those cells slowly, and AFR's have dropped but not by as much as the rest of the map.

chris2712au · Post by **chris2712au** » Sat Jan 17, 2009 5:45 pm

you will have to do a map and pay attention to your duty cycle on the injectors. It cant suddenly go lean if you have all the same values around this lean spot.. maybe a little but not much..

check your load scales and watch your TP as well to make sure you are still on the map.. Also check that the engine isnt missing as it kind of stuffs up AFR readings..

Jumping to knock maps can be a problem so do a knock copy and an ECU sync just to be sure that value is in there.. try giving the surrounding cells a little more as it is averaging.. other than that you could bump the K value a little.. normally that does not need to be touched though.. also check your AFM ? how is the voltage ? is it standard ?

It would not hurt you to connect a fuel pressure gauge.. not saying that the pump could be stuffed but sometimes the in tank hoses and pump connections break / come loose.. and maybe the reg is stuffed..

modulation · Post by **modulation** » Tue Feb 03, 2009 11:21 am

chris2712au wrote: check your load scales and watch your TP as well to make sure you are still on the map..

I thought of this possibility, but from what I saw with my logs I was getting close to the max TP on the load scale (last column) but never on it.

However since you also mentioned it, I said F' it and decided to just change the last couple of numbers in the load-scale table.

Sure enough I'm still hitting the next to last column (not the last one,) but my AFR's are now what they should be and I've had to remove a ton of fuel.

Thanks for the awesome advice. I wonder why the logs wouldn't show me hitting the max TP loadscale number (it was 108, logs showed me at ~104.)

PL · Post by PL » Tue Feb 03, 2009 1:22 pm

I'm coming in a bit late here but check your TTPmax. Theoretically it shouldn't be a problem for you, but the symptoms certainly sound like it is.

I got caught by this recently. Doing an S13 SR20DE+T. Couldn't get it rich enough on full boost by adjusting values in fuel maps. Increasing K would work but that messed everything else up. Very frustrating.

Increased TTPmax and suddenly the injector pulsewidth would respond to values in fuel maps again...

PL