bank 2 o2 sensor affect command a/f table
03-05-2012, 09:55 AM
#1
bank 2 o2 sensor affect command a/f table
Guys I have been following a thread in the tech section and the op is working on a few things trying to improve milage etc. He is doing a pretty good job of researching it seems.
Anyway he founf this concerning the bank 2 o2 sensor ( one after the cat):
He states when I asked him if he was certain that o2 sensor was directly influcing his a/f table:
"100% positive. I was directly offsetting the wideband signal to the ECU, and the Narrow Band was causing the ecu to drop commanded AFRs. With the signal wire pulled from the Narrow band (0volts to ECU or ultra lean.) my commanded AFRs dropped to 14.3 max instead of 14.8. When the ECU registered the NB as a fault AFRs went to 14.8 as usual.
It appears the ECU is using the NB as a go/no go test when at stoich. This had no effect to my ability to lean AFRs at idle, because the ECU expects to see a lean condition. Probably to prevent a lean condition that could cause engine damage."
I have never seen this discussed and I thought it may be of some interest.
__________________
Quote:
Anyway he founf this concerning the bank 2 o2 sensor ( one after the cat):
He states when I asked him if he was certain that o2 sensor was directly influcing his a/f table:
"100% positive. I was directly offsetting the wideband signal to the ECU, and the Narrow Band was causing the ecu to drop commanded AFRs. With the signal wire pulled from the Narrow band (0volts to ECU or ultra lean.) my commanded AFRs dropped to 14.3 max instead of 14.8. When the ECU registered the NB as a fault AFRs went to 14.8 as usual.
It appears the ECU is using the NB as a go/no go test when at stoich. This had no effect to my ability to lean AFRs at idle, because the ECU expects to see a lean condition. Probably to prevent a lean condition that could cause engine damage."
I have never seen this discussed and I thought it may be of some interest.
__________________
Quote:
03-05-2012, 10:27 AM
#2
Bank 1, sensor 2 short term fuel trim is added to the closed loop target.
That is OBD-II parameter 0x15, so you can monitor it with lots of tools.
Obviously, the ECU handles it fairly gracefully after it detects a fault, so all the people without a rear sensor haven't had many problems. This has lead to the myth that the sensor is purely for diagnostics.
edit:
It seems I need to get off my *** and finish a definition file for the rom. The final corrections to the closed loop target are the lean and rich limits, which are 1.05 and 0.85000002. Seems that some people might be interested in changing these. It is pretty incredible how lean the Renesis can run at low loads.
That is OBD-II parameter 0x15, so you can monitor it with lots of tools.
Obviously, the ECU handles it fairly gracefully after it detects a fault, so all the people without a rear sensor haven't had many problems. This has lead to the myth that the sensor is purely for diagnostics.
edit:
It seems I need to get off my *** and finish a definition file for the rom. The final corrections to the closed loop target are the lean and rich limits, which are 1.05 and 0.85000002. Seems that some people might be interested in changing these. It is pretty incredible how lean the Renesis can run at low loads.
Last edited by oltmann; 03-05-2012 at 10:48 AM. Reason: also
03-06-2012, 07:24 PM
#5
mmm... tastes like jesus
Join Date: Oct 2003
Location: Jacksonville, FL
Posts: 165
Likes: 0
Received 0 Likes
on
0 Posts
I guess I stand corrected in another thread in which i so blatantly stated that is was nothing more than a monitoring sensor.
If my understanding of what you are describing is correct then the function is simply a test for a sensor fault or responsiveness. The change to STFT would lead to a corresponding change in sensor output allowing the system to check itself - no?
So while its response and feedback can alter afr's it happens through a primarily diagnostic subroutine?
Ray
If my understanding of what you are describing is correct then the function is simply a test for a sensor fault or responsiveness. The change to STFT would lead to a corresponding change in sensor output allowing the system to check itself - no?
So while its response and feedback can alter afr's it happens through a primarily diagnostic subroutine?
Ray
Bank 1, sensor 2 short term fuel trim is added to the closed loop target.
That is OBD-II parameter 0x15, so you can monitor it with lots of tools.
Obviously, the ECU handles it fairly gracefully after it detects a fault, so all the people without a rear sensor haven't had many problems. This has lead to the myth that the sensor is purely for diagnostics.
edit:
It seems I need to get off my *** and finish a definition file for the rom. The final corrections to the closed loop target are the lean and rich limits, which are 1.05 and 0.85000002. Seems that some people might be interested in changing these. It is pretty incredible how lean the Renesis can run at low loads.
That is OBD-II parameter 0x15, so you can monitor it with lots of tools.
Obviously, the ECU handles it fairly gracefully after it detects a fault, so all the people without a rear sensor haven't had many problems. This has lead to the myth that the sensor is purely for diagnostics.
edit:
It seems I need to get off my *** and finish a definition file for the rom. The final corrections to the closed loop target are the lean and rich limits, which are 1.05 and 0.85000002. Seems that some people might be interested in changing these. It is pretty incredible how lean the Renesis can run at low loads.
