 |
  |
|
 |
      |
|
 |
|
|
 |
     |
|
|
|
|
|
|
|
|
|
|
|
|
VII. Fuel Injection C. Diagnosing Bosch KE-Jetronic Injection If you've done much work on European cars, you've encountered Robert Bosch K-Jetronic mechanical fuel injection, which first appeared in the early seventies. It's an efficient, multi-port, continuous (the "K" stands for that word auf Deutsch) spray system that works so well it allowed some car makers to avoid adding catalytic converters for a few years after everybody else needed them to comply with emissions regulations. Yet, it's less expensive than regular electronic port injection, better than TBI (Throttle Body Injection) because it doesn't allow stratification of the blend in the manifold, and relatively simple -- that is, once you understand it. Mechanical, but electronic, too As with just about everything else automotive, K-Jet didn't escaped change in the last decade. When three-way catalysis became necessary in order to meet NOx limits, a means of keeping the air/fuel mix closer to the ideal stoichiometric ratio (that is, 14.7:1 air to fuel by weight) than could possibly be achieved mechanically was needed. So, this injection set-up got add-on electronics to give closed-loop operation and satisfy the Feds. First, this was called K-Jetronic with Lambda (that Greek letter has come to represent the stoichiometric blend), but later came a more elegantly simple version, KE-Jetronic. You'll find it on several Volkswagen and Audi models from '84 up, even the entry-level Fox (a neat little car that, in our opinion, outclassed its competition). Since VW has made more vehicles with this system (which the company calls CIS-E for "Continuous Injection System-Electronic) than everybody else put together, I'll give you the gospel according to Wolfsburg and use the Fox as my example. Although I'm being very model-specific here, a lot of what I say will give you a better understanding of K-Jetronic in general. Flow The little Volks has a fuel supply system with a small in-tank transfer pump that sends gasoline up to a reservoir. From there, the main roller-cell electric pump (which should never be allowed to run dry) picks it up and pushes it into an accumulator, which protects the metal diaphragm in the fuel distributor from the shock of the pump turning on, and maintains residual pressure after the engine is shut down to help avoid vapor lock. After passing though a big, lifetime filter mounted on the same bracket as the pump and accumulator, the flow reaches the fuel distributor, and here's where you'll find the main differences between KE and other CIS's. As you know if you read the previous section, the basic principle of Bosch continuous injection is that a metal disc floating in a cone moves in accordance with the amount of air entering the engine, and the disc's lever is connected to a control plunger that rises in a barrel progressively uncovering fuel-metering slits. The more air being taken in, the more the disc moves, and the more gasoline is delivered. This much remains the same. Mixmaster Differential pressure valves still serve to maintain the pressure drop at the metering slits regardless of how much fuel is flowing through them, and here's where KE's closed-loop mix control comes in. An electro-magnetic differential pressure regulator (a small plastic box screwed to the side of the fuel distributor -- its presence will let you know the system is KE) opens and closes as commanded by the ECU, varying pressure in the lower chamber of the fuel distributor. When the computer gets a lean signal from the oxygen sensor, it sends additional current to the pressure regulator, increasing differential pressure, hence fuel flow, so the blend richens. And vice versa. Although the whole differential pressure versus fuel flow concept is a little hard to grasp, what I've just said is really all you need to know for diagnosis. The injectors themselves, by the way, are of the air-shrouded variety that provide a very finely atomized spray. Air flow through passages in the cylinder head really breaks up those droplets. Instead of the combo of a control pressure regulator and a pressure relief valve, KE-Jet has only a diaphragm-type pressure regulator, which maintains about 78 psi in the system by venting the excess back to the tank, and also keeps residual pressure available after the engine's shut down. A cold start valve sprays extra gas into the intake manifold when it gets voltage from the thermo-time switch, just like previous systems. Also similar is the auxiliary air regulator, which is plumbed across the throttle plate to let in extra air for fast idle. With that operational explanation taken care of, I'll look at service. As with every other modern engine control or fuel injection system, don't assume a driveability or performance problem is due to KE-Jet until you've checked for ignition problems, vacuum leaks, poor compression, etc. If it's not tampered with, this set-up should stay in calibration and last just about forever. But that's not to say you'll never run into any problems with it. Closed loop? The first thing to find out is if the closed-loop mode is ever being entered, and for that you'll need a milli-amp meter and either a special VW wiring adapter (I'd be surprised if you bought such a thing) or some home-made leads. With the ignition off, unplug the two-terminal connector from the differential pressure regulator. Those Bosch plugs can be difficult if you're not used to them -- pry off the metal clip completely with a small screwdriver, then separate the connection. Using very narrow spade terminals, crimp a male to one end of a six-inch wire, and a female to the other. Attach a similar pair of spades to the leads of your meter (I found that an ordinary analog multi-meter worked fine). Since a short here can blow the electronics, I wrapped the spades with electrical tape to make sure they wouldn't make contact with each other. Now, connect the male end of your jumper to one side of the plug, the female end to the corresponding spade of the differential