7% fuel increase upon gear shifting ... Why?

[zbrskz]

Hi,

I've recently noticed that after shifting a gear the ECU increases the fuel around 7%, this will progressively decrease to the fuel requested in the fuel maps some 0.5-0.7s after TPS reaches 100%.

Why does this happen?, of course I see the AF getting quite rich during this period. How may I deactivate this on KManager?

Thanks,

Luis

[Spunkster]

http://www.hondata.com/forum/viewtopic.php?f=12&t=3629

[zbrskz]

Thanks Spunkster,

I race an Euro 2004 Civic Type-R (EP3), PRA ECU, 2008 KPRO (can't give you the serial #s at the moment, as I am 5000Km from the car), RRC manifold, 70mm TB, 4-2-1 exhaust header, 3" exhaust line, IPS cams, 440 Precision Injectors.

I've attached my last race kal & log. You may see the problem in any WOT 2nd to 5th gear acceleration.

Regards,

Luis

[Hondata]

What type of wideband do you have? Then I can overlay the actual lambda.

1. Your low speed cam angles are set to 50 degrees everywhere from 2500 rpm and up. At low loads this will make the AF very sensitive to small fuel changes.

2. Your high speed fuel tables are a real mess. The fuel values are not consistent from column to column, nothing above column 7 is tuned, values change wildy from cam angle to cam angle. In short, it needs to be tuned again.

3. The low cam tables are better, but see will need tuning after #1 is fixed.

4. The ignition tables need tuning too.

5. Error P0341 is disabled. This usually means the cams are off by one tooth.

[zbrskz]

Hondata,

I have an Inovative LC-1. Unfortunatelly, the 0v=7.36, 5v=22.390 "spec" does not work. Same in all our 3 race cars.

You may use the following table:

1.39v = 7.35
2.38v = 12.00
2.53v = 12.60
2.63v = 13.00
2.68v = 13.20
3.04v = 14.70
4.93v = 22.39

1. Yes, I set 50º everywhere. While racing we rarely do less than 5k. That log is from a city circuit, in a race track we are always above 6.5k. Anyway, the report problem occurs after gear changes, that's 6.5-7k onwards...

2. My race weekend works like this: practice - small tune - practice - small tune - qualifying - big tune - race1 - small tune - race2. In some cases I don't even get out of the car between driving and tuning it. So, I only pay attention to areas of the maps I actually use while racing. That said, this last race was at 1800m altitude, MAP values always below 820mbar, that's why columns 9 & 10 remain untouched from sea level races and are so "far" away from columns 6&7. I use a big TB, which causes a lot less airflow for MAP readings below WOT columns, that's why the high fuel tables look strange to you. If you open the attached log and use the lambda overlay you'll understand.

3. The effect of the big TB in the low cam isn't severe as in the case of the high cam.

4. At 800-820mbar of MAP I find the car to respond well to 31º@5775-6425, 32º@6425-6725, and 30º from 7000 onwards.

5. Ups, P0341 was disabled by mistake, thanks!

If you open the log, look at for instance frame #2264, inj=8.54ms and inj. decreasing progressively to 8.00ms (frame #2270), that's from 6900 up to 6970rpm. This is 5th gear, if you go back to frame #2210 (6963rpm in 4th gear) you'll inj=7.99 as it should. Same happens in 2nd and 3th gear immediately before, and off course, you notice lambda to go richer in these areas ... why does the overfueling occur?

Thanks,

Luis

[Hondata]

It essentially is a tuning problem - you still need the non-WOT parts of the fuel tables to be correct, otherwise you will get odd fuel trims. I think that you are going about tuning the wrong way. There are huge jumps between columns 5 and 6 in the high cam tables. The fuel tables should be consistent from one cam angle to another - compare the 40 and 50 degree high cam tables. Tuning changes made by rpm should always change a whole row, so that columns 9 and 10, even if not used, are consistent with the rest of the table. Setting the low cam angles all to 50 degrees will cause problems, as the airflow at that cam angle will make the lambda very sensitive to fuel changes.

The only other thing I notice is that the injectors are slow to cut off during a gearchange. Increase the high cam overrun cutoff pressure a little and it also might pay to physically move the TPS so that the TPS calibration is closed to 0%/100%.

[zbrskz]

It essentially is a tuning problem - you still need the non-WOT parts of the fuel tables to be correct, otherwise you will get odd fuel trims. I think that you are going about tuning the wrong way. There are huge jumps between columns 5 and 6 in the high cam tables. The fuel tables should be consistent from one cam angle to another - compare the 40 and 50 degree high cam tables. Tuning changes made by rpm should always change a whole row, so that columns 9 and 10, even if not used, are consistent with the rest of the table.

I totally disagree, apart of columns 9&10. Check the lambda values through the all map, it is tuned for AFr 12.6-13.0 at WOT (high downto low ECT), and 12.6-14.7 for columns 1-6. The fuel lines result on tuning the mixture to those values, that is, I just tune the maps to make sure lambda results are as expected.

Setting the low cam angles all to 50 degrees will cause problems, as the airflow at that cam angle will make the lambda very sensitive to fuel changes.

Not really important, while racing the engine is rarely in the low cam zone. But if it get's there for any reason, there is no time spent to move to 50º whem going into the high cam.

The only other thing I notice is that the injectors are slow to cut off during a gearchange. Increase the high cam overrun cutoff pressure a little and it also might pay to physically move the TPS so that the TPS calibration is closed to 0%/100%.

Yeap, well seen ... The TPS is calibrated to 11% minimum reading, which is not good ...

Image4.jpg

My problem refers to the areas outlined in the red circles, although cam lines are not in the graph, in those areas they are "stucked" at 40º. MAP is constant (red line), injection (blue line) is higher than in the maps, and thus lambda (green line) is lower than tuned for.

Thanks Hondata,

Luis

[Hondata]

Do you see why the whole table needs to be tuned, even for a race car, and furthermore why each cam angle table should not be radically different from the prior table? I'm just pointing out what is wrong with your calibration. It's up to you to accept my advice or not.

[zbrskz]

Hondata,

I understand your reasoning, I honestly do, and above all I appreciate your (good) advice(s).

Nevertheless, it doesn't explain the "overfueling", that clearly occurs.

A different "perpective" from the datalog I posted:

frame #1949
rpm=7267
map=816mbar
cam=38º
inj=8.32
map fuel[40º@7250rpm@810mbar] = 2073
map fuel[30º@7250rpm@810mbar] = 2106 (+1.6%)

frame #1896
rpm=7234
map=813
cam=39º
inj=7.95

In frame 1949, in pretty much the same conditions than #1896, there is +4.6% fuel from the injectors. With the "bad" tuning, worst case scenario, would be a 1.6% of fuel increase going from 40 to 30º advance cam. Cam advance is basically the same, 38 vs 39º. There are no fuel trims being applied, it's working in open loop above 700mbar. Why does the ECU order +4.6% of fuel?, is some areas it gets to 7% …

Thanks,

Luis