6 tuning questions - AP2 S2k (o2 CEL, s.trim, etc)

[jun1or]

Greetings,

After a trip on the dyno and a week of self-tuning, I have a list of questions. I'm a 2 year user of s300 with 25+ NA tunes, but new to flashpro for S2000.

I also know how to post properly, hopefully inducing replies by spunkster and or hondata.

Serial number of flashpro: FP-S2K-US-30023
FlashPro Manager version: 1.1.8.5
Calibration: Attached
Datalogs: Attached
Vehicle: 2006 Honda S2000
ECU: Stock PZX
Mods: Airbox lid off (don't flame, tuned this way), otherwise bone stock.

1.) It is my understanding that there is no need to to force open loop, and that s.trim and l.trim are compensated during tuning. I've monitored while livetuning and it appears that when l.trim is <>0, it does make an influence in lambda overlay. Maybe it's just my imagination. So... is the first sentance in this question true?

2.) Is there a way to clear lambda overlay like in sManager (control+L)? I loved this feature and it's gone. When I live tune the fuel maps, I want to see the changes without stopping recording and starting again (thus generating a large number of autosaved datalogs - a cool new feature nonetheless)

3.) Is there a way to show target lambda overlay values like in sManager (Shift + F4)? If I add or change load indexes, I have to flip back several times to settings to see what I'm actually tuning to (target lambda).

4.) Are there *any* differences between a "stock" uploaded flashpro config and a true oem untouched ecu? I thought I read that knock sensor sensitivity was reduced, to aide in tuning, even in the claimed "stock" configuration in FlashPro Manager.

5.) After 40k miles and no check engine lights in a bone stock 2006 S2000, after starting with a "stock" config with only two changes (ELD disabled and wideband input configured on ELD), I now have a DTC/check engine of P2A00 (air fuel sensor [sensor 1] circuit range/performance problem). Coincidence that this happened on day 1 of using Hondata? I won't rule it out since it was a service bulletin. So...does the "stock" flash take into factor the o2 heater delay described in the bulletin? Any other comments? DTC always comes on while jake braking (engine braking) on the highway offramps. Datalog attached "06-21-11 P2A00 DTC at 25-30sec.fpdl"
http://ww2.justanswer.de/uploads/hybrid ... _p1172.pdf

6.) I noticed +10 s.trim on highway cruising and no l.trim adjustment, coupled with a steady 13.4 ish AFR. This seems strange to me but could be normal. I'm used to near 14.5 bouncing in closed loop with s300 in my CRX. Is this normal? Datalog attached "06-23-11 Large s.trim on highway, no l.trim.fpdl"


Thanks in advance,

Mike

[jun1or]

Wow, that is one long post. Grab a beer!


I lied. Question #7.)

Is it normal that my AF.corr and Wideband are .5 AFR off? If so, my AF.corr seems dangerously lean at peak torque, 6800rpm. See attached picture and datalog.

WB = AEM 30-5130B, 0.77v offset.

Thanks again,

Mike

[Hondata]

1. Long term fuel trim is not used in open loop.
2. Not currently.
3. No. but we intend to change the target table from three values at different loads to a full sized table.
4. The 'return to stock' calibration is identical to the stock ECU calibration. The 'stock equivalent' is very close - the tables are all the same, but the ECU code has features like boost control added. In our testing the dyno between a return to stock and stock equivalent was identical.
5. We use the most recent version of ECU firmware (calibration) from Honda. It appears that Honda modified the parameters for the o2 sensor operation between early and later firmware versions, possibly as a result of the service bulletin, and it seems that the more recent firmware throws the P2A00 error more easily. Since we use the latest firmware, it should have the moisture protection as described in the service bulletin. You get the DTC upon overrun because this is when the ECU tests the o2 sensor.
6. No, 14.5:1 would be normal. From the datalog the ECU does not appear to be running in closed loop at all. It could be from the P2A00 error, but first I would suggest testing with the 'return to stock' calibration to see if the AF is closed to 14.5:1
7. If you have any offset at all, then the datalogging wideband reading is not accurate. Initially I did not include the option to enter a voltage offset in the software but people forced me to add this. You need to keep the wideband controller ground wiring as short as possible, and use the same ground point as the ECU, which is on the back of the engine. In some cases tapping into the ECU ground wiring close to the ECU is even better, as the potential ground loop length is shorter still.