pressure regulator, then attach your meter in series across the other terminals, and set it to the milli-amp scale. Start the engine, let it get warm, and make sure all electrical accessories are off. Once normal operating temp is reached (remember, any oxygen sensor has to be warm to work, so you may need to hold rpm at 2,500 or so for a while to heat it up) look for a reading of 10 mA + or - 6, and it should fluctuate over a 1-3 milli-amp range. That fluctuation is the tip-off that closed-loop operation is occurring -- in open loop, the ECU sends a fixed 10mA signal, which won't cause any drivability problems, but will increase emissions. A CO test should ideally go along with this. If you happen to access to an exhaust analyzer, pick up the sample at the convenient tube you'll find in front of the manifold (it has a blue rubber cap), make sure the timing is at 6 degrees BTDC and the idle speed is 900 rpm, then look for 0.3% to 1.2%. Altering the mixture requires drilling out a plug between the air sensor and the fuel distributor, then inserting a long metric allen wrench, but normally there'll be no reason to fool with this. Garbage in . . . If the computer isn't sending a proper signal to the differential pressure regulator, check the inputs. First, the coolant temperature sensor, which is of the NTC (Negative Temperature Coefficient) type that loses resistance as it gets hot. At 32 degrees F., there should be 6K ohms across its terminals, at 68 degrees 2.5K ohms, and at 176 degrees a mere 300 ohms. Next, measure oxygen sensor voltage, which should range between about 1/10 and one volt. A good, straightforward test is to unplug the temperature sensor connector and bridge its leads with a jumper (no resistance, so the computer will think the engine's warmed up), then separate the oxygen sensor lead from the green wire, and observe the milli-amp reading at the differential pressure regulator as before. You should get 9-11 mA. Now, ground the green wire. Within 20 seconds, expect the meter to show 19-22 mA. No? Then check the wiring from the ECU to the oxygen sensor. If that's okay, the computer's probably bad, which is an unusual, but expensive, occurrence. A complete lack of current to the differential pressure regulator won't make the car undriveable, but there may be a little roughness because the mixture will be lean. My VW Service Training contact warned me about a common cause of this electrical failure: "The 10 amp system fuse is auxiliary, mounted on the outside of the fuse and relay panel," he said. "People think it's a spare fuse and take it out to use it elsewhere. I've fixed cars that have been worked on for days simply by installing a fuse." Won't hold Hot restarting trouble can usually be traced to a loss of residual pressure in the injector lines that promotes percolation -- you'll have to crank for quite a while to purge all those bubbles. To check, attach a pressure gauge to the fuel line that goes to the cold start valve, run the engine, shut it off, and watch the needle. If it doesn't hold at about 38 psi for at least 10 minutes, the check valve in the pressure regulator is probably leaking. The KE-Jet control plunger inside the fuel distributor seats on an O-ring, so a small amount of clearance between the air sensor lever and the plunger (engine off) is required or residual pressure will be lost. If you're used to other CIS's and you notice this clearance, you may jump to the unfortunate conclusion that the sensor plate is out of adjustment and get yourself in trouble. That little bit of free-play is supposed to be there. To see if it is, make sure there's pressure in the system (you can run the fuel pump by removing the pump's electrical relay and bridging panel terminals #30 and #87 for about five seconds), then try to lift the sensor plate using a magnet on its 10 mm center bolt. You should feel somewhere between just noticeable clearance and .080 in. (that's 5/64 in., or 2mm) of movement. While you're there, you might as well raise the plate through its full range to find out if it's binding, which condition can be corrected by loosening the bolt, centering the plate so a .004 in. feeler gauge can go all the way around it, and retightening the bolt (four ft. lbs. will do, and thread locking compound is a good idea here). Coping with cold When you encounter a hard- or no-start-while-cold problem, think about the thermo-time switch and the cold start valve. The points in the switch open at 86 degrees F., so make sure the engine's below that temperature, then remove the connector from the cold start valve, ground terminal #4 of the ignition coil, and put a test light across the connector's terminals. Run the starter and the light should glow for between one and eight seconds, then wink out. To test the fuel delivery of the valve itself, remove the two Allen head screws that hold it to the intake manifold, detach the valve and aim its business end into a suitable container, energize the fuel pump as explained above, then connect a jumper from the positive battery post to one of the valve's terminals, and another jumper to ground (be careful not to make any sparks!). You should see a steady cone-shaped spray pattern. Disconnect the wires and no drops should appear at the valve tip for at least a minute. If the powerplant tends to stall when cold because there's no fast idle, examine the auxiliary air regulator. With the coolant below 80 degrees F., unplug the regulator's electrical connector, start the engine and let it idle. Pinch either hose that's routed to the regulator and idle speed should decrease slightly. Let the engine reach normal operating temperature, plug the connector back in, and pinch again. You should hear no change in rpm. In cases where fast idle never kicks out, see that current is reaching the heating element in the regulator (that's what warms the bimetal and shuts off the flow of air) by using a test light across the connector terminals with the key on. No light means there's an open somewhere between the fuel pump relay and the plug.
|