EfiMini O2 Sensor Operation

From efiHacks Wiki
Jump to: navigation, search

And More Details

Well ... the O2 sensor in narrow band flavor or wide band.

While lots of experts might say the only way to tune is with a wide band O2 sensor ... yes it's very convenient ... maybe a bit slow. Depending on the sensor and the controller ... we try to compare the 2 setups.

We believe that a stock narrow band O2 sensor (NbO2) can deliver enough information of where it's still to lean and we'll get to this in a moment.
We want to point out some technical and logistic differences between NbO2 vs WbO2.

Narrow band O2 sensing :

  • plus - sensor already installed.
  • plus - sensed data in voltage form, range 0 V to 1.0 V.
  • plus - tap-ing into the existing O2 sensor to ECU wiring possible.
  • plus - no complex controller necessary, voltage input sufficient.
  • plus - low sensor cost for replacement.
  • plus - sensor seems fast enough for A/F indication during acceleration.
  • minus - only very narrow bandwidth info.

Wide band O2 sensing :

  • plus - wide measurement range.
  • minus - original location of NbO2 sensor position within the exhaust stream too hot, relocation necessary.
  • minus - additional cost of more sophisticated extra sensor.
  • minus - additional cost of required sensor controller (2 internal closed control loops).
  • minus - possibly additional NbO2 simulation for ECU necessary (built into WbO2 controller).
  • minus - slow response time from sensor and / or sensor controller.

Now the only positive aspect of a wide band setup, is the WIDE BAND range of A / F ratio measurements and the need thereof, which is why this sensor was designed at the 1st place.
At the extra cost of approx $ 70 plus for the controller and at least the same for the sensor.
Most likely relocation of the sensor to a cooler place. Additional expense and / or more mechanical work.
Additional NbO2 simulation and the chance of incompatibilities with the ECUs sensor detect strategy, tripping the engine failure light.
Yes, we do value the capabilities of the wide band measurement results, but at a significant expense. i.e. 2 cylinder banks with 2 O2 sensors.

Compared to the ease of electrical and mechanical setup, no additional expense, but the limited narrow band A / F ratio indication.

1st hand experience with O2 sensors. From our testing 2 cases come to my mind.

A KIA Sephia as one of our 1st test vehicles. With two 4 wire NbO2 sensors. Only the pre-cat one needs attention, as long as the post-cat one doesn't trip the engine failure light. Otherwise the 2nd one needs simulation as well.
The weird thing was that the O2 sensor ground was kept by the ECU at about 2.5 V measured against ground. Now for the measuring this was ok since the efiMini does differential voltage measurements with a floating ground range from 0 V to 5 V .
However we could not do an NbO2 simulated output since the efiMini version at this time had the O2 sensor output fix at ground and could not handle the referenced ground to 2.5 V.

Most WbO2 controller sim ourputs would not be able to accommodate the 2.5 V ground reference for the KIA Sephias O2 simulation requirements.

Lesson learned : changed the O2 sensor out circuit from fixed ground ref to floating ground ref to accommodate virtual ground setups.

The 2nd learning experience was on a KTM 690 single cylinder bike. The efiMini was tap-ed into the NbO2 sensor to ECU signal and everything appeared fine. Nice 100 mV to 800 mV sensor voltage swings during closed loop operation. But we wanted to find out more.

Voltage Measurement Setup : O2 sensor to ECU, no efiMini connected.
Measuring the voltage on a cold sensor (high high impedance) we measured 3.2 V on the sensor. The shops cheap multimeter had an input impedance of 1 MOhm, so the ECU has an internal pullup to 5 V with about 500 kOhm cause remeasuring with our meter with 10 MOhm read 4.8 V. Measuring it again with a hot sensor right after stopping the engine brought low readings cause of the low impedance of the hot sensor.

Voltage Measurement Setup : O2 sensor to ECU, with efiMini connected.
Even on a cold sensor and ignition on we measured about 450 mV. We were under the impression that our non inverting efiMini O2 sensor input had 133 kOhm impedance which would get the reading to about 1 V. Why 450 mV which would imply a 60 kOhm impedance. However, running the engine showed the 100 mV to 800 mV swings.

After we took the ECU off the sensor and kept it happy with a 1 MOhm resistor to ground, strange things were observed.
The voltage on the sensor barely reached 500 mV after engine start and on a 10 sec full power accel run from almost idle to max RPM the Sensor voltage slowly dropped down to 300 mV even our tail pipe WbO2 sniffer indicated total rich A / F ratio.
We need to do way more investigation into this test scenario.

Done : changed the diff OpAmp input resistors from 33k to 100k ... got now 0 V - 3.3 V input range and 200k total input resistance on non inverting input. Added 1M pullup to 3.3 V.

Observations : at low O2 sensor temp (engine at idle for 38 secs after engine start) we reached a max of 856 mV. Decreased to 580 mV after about 100 secs and had their maxs during up-shifting at same 580 mV. Dropped to 150 mV during 0 throttle to wait for RPMs to decrease. Increased to 530 mV before 10 sec full power (dyno) run and stayed there during the run (2800 RPM to 8500 RPM, 10 sec, 6th gear).

Now we took the O2 sensor off the efiMini rev 1 (efiMini 1M pullup at 3.3 V and 200k input impedance) and measured the O2 sensor voltage with a 10M DVM and nothing else hooked up to the O2 sensor. Did a similar power run (10 sec, 2800 RPM to 8500 RPM) and observed a steady about 900 mV on the meter.

Later an average range 700 mv to 800 mV on the meter got a 20 % drop once a 200k resistor was put across. Additional 20 % were observed during very high sensor temps at full power output.

Lesson learned : efiMini rev 1, raising the efiMini O2 sensor input impedance by a factor of 5 where the OpAmp resistor value changes also bring an input voltage range change from 0 V - 1.2 V to 0 V - 3.3 V. We also assume that a really hot O2 sensor seems less capable to provide it's 800 mV on a 200k load and the output value drops at lease by 20 %.

More to come ...

Attention :

As soon as anything related to an A / F ratio change caused by the outside of the ECU happens there is a chance that the ECUs internal short and long term adjustment might be effected, if the A / F ratio happens to be within the closed loop operation range of the ECU. This now has an inverse effect on the desired A / F ratio change if one falls within this closed loop range of the ECU.

Also if any internal long term adjustments in the ECU occurs after an external adjust mapping has been placed, this internal long term ECU adjustment might render the external adjustment less effective. We assume that this is the reason why many piggy-back controller manufacturers suggest the O2 sensor eliminator kit in form of a disconnected O2 sensor signal wire and a 1 MOhm resistor on the ECUs O2 sensing terminal to gnd to keep the ECU happy and preventing it to perform any internal short or long term A / F ratio adjustments.

Note : All published design documentation is Copyright efiLabs.com and its use is without exception FOR NON COMMERCIAL PURPOSE ONLY. For commercial licenses contact efilabs.com ... try us, we're reasonable :)

Personal tools