[jun1or]

1. Long term fuel trim is not used in open loop.
2. Not currently.
3. No. but we intend to change the target table from three values at different loads to a full sized table.
4. The 'return to stock' calibration is identical to the stock ECU calibration. The 'stock equivalent' is very close - the tables are all the same, but the ECU code has features like boost control added. In our testing the dyno between a return to stock and stock equivalent was identical.
5. We use the most recent version of ECU firmware (calibration) from Honda. It appears that Honda modified the parameters for the o2 sensor operation between early and later firmware versions, possibly as a result of the service bulletin, and it seems that the more recent firmware throws the P2A00 error more easily. Since we use the latest firmware, it should have the moisture protection as described in the service bulletin. You get the DTC upon overrun because this is when the ECU tests the o2 sensor.
6. No, 14.5:1 would be normal. From the datalog the ECU does not appear to be running in closed loop at all. It could be from the P2A00 error, but first I would suggest testing with the 'return to stock' calibration to see if the AF is closed to 14.5:1
7. If you have any offset at all, then the datalogging wideband reading is not accurate. Initially I did not include the option to enter a voltage offset in the software but people forced me to add this. You need to keep the wideband controller ground wiring as short as possible, and use the same ground point as the ECU, which is on the back of the engine. In some cases tapping into the ECU ground wiring close to the ECU is even better, as the potential ground loop length is shorter still.

Thanks. Really good info here.

Regarding the P2A00 error(#6 & #7 above), I will first test with "return to stock" and also purchase new primary o2 and keep this thread updated.

Followup questions:

a.) Shall I add the two features mentioned to the feature request thread? (clear lambda overlay & display target lambda table)
b.) Regarding .77v offset and (related?) wideband innaccuracy, I did tap into sensor ground at ECU, pin E4, for accuracy. Other suggestions? I figured that would be most accurate. Perhaps I should just choose chassis ground near ecu?

Thanks,

Mike



EDIT: Just realized, hours later, that if I chose the wrong pin for ground reference on ECU (pin E4) that could maybe trigger the DTC and/or cause AEM WB innacuracy.

[jun1or]

6. No, 14.5:1 would be normal. From the datalog the ECU does not appear to be running in closed loop at all.

Wait which datalog are you looking at?

The picture is a full throttle run, so it shouldn't be closed loop, but the datalog attached above titled "06-23-11 Large s.trim on highway, no l.trim.fpdl" should be closed loop. You're saying it's not?

[Hondata]

E4 on the S2000 side or RSX side? I've found a difference in the ECU between the signal grounds, logic grounds and power grounds, and still recommend the main ground at the back of the engine.

Remove the analog input, then wire a 1.5V battery across the input pins (ELD and the ground). Measure voltage with a voltage meter and check that the KManager datalogged voltage matches. If that works, test with the wideband - some early widebands don't have enough output current to overcome the input resistance of the ELD input (and it has a fairly high resistance too).

[jun1or]

E4 on the S2000 side or RSX side? I've found a difference in the ECU between the signal grounds, logic grounds and power grounds, and still recommend the main ground at the back of the engine.

Remove the analog input, then wire a 1.5V battery across the input pins (ELD and the ground). Measure voltage with a voltage meter and check that the KManager datalogged voltage matches. If that works, test with the wideband - some early widebands don't have enough output current to overcome the input resistance of the ELD input (and it has a fairly high resistance too).

RSX side? This is a flashpro.

