View Single Post
Old 05-23-2004, 01:09 PM   #9
Jon [in CT]
*** Banned ***
Member#: 2992
Join Date: Nov 2000
Chapter/Region: NESIC
Location: Connecticut, USA
02 WRX Sedan


Here's a summary of what I think is going on here. OFSTI buggered his front A/F sensor or one of its leads, probably during the install of his new headers. He works at a Toyota shop. When he discovered how expensive a new sensor from Subaru was, he decided to try and use a Toyota sensor instead because he can get one of those "cheap." For some unknown reason, OFSTI selected a Denso planar wideband A/F sensor that's started appearing on Toyotas a year or two ago, instead of an older Denso wide range A/F sensor like Toyota part number 89467-33010. I bet the STi sensor looks exactly the same as that older Toyota part. Anyway, OFSTI cut off the new sensor's Toyota connector and soldered on the connector from his bad Subaru sensor, and it apparently works.

First, the information I have on the Subaru front A/F sensor is for a 2002 WRX. I suspect the STi uses the same sensor, but only a check of Subaru's parts database can confirm that.

The WRX sensor is probably identical to that older Toyota sensor whose part number i cited. Unlike NTK and Bosch wideband sensor which have two cells (one of them is called the pump cell) and 5 wires in the pigtail, the Denso sensors used by Subaru and Toyota have only a single cell and 4 wires in the pigtail. Rather than relying on a pump cell current, the Denso sensors rely on a limit current. The WRX sensor's principle of operation is described in this paper, Development of Air Fuel Ratio Sensor for 1997 Model Year LEV Vehicle, and the Toyota planar wideband is described in this paper, Development of Planar Air Fuel Ratio Sensor.

The sensor's "signal" is current magnitude (milliamps) and direction (+/-). Toyota ECUs send 3.5 volts to the sensor's positive terminal and 3.0 volts to the sensor's negative terminal. Thus, the sensor "sees" 0.5 volts. The detetector circuit in the Toyota ECU converts the current generated by the sensor into a voltage such that 3.3 volts corresponds to 0 milliamps (stoichiometric A/F ratio) and sends that voltage on to an A/D converter in the ECU. Toyota scan tools can report the voltage output from the detector circuit. This is described beginning on p. 6 of

The Subaru ECUs work differently. They send a variable voltage (2.8 to 3.2 volts) to the A/F sensor's positive terminal and a variable voltage (2.4 to 2.7 volts) to the A/F sensor's negative terminal. Thus the sensor "sees" from 0.1 to 0.8 volts. The Subaru ECU's sensor detector circuit also converts the sensor's current to a voltage, but the details about it are unknown and aren't reported by the Subaru scan tool. Instead, the Subaru scan tool reports the result after the signal has been digitized and transformed to a lambda value.

Both the "cup type" A/F sensor used by the WRX and the planar A/F sensor produce the same current at stoichiometry, 0.0 milliamps, so this may explain why the planar sensor seems to work in the STi. However, heater operation is very important and improper operation can damage/crack the A/F sensor. So I would be concerned about this since the heaters are radically different in the two types of sensors. Although, the planar A/F sensor can accurately read much richer A/F ratios than the cup type A/F sensor in the WRX, I doubt whether the STi's ECU will ever report anything richer than when the more-narrow range cup type A/F sensor was used.

The MATCO scanner apparently can report the Mode $01, PID$024 OBD parameter that Subaru ECUs use to report data from the front A/F sensor to generic OBD-II scan tools. Subaru has never defined what, exactly, is being reported there and few scan tools report that PID, so that is still a mystery to me.
* Registered users of the site do not see these ads.
Jon [in CT] is offline   Reply With Quote