Semi-Open Loop Questions

[ferocity02]

So I think I got semi-open loop to work... if the short term fuel trims are anything but zero does this mean it's working? They also go straight to zero once EQR is anything but 1.0.

Something I'm confused about... So I have B3601 set to 14.17 AFR (assuming E10), and I'm commanding an EQR of 1.0 using the STFT's for trimming... This gives me an average wideband lambda of 1.012. Since my NBO2 sensors are attempting to find stoich, does this mean my fuel's stoich is actually 14.17*1.012=14.34?

Or what is the significance of lambda not being 1.0 when the STFT's are trimming to stoich?

Here are the settings I used that are different than stock:
B3801 Long Term Fuel Trim Correction: Disable
B4108 STFT Idle Enable: Disable
B4205 Closed Loop Temp Enable: All maximum (284F)
B3647 Commanded Fuel vs RPM (Normal): Set to 1.0 in area's where I want semi-open loop, which is every column 90kPa and under

Does anyone have tips on adjusting the STFT settings to improve their trimming ability? They seem to be working okay, but maybe they can be better?

My first impression is that I like semi-OL. It helped quite a bit with the light throttle and decel bucking. Everywhere else I haven't noticed a difference.

Here's a log from today, my first log with semi-OL. And a screenshot of the STFT's cycling.

[joecar]

Try setting B3601 to 14.17/1.012 = 14.0

[ferocity02]

Try setting B3601 to 14.17/1.012 = 14.0

Thanks Joe. Could you explain the logic for doing this?

My wideband is outputting lambda as a voltage. I use this lambda to calculate AFR: WB_AFR=WB_Lambda*B3601

So using my previous example, WB_AFR=1.012*14.17=14.34. Since the NBO2's target stoich, does this mean 14.34 is stoich?

Or is there just a 1.2% error in what the wideband thinks in stoich versus what the NBO2's think is stoich? If this is the case, perhaps adjusting the switchpoints is a fix. Mine are all 450mV at the moment.

Coincidentally I was dialing in the VE tables in OL commanding an EQR of 0.99, which is a lambda of 1.01, pretty close to what I'm getting now in semi-OL.

[joecar]

My logic was to try B3601 set to 14.0 and see what thw wideband reports... and then go from there (you may be correct in saying stoich is 14.34, but I just wanted another data point).

[ferocity02]

My logic was to try B3601 set to 14.0 and see what thw wideband reports... and then go from there (you may be correct in saying stoich is 14.34, but I just wanted another data point).

Ok I set B3601 to 14.0 and logged again this morning, the average cruise lambda was 1.008. Not sure what to conclude from this if anything?

[joecar]

Lol...

Try setting B3601 to 13.9 and see if it is consistent.


What fuel are you running...? Have you checked its alcohol content (using the cheap $20 test kit)...?

[ferocity02]

Lol...

Try setting B3601 to 13.9 and see if it is consistent.


What fuel are you running...? Have you checked its alcohol content (using the cheap $20 test kit)...?

91 pump gas. Have not checked the ethanol content yet, perhaps I should. I'll try a new B3601 on the way home today.

Perhaps there is just an error between what my wideband and my NBO2's think stoich is. I mean, it's only 0.8-1.2%. I'll log my NBO2's today also.

[ferocity02]

I bought a testing kit so I'll know for sure what I'm getting out of the pump soon.

[5.7ute]

I just modify the switch points for the o2 sensors until the trims cycle around stoich.

[ferocity02]

I just modify the switch points for the o2 sensors until the trims cycle around stoich.

What switchpoints are you running now?

[5.7ute]

I couldn't say for certain as I do not have efilive on this computer. I logged the wideband & narrowband sensors & kept lowering the o2 switch points until I had an average of stoich while trimming. From memory I was down to around the 400 - 420 millivolt range. You need to make some custom pids to work out which CL mode you are in to get it tuned in accurately. There is a how to on here somewhere but I wont have time to look for a little while.

[5.7ute]

Look here for how to set up the CL mode pid http://forum.efilive.com/showthread.php?8142-B4105-how-do-i-link-this-table&highlight=lookup+cl+mode

[joecar]

91 pump gas. Have not checked the ethanol content yet, perhaps I should. I'll try a new B3601 on the way home today.

Perhaps there is just an error between what my wideband and my NBO2's think stoich is. I mean, it's only 0.8-1.2%. I'll log my NBO2's today also.If you use wideband lambda and commanded EQR then the only stoich AFR that matters is what you enter in B3601.

[ferocity02]

If you use wideband lambda and commanded EQR then the only stoich AFR that matters is what you enter in B3601.

Makes sense. I guess what I meant to say is that perhaps what my NBO2's think lambda=1 is different than what my wideband thinks.

But I'm having an issue now, and it might be related to lean cruise being enabled. In B3647 I'm commanding 1.06 EQR at 95kPa. On the highway (when all lean cruise enablers are met) if I hit 95kPa it's still commanding 1.0 EQR. Lean cruise is not removing any fuel at this point according to B3639. It seems like if lean cruise meets the enablers for speed and timers, then it will run on B3639 and ignore B3647. Is this correct? STFT's also do not activate when EQR=1.0 and the lean cruise enablers are met. (see the log 101 frames 7320-7410)

Also, the STFT's seem to go to zero under deceleration at high-ish RPM's >3000 or so even with DFCO disabled. Is it possible that it thinks DFCO is active even though its not cutting fuel or spark? Log 99 has some good examples of this.

[EagleMark]

Makes sense. I guess what I meant to say is that perhaps what my NBO2's think lambda=1 is different than what my wideband thinks.
Yup!

Been playing with this for months. Can run OL on my tune and lambda gauge and serial WB steady at 1.0

Closed loop:

Bosch O2 sensors = .97
Second set of Bocsh O2 sensors = .99
AC Delco O2 sensors = 1.02
Denso O2 senor = .99

All sensors were new when installed running same tune.

[ferocity02]

Yup!

Been playing with this for months. Can run OL on my tune and lambda gauge and serial WB steady at 1.0

Closed loop:

Bosch O2 sensors = .97
Second set of Bocsh O2 sensors = .99
AC Delco O2 sensors = 1.02
Denso O2 senor = .99

All sensors were new when installed running same tune.

Interesting results! Sounds like it's just an error between the NBO2's and wideband.

I tested my fuel today and it's about 8% ethanol. If my calculation is correct I should set my B3601 to 14.236, sound about right? It's currently 14.17.

[EagleMark]

Stoich for gas is 14.7 AFR. Stoich for Ethanol is ~9.00AFR. (14.7 x .92) = 13.524 = (9.00 x .08) = .72

13.524 + .72 = 14.244 for 8% ethanol