[jun1or]

I moved the AEM ground to chassis ground near the ecu, no difference - the AEM shows about .5 AFR rich as compared to the stock corrected sensor. Dang.

[jun1or]

6. No, 14.5:1 would be normal. From the datalog the ECU does not appear to be running in closed loop at all. It could be from the P2A00 error, but first I would suggest testing with the 'return to stock' calibration to see if the AF is closed to 14.5:1
7. If you have any offset at all, then the datalogging wideband reading is not accurate. Initially I did not include the option to enter a voltage offset in the software but people forced me to add this. You need to keep the wideband controller ground wiring as short as possible, and use the same ground point as the ECU, which is on the back of the engine. In some cases tapping into the ECU ground wiring close to the ECU is even better, as the potential ground loop length is shorter still.

I performed "return to stock" and still have P2A00 (air fuel sensor [sensor 1] circuit range/performance problem), so I will install a new primary o2.

I do have what appears to be normal closed loop operation as well, but I still have an offset in o2 sensors. The stock sensor AFR.c shows 0.4 to 0.5 points leaner than my aftermarket AEM wideband, even after moving ground location. I'm contemplating trying another input on the ECU at this point. Open for suggestions either way.

I'd be amazed if a new o2 sensor magically fixes this offset, but I'll try regardless.

[jun1or]

Update to this thread:

SOLVED DTC issue - I've purchased and installed a new primary o2, no CEL/DTC P2A00 anymore.

SOLVED WB offset, kind of - I've replaced the AEM wideband gauge and moved the ground location as well as replaced the defective oem primary o2 as mentioned above. Now the difference between factory wideband and aftermarket wideband is less, but still there (.3AFR difference comparing AF.Corr and WIDE <datalog attached> versus .5AFR prior). Which should I trust? Anyone else have this offset? Which sensor do you tune with?

ISSUE - I still have a problem with highway cruising. I am either not entering closed loop (only on highway) or something else is wrong with my cal. I'm seeing 13.5 AFR at 60+ mph speeds. I am getting proper 14.4-14.8 AFR bouncing in low load city cruising. I have attached a new datalog of highway cruising, along with my full calibration. When I use return to stock feature, I get expected 14.4-14.8 AFR's on highway. Comments welcome.

Mike

[Hondata]

Thanks for giving feedback for the problems you have solved - this helps other people with the same problems.

Start with the stock equivalent calibration, check behavior, then import one table at a time until you narrow the problem down. Also try running the calibration with live tuning disabled in case that is the cause.

[jun1or]

Thanks for giving feedback for the problems you have solved - this helps other people with the same problems.

Start with the stock equivalent calibration, check behavior, then import one table at a time until you narrow the problem down. Also try running the calibration with live tuning disabled in case that is the cause.

Final update - I turned off live tuning and my AFR's on light load highway cruising are now normal. Why would this be affected by simply turning on/off live turning?

At this point all my issues are resolved.

Thanks,

Mike

[bunger78]

3. No. but we intend to change the target table from three values at different loads to a full sized table.

Please! =)

[jun1or]

3. No. but we intend to change the target table from three values at different loads to a full sized table.

Please! =)

Whoa whoa, bringing my thread back from the dead. I think you need to post in the feature request thread.

Wait, now that I think about it, I may have alread done that (for request of target lambda overlay.)

[jun1or]

I hate bringing threads back from the dead, but after a year I have finally eliminated the Wideband Voltage Offset. (plus 2600 views on this thread dictates this should help others in the future)

Solution = RTFM... Read The F*ing Manual. (your wideband manual), and input the correct values in the lambda vs volts table. That simple.

I found this out only after spending hours logging the differences between my AEM gauge and the hondata display values. After I entered the correct values for AEM UEGO 30-5130, voila, everything matched perfectly with 0v offset.

Mike


(upper portion of picture is BEFORE I applied the correct values in Volts vs Lambda table, obviously)