When I pulled the old O2 sensor out of the drawer, I discovered that much of the soot on the outside could be wiped away so I started by hosing the thing down with brake cleaner, and got quite a bit of it off that way, and then clamped it in a pair of mole grips and heated it with a butane torch - after some working around it I found it quite responsive rising to a shade over 900mV and dropping rapidly once the torch was removed.
I'll reinstall it, probably on the week end and see what difference if any it makes.
That's interesting - I have the old O2 sensor I took off of my Mitsubishi - all black & sooty on the inside, I'll give it a shot - like you I have nothing to loose if it doesn't work.
I have been having problems with my 1998 Jimny G13BB recently and the latest problem was a "Check Engine" light which turned out to be a DTC Code 13 - HO2S Sensor Low Voltage.
Suzuki Australia wanted A$407 (~GBP285) for a new "genuine" HO2S sensor and the local after market part supplier wanted A$177 (~GBP124) for a Bosch replacement.
I recently read in a professional repair manual that faulty HO2S sensors could sometimes be given a new lease of life by carefully heat treating the element to remove contamination. The Suzuki workshop repair manual recommends replacing the HO2S sensor at 100,000Km intervals (at 400 bucks a pop, I don't think so!) and my Jimny has done over 330,000Kms. So what did I have to loose?
DISCLAIMER: This information is provided as just that, information. I make no guarantee that this procedure will work in every, or any case, and I stress that you MAY damage your sensor by carrying out this process. It worked out well for me, you may not be so lucky.
CAUTION: The HO2S sensor uses a ceramic element. DO NOT subject it to sharp knocks or drop it on the concrete floor of the garage. If you do, it will suffer the same fate as your favorite china mug treated the same way - smashed beyond resurrection! But enough of the doomsayer stuff...
The first drama was actually getting the HO2 sensor out of the manifold. The electrical connector ("coupler" in Suzuki speak) is hidden away between the head and the firewall, underneath the camshaft position sensor housing and the EGR valve. The lock tab has to be released with a screwdriver and that is a challenge of its own! The job would have been a lot simpler with two people, one on the screwdriver and the other to wrestle with the connector. Once the sensor cable is free of the harness, you will need a 22mm ring spanner or dedicated HO2S socket to remove the sensor itself. The dedicated socket is a deep socket with a slot in the side to accomodate the cable. I found the ring spanner OK and actually had to help it along using a piece of wood as a hammer to "crack" the seal. I had previously applied a liberal supply of penetrating spray (WD40, etc) to the sensor and manifold.
Once free of the vehicle, I bench tested the heater element resistance (the two black wires) and found the heater was open circuit. It shold be around 10-15 ohms depending on the ambient temperature. I decided to continue trying to resurrect the sensor as early O2 sensors were not heated anyway. Electrical heating was added later in the EFI development phase to bring the O2 sensor up to its required operating temperature quicker to minimise overfueling when the engine is first started.
But that is getting ahead of ourselves. First, a quick look at how the O2 sensor works and how it is used in an EFI system. This explanation is very basic, there is a very good treatise on wikipedia under "oxygen sensor" if you would like more information.
The basic zirconia O2 sensor is a fuel cell which generates a voltage like a battery when it is heated in the presence (or lack) of oxygen. When there is lots of oxygen, the voltage is low, around +100 millivolts (mV). When there is not much oxygen, the voltage rises to around +900mV. The ECU monitors this voltage. If the mixture is lean, there is an excess of oxygen and the voltage output is low. The ECU increases the time the injector is open to let in more fuel. If it is too rich, there is not much oxygen and the voltage is high, so the ECU reduces the time the injector is open to reduce the amount of fuel. In practice, the ECU is constantly varying the injector timing to create a rythmic swing between slightly rich and slightly lean around the ideal stochiometric air/fuel ratio of about 14.1:1 about 2-3 times per second. There are various sites on the net showing this voltage swing when viewed with an oscilloscope, engine analyser or graphed with some of the newer OBD scan tools if you are that way inclined.
When the sensor becomes contaminated with combustion byproducts, the output voltage can decrease and/or the response time of the sensor can get progressively more sluggish. In the case of the Jimny G13BB ECU, a DTC Code 13 indicates an O2 sensor low voltage output. Other ECUs on other vehicles and brands may report more detailed information. As it turned out, my O2 sensor suffered both low output and VERY sluggish response time.
As mentioned earlier in this article, I had only read that it may be possible to resurrect a contaminated O2 sensor. The instructions were pretty sketchy, but I figured it was worth a try anyway as my sensor was effectively useless as it was.
Photo 1 shows the sensor mounted in a vice ready for the heat treatment. Unfortunately, I did not take a photo of the sensor before I started the heating process and some of the tan coloured contamination has already been burned away. The whole element was the same "ideal spark plug tan" colour inside and out that can be seen where the element meets the body before I started the heat treatment.
Photo 2 shows the gas torch I used. This torch uses a mix of butane and acetylene gas and I bought it some years ago to do some heavy brazing as it burns hotter than straight butane or propane. Straight butane or propane should work OK, I only used a very low flame. If you use oxy-acetylene, use a soft neutral flame. Too much actylene and you risk contaminating the element further with carbon, too little and you risk it being too hot. The important thing is to use a LARGE, SOFT flame and keep it back from the element. Too much heat and you WILL melt the sensor outer case and probably damage the ceramic element beyond repair. Photo 5 should give you a good idea of how close to get.
Photo 3 shows the test setup. Mount the HO2 sensor in a vice and connect a digital voltmeter to the Blue and White wires of the sensor, positive to Blue, negative to White. Make sure you keep the wires well away from the element before you start to heat it up to avoid burning them. Apply the heat gradually and evenly around the element to raise the general temperature of the element as a whole rather than shocking one side and risking fracturing the element. Once it is warmed up faily evenly, you can concentrate on one spot at a time. In my case, I could see the contamination changing colour from tan to dark grey as it burned off. Initially, the voltage output remained very low and did not change much when the flame was applied and removed. But gradually as I kept working around the element, the contamination became less and less and the output began to rise to the point where it rose to over 900mV. But the response time was still very sluggish. It took several seconds for the voltage to rise and then die away as I removed the heat.
But perseverence paid off. As I continued to work around the element, the response time got quicker and quicker. So much so that it became a real problem to take photos of the setup to show the output voltages under different conditions. It is important to note that whilst the O2 element needs to be hot to give an output, the output voltage is dependant on the oxygen level at the element, not the temperature. Photo 5 shows the high output voltage in Hot Rich mode, you can see the flame is directed at the element and "breathing" oxygen starved Carbon Monoxide over the element. In photo 6, the flame has just been removed and the element is still glowing bright red, but now the element is bathed in oxygen rich air and output voltage has dropped to a very low level. The response time was now so good, it followed the flame as it moved. Photo 7 shows the mid level output as the gas stream passes the element.
I refitted the HO2S sensor to the vehicle, reset the DTC code by disconnecting the battery negative lead for a few minutes (the manual says 30 seconds at an ambient temperature above 0C), and started the car. The "Check Engine" light came on during cranking and disappeared as the engine sprang to life. I took the car for a spin up the highway, sitting on 100Kph. About 15Kms out I turned around and stopped the engine for a few minutes. I started the car again and drove home at the same speed. And no "Check Engine" light. I am one very happy little Vegemite, having saved myself a couple of hundred dollars. And the little car even feels more responsive, but that could just be euphoria setting in!
I guess a YouTube video would have been a bit more instructive, but I am not up to that level of technical sophistication yet. Hopefully these photos will be enough to get some of you experimenting with those "dead" O2 sensors.
