South Buckner Auto Clinic, Inc.

6414 C.F. Hawn Frwy. Suite 6

Dallas, TX 75217 (214)398-6097 Fax (214)398-1660

June 7, 1997

Bob, here's a story almost four years old, but it sticks clearly in my mind. We worked on it for seventeen days, our record, but it did leave fixed!

A large and well respected used car lot in Dallas, Texas, sent us this car, a 1988 Chevrolet Beretta with a 2.8 V-6 engine. Their shop had worked on the problem unsuccessfully for sometime before we saw it. The complaint was a very bad miss and erratic engine operation. The problem only occurred intermittently, usually only occurred when the engine was very hot, and when it was under a heavy load.

First, a little information about Chevrolet 2.8 engines. Anybody can fix a Chevrolet, right? Wrong! The 1988 Chevrolet 2.8 is a strange bird. It is a multi-point (but not sequentially) fuel injected engine. This means that it has an injector for each cylinder, but they are not injected individually. There are two fuses that power three injectors each. There are two injector drive circuits in the computer which grounds three injectors at a time. This does not present a problem with how the engine runs, but it does present a problem for a technician trying to diagnose an injector problem. The injectors in this engine are covered by the intake manifold. A technician looking for a bad injector cannot get to the injectors for individual testing. He must test the four- wire harness coming out from under the manifold. By doing so, he is testing three injectors at a time, not individually.

Now, having said all this, this is not the really strange thing about the 1988 2.8. Originally that year (and some 1987 models), the vehicle came equipped with both a map sensor and a mass air flow sensor. Both sensors are used to measure engine load, but this is the only engine I know of that came equipped with both! All other vehicles have one or the other. The on-board computer used information from both of these sensors for it's calculations. A map sensor measures engine load by measuring manifold pressure, and a mass air flow sensor measures engine load by measuring air volume entering the intake manifold.

Immediately, GM had some sort of problem with this arrangement. They began manufacturing "speed density software" to update these vehicles. This speed density software consisted of a new prom, a new idle speed control motor, and a label to attach to the ECM (computer) that the vehicle had been updated with the speed density software. I am unsure of exactly what this corrected, but it told the ECM to ignore all signals from the MAF sensor. The MAF (mass air flow) sensor was still there, it just didn't do anything anymore. It presents problems for a repair technician if he is unaware of whether the vehicle has been updated or not. A scanner plugged into the ALDL link will give erroneous data and trouble codes if the vehicle has been updated and the scanner is not made aware of it!

The reason for all this explanation is to make it known that this is an unusually difficult vehicle to troubleshoot. They seem to always have the strangest problems that are rarely seen. Even the technicians at the large Chevrolet dealership near our shop tell us that they try to run from them whenever they come in.

When test driving this vehicle, it appeared okay most of the time. On some occasions though, under hard throttle conditions, the engine would backfire rapidly and repeatedly. It would misfire and generally run real bad until the operator released the throttle, at which time it would immediately run good again. These symptoms usually indicate an ignition problem. Tests on the oscilloscope day after day revealed nothing. The three ignition coils and ignition module assembly were removed and a known good set were installed for testing. These are very difficult to change because of their location on the front of the vehicle. No improvement was noted.

The other shop had already changed the spark plugs and wires before giving it to us, but no improvement was found. Fuel pressure gauges were hooked up, and run under the hood to the windshield so the driver could watch it constantly while driving. In this way, he could note any changes in fuel pressure if and when the problem would occur. No fuel pressure problems were ever noted.

All tests and troubleshooting charts on modern vehicles inevitably lead to a final conclusion at the end of the chart if no other problems are found. "Substitute vehicle computer with a known good one, and retest. If this corrects the problem, replace the computer." The problem with this statement in troubleshooting charts is that to obtain a "known good computer" you have to buy one. After you buy one and break the seal on the box, you may not take it back for a refund. This is an expensive way to test, since the customer will not generally be willing to pay for a computer that he doesn't need, regardless of the fact that it is the manufacturer's officially published way of diagnosing the problem. Regardless of this, we finally got to that point and purchased a new computer. No improvement!

This vehicle had none of the classic symptoms of bad injectors, but certain tests were performed anyway. Getting back to the beginning of this story, each injector driver controls three injectors, wired in parallel, at a time. Battery voltage is always present at the injectors, and a transistor in the computer provides a ground for a few milliseconds for the injector solenoid windings when the computer wants the injector to open. Each of these transistors can handle only a certain amount of current. The injectors in this particular vehicle should have approximately twelve ohms of resistance across the coils. That means if twelve volts were applied to the injector, not counting the voltage spike when the circuits are released, Ohm's law says that about one ampere of current should flow through the injector. Since there are three injectors wired in parallel, approximately three amps should flow through the injector driver transistor when the circuit is closed (approximate round figures).

It is not easy to measure the current flow in these two circuits, so the best way to test is to measure the resistance across the two injector circuits at the wiring harness coming out from under the manifold. Three twelve ohm injectors wired in parallel should measure about four ohms with the ohmmeter, and both circuits did measure 4.3 ohms each. Since we tried every test we could think of many times each, we tried this injector resistance test many times also, always with the correct results. The problem was, the vehicle never exhibited the problem in the shop, and only did it on the road under very hot and heavy load conditions.

Eventually, we began disabling the radiator fan to intentionally make the engine overheat in the shop in an attempt to duplicate the problem, and it worked. After successfully getting the engine to begin overheating under load conditions in the shop with test equipment hooked up, we began retesting, but still found no problem. Finally, on the eighteenth day, we made this injector test for the third time under overheated conditions. On the third try, one of the injector circuits measured only one ohm! We quickly pulled the very hot manifold off the car, and individually tested the injectors on that circuit. The number two injector measured less than one ohm, and the other two still measured twelve. Ten minutes later, retesting the bad injector again revealed that it was back to 12.6 ohms, and of course, passing.

The number two injector was replaced, and the car ran happily ever after. The chances of ever finding this intermittent shorting problem (I haven't seen it before or since) were very small, and required as much luck as anything to find it. Ironically, the injector functioned properly even when shorted, but the excessively high current reaching that transistor was causing the computer to do strange things.

Fifty man hours went into this job, and we collected thirteen from the customer. I have a hard time deciding whether or not to take in jobs like this one. Should we not take them at all? In this case it is one of our very best and most loyal customers.

On the other hand, these kind of problems are not all that rare. Several jobs like this one in a row, and the shop would go broke. On this particular engine, it is common for an injector to short, but they stay shorted and are easily diagnosed. They do not have the symptoms that this particular car had. They simply have a dead miss that leads you right there after the ignition is found to be okay. Remember, this car did not have bad injector symptoms at all.

One can never collect all that time spent on such a job. Another shop might have begun immediately replacing parts, including all six injectors (when you're guessing, you always do them all) and found the problem sooner. They would have appeared to be more heroic, in spite of the fact that they would never really have known what was wrong, only that it was fixed.

Most customers could not leave there car there for three weeks either. In this case, if the shop worked hard on it for three days, and the customer demands it back even though it is not yet fixed, does he owe for that three days? If he paid for that three days after the fight, and then took it someplace else that began guessing, being lucky enough to try the injectors pretty quick, how would that make the first shop look in the eyes of the customer?


Back to Brother Bob's Home Page

Back to the index page (top level)

Copyright © 1997 by Bob Hewitt - All rights reserved