03-06-2012, 07:48 PM
#6
There is no commanded value related to the rear O2 sensor in the RX-8 ROM that is active.
The legacy code has no direct lookup.
Try it.
You can send any value you want to that input and there is no direct correlation to the commanded fueling or ignition timing.
The legacy code has no direct lookup.
Try it.
You can send any value you want to that input and there is no direct correlation to the commanded fueling or ignition timing.
03-07-2012, 12:23 AM
#7
It is added at the end of the closed loop target subroutine. No branches over it, nothing "legacy." Like I said, it is the exact value that is returned by standard OBDII parameter.
You are (once again) making a type II error.
Narrow-bands are faster, more accurate, and more reliable than widebands. They are used to correct for the aging of the wideband on most cars with this setup. You can get by without this sensor, and you can also live with one kidney.
03-09-2012, 04:42 PM
#9
I did try it and you are wrong. There is some sort of failsafe limit the ECU has. It does effect commanded AFRs. I left my logs in the other thread, but I have seen it with my own eyes. As you try to lean burn out of idle the ECU fights you to maintain stoich, and it is from the rear o2 sensor.
03-14-2012, 07:40 AM
#10
Registered
Bank 1, sensor 2 short term fuel trim is added to the closed loop target.
That is OBD-II parameter 0x15, so you can monitor it with lots of tools.
Obviously, the ECU handles it fairly gracefully after it detects a fault, so all the people without a rear sensor haven't had many problems. This has lead to the myth that the sensor is purely for diagnostics.
edit:
It seems I need to get off my *** and finish a definition file for the rom. The final corrections to the closed loop target are the lean and rich limits, which are 1.05 and 0.85000002. Seems that some people might be interested in changing these. It is pretty incredible how lean the Renesis can run at low loads.
That is OBD-II parameter 0x15, so you can monitor it with lots of tools.
Obviously, the ECU handles it fairly gracefully after it detects a fault, so all the people without a rear sensor haven't had many problems. This has lead to the myth that the sensor is purely for diagnostics.
edit:
It seems I need to get off my *** and finish a definition file for the rom. The final corrections to the closed loop target are the lean and rich limits, which are 1.05 and 0.85000002. Seems that some people might be interested in changing these. It is pretty incredible how lean the Renesis can run at low loads.
I need to go rich fast from that when any more load is applied....
03-14-2012, 01:26 PM
#14
You can "trick" the ecu into delivering leaner fueling several ways. Run open loop at low loads and lean out the VE table. Set the closed loop fueling target beyond the max correction threshold (0.23). Alter the wideband signal as Harlan has done. The are also off-the-shelf devices for this.
Personally, I've never found big gains from running a leaner mixture, at least when averaged over several tanks of gas. It may require some calibration tricks I don't yet understand. For example, I've read that by changing fuel injection timing it may be possible to stratify the air/fuel charge such that the mixture near the spark plugs is richer. Also, running lean properly in closed loop without any ugly hacks would probably help.
There is still enough data to satisfy me that substantial gains are possible, and I'm very interested in the ongoing experiments.
Also, a couple of notes. The basic fuel injection amount is set by the VE table, closed loop tables and numerous other corrective factors under all circumstances. Exiting closed loop means turning off short-term feedback correction and turning on "open loop" enrichment. Mazda calls open loop the "High load volume increase zone" and describes it thusly:
VE table is also a bit of a misnomer. The load value is the ecus estimate of volumetric efficiency. The VE table is a corrective term. I think it is mainly a compensation for the ratio of fresh air to residual exhaust gas in the chamber. So inverse residual gas fraction model might be a better name for that table... or maybe not. 
In any case, you can use it to change injector pulse width, but it bears little relation to the VE tables that are used in speed-density systems.
Personally, I've never found big gains from running a leaner mixture, at least when averaged over several tanks of gas. It may require some calibration tricks I don't yet understand. For example, I've read that by changing fuel injection timing it may be possible to stratify the air/fuel charge such that the mixture near the spark plugs is richer. Also, running lean properly in closed loop without any ugly hacks would probably help.
There is still enough data to satisfy me that substantial gains are possible, and I'm very interested in the ongoing experiments.
Also, a couple of notes. The basic fuel injection amount is set by the VE table, closed loop tables and numerous other corrective factors under all circumstances. Exiting closed loop means turning off short-term feedback correction and turning on "open loop" enrichment. Mazda calls open loop the "High load volume increase zone" and describes it thusly:
Corrections are added to the basic injection amount and the high load coefficient is calculated according to the engine speed, mass intake airflow amount and the throttle valve opening angle.
In any case, you can use it to change injector pulse width, but it bears little relation to the VE tables that are used in speed-density systems.
Thread
Thread Starter
Forum
Replies
Last Post
