 | |||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
|
Remain Flexible In Your DiagnosticsPosted 3/15/1997By Mark Warren, AAM
The 1990 Celica came into the shop for an oil change. After a routine lube, oil and filter, the tech pulling the car out noticed black smoke from the tailpipe. A quick check with the infrared showed the Celica running at 8 percent carbon monoxide (CO). This was way too rich and needed to be dealt with right away to prevent converter damage. The customer was called and informed of the situation; of course, "it didn't do that before you worked on it." A wire could have been bumped off during the oil change -- anything is possible -- we'll check and call you back. A check of the service records showed the car was in three months ago for a major service and the emissions were perfect then. This wasn't at my shop, but it was at a great shop with good techs and good equipment. The techs are also trained in my "begin at the end" technique. They started by checking codes. Sure enough, there was one for the O2 sensor -- too rich. Next, they checked the O2 signal, the injector pulse and the tailpipe. The O2 was pegged rich at one volt, the pulse width was "normal" at 2.19 ms and the tailpipe was 8 percent CO. They knew the tailpipe was rich and they knew the O2 was aware of that. Shouldn't the computer respond by narrowing the injector pulse width? Down, to say, 1.0 to 1.5 ms? If the pulse width got down low and we still ran rich, we could assume fuel from somewhere else -- blown regulator diaphragm, high fuel pressure, leaking injectors or something such as that. They checked fuel pressure and it was normal. They decided to clean the injectors -- maybe one was sticking. Nice try, but no good. They decided that possibly the computer was bad, since no scan data means no way of knowing what the computer may be "thinking." A new computer changed nothing but the weight of their pocketbook. Finally, they tested component by component using the Toyota factory book ... no problem found. I've got to hand it to their guys, they really did try everything before they called me. What went wrong? OK, lets review. A "normal" pulse width is 1.5 to 3.0 ms, engine warm and in closed loop at idle or 2,500 RPM no load. Less than 1.5 ms is taking away fuel (enleanment) and more than 3.0 ms is adding fuel (enrichment). This is our general rule. It may not always be true; however, it should have been true for this Toyota, in my experience. Keep in mind, my job has been made easier by the fact that they have eliminated many possibilities for me. What was the key to fixing this Toyota? It was obvious the O2 was not in control. If the engine is operating outside the command authority of the O2, then I assumed that some sensor with higher command authority was out of specification. This shop had a SimuTech connected, so I went ahead and checked the computer grounds to be sure they weren't high, causing a sensor shift. No problem there. I checked the coolant temperature sensor next because of its high command authority. It was OK. Throttle position sensor and air charge temperature sensor specs were OK also. Down to that pesky MAP sensor. The tech working on the car had documented every step and recorded the measured values and the book values. The MAP sensor was OK according to the specs in the manual; however, the test involved only voltage changes relative to changes in vacuum applied to the sensor. There was no baseline spec on the sensor, only relative changes. Was this sensor shifted to a higher voltage reading (add fuel)? Without adequate data, we decided to boldly go where no man had gone before! We installed a variable resistor in place of the MAP sensor so that we could control the input to the computer. The baseline on the MAP sensor was 3.37 volts key-on-engine-off and 1.56 volts at 18" of vacuum. Using the resistor, we slowly cranked down the voltage. At 1.40 volts, the O2 started sweeping just fine and the tailpipe emissions dropped to near zero. We cranked further to see where the lean limit of the O2 command was reached. At 1.21 volts, the O2 stuck lean and could not compensate for the MAP enleaning the system. So, now we know that for the O2 to be in control, the MAP sensor voltage at 18" of vacuum (idle) needs to be between 1.21 and 1.40 volts. We figured 1.3 to be perfect. A new MAP sensor from Toyota was $400. Ouch! The customer wanted to know for sure that this was going to fix the problem and had already spent some pretty good money on diagnostics. He was not billed for the parts that did not fix the car. A Mitsubishi MAP sensor (MD 178243) costs $8.75. Just maybe its output was the same. We decided to install one for a confirmation of the fix. Well, it changed voltage at the same rate, but it was shifted by .2 volts the other way. Playing with various resistors, we found we could shift the voltage to where we wanted by installing a 22k ohm resistor in one of the three wires going to the MAP sensor. You may ask which wire? Frankly, I don't know -- we got so confused at this point, we just found which wire shifted the voltage in the direction we wanted and then substituted resistors until it worked. Well, the car ran great and responded well under all driving conditions. The owner opted for the cheap fix and six months later all is well. The point is to remain flexible in your diagnostics. I would have never imagined that the difference between 2.04 ms, 2.14 ms and 2.19 ms injector in time would cause an engine to go from dead lean to perfect emissions to 5 percent CO. This is what threw the tech working on the car. According to what I had taught him, these were all normal idle specs. The key was noticing that the O2 did not have control. Often when something goes out of spec, the computer will substitute a fixed strategy and feedback control is lost. From this point, we must backtrack and find the bad signal.
Later we found the specifications on the MAP sensor we needed in a Toyota training manual. The graph of the MAP voltage vs. the vacuum showed that the MAP signal at 18" of vacuum should be 1.3 volts, just as our testing showed. We need to be looking at these sensor voltages to the hundredth of a volt sometimes.
|
||||||||||||||||||||||||||||||||
Copyright (c) 1996-2008. Automotive Service Association. All rights reserved.
 | |||||||||||||||||||||||||||||||||
