Fuel leaking into air flow sensor KE-Jetronic

[brabus3.6]

After being directed towards the fuel distributor as the possible cause of my hesitation below 2000rpm/misfiring issues (m103 engine) I noticed that the inside of my air flow meter was wet with petrol and smelt very strongly of it.
Is the leaking into the air flow sensor caused by the fuel distributor or the fuel pressure regulator? Should I still have my fuel distributor cleaned and recalibrated?

[dpetryk]

The leakage you describe is coming from the fuel distributor's control piston. Its worn out.

[saumil]

There is a serviceable gasket (from MB dealer ?) for the fuel distributor shaft, if your replace this the fuel leak will stop if it is coming from the shaft. You will need a special tool like a split screwdriver to open the screw because the shaft passes thru the screw. Also you will have to exactly note down the length of the shaft with respect to the outer frame because turning the screw that allows you to replace the gasket also alters how far the shaft drops when you hold the distributor in the right (up/down) position. Take your distributor to a dealer's part store and they may loan you the tool and you can change the gasket right there and return the tool.

[saumil]

Sorry, the measurement of the shaft should me made by keeping the shaft up and distrbibutor down. But here are the instructions.

[brabus3.6]

Thank you for your replies but a new seal and recalibration has not solved this problem. All is outlined below...
Okay, brabus has been suffering from hesitation just before around 2000rpm for a while now which clears up after it. It feels like a potential barrier which is overcome and then its fine.
I first thought it was and ignition system fault so I had the car scoped and it checked out fine according to Mercstar, with all cylinders receiving the same spark. I have tried a new coil but to no avail.
I then moved onto the duty cycle,this reads 50% as it should for a non-cat car such as mine. I also measured the EHA current which appears to enrich the mixture under quick acceleration. However, when I disconnect the coolant temperature sensor (CTS) my problem seems to magically dissapear, although the acceleration past 2000rpm isn't quite as strong (due to the lack of enrichment I guess).
I resistance tested my CTS and it checks out. I then noticed fuel in my air flow potentiometer (which I have also previously checked and readjusted to Bosch/Merc spec), and upon removing this my throttle body. So I though leaking fuel distributor.
I bought my pressure regulator, fuel distributor, air flow sensor, injectors and cold start valve to specialist Ken Mills Injection. The fuel distributor was not leaking but he changed the seal anyway and recalibrated it as it was only giving 150cc instead of 175cc. All other components checked out fine. So today I put it all back together thinking all would be okay but , this car is having none of it, same problem. Where do I go from here guys?

[saumil]

Did the air flow sensor turn out to be o.k. ? If this is a potentiometer type, there is a possibility that there is an open circuit at certain section of the pot contact. With car off, when the air flow sensor is at the top, the resistance should be between 1 and 2K. Now if you gradually depress the plate, the resistance should gradually increase, there should be no jumps greater than 4K. If for certain position of the plate, the sensor resistance is unreliable, you could get hesitation for those positions. Maybe, disconnecting the temp sensor puts your car in limp mode (this thing is car specific, and I dont know about your car) ? And in this mode, the air/fuel ratio is controlled in an open loop maner by the computer, no sensors are read, and this actually tells you that the problem is in one of the sensors. The other possibility is if the position sensor is giving unreliable signal around those rpms. I cannot upload the manual's pages that outline steps to test electrical components of the KE system, send me an email at saumil@swbell.net if you dont have these steps and need them.

[brabus3.6]

Today I decided to change the CTS since it is fairly cheap and the resistance values accross its pins were slightly different to each other. This did not change the cars behaviour at all, hesitation at 2000rpm is still there.
Thanks to saumil I have the ECU pin-out and KE system test procedures so I dabbled a little. First of all I checked the harness to air flow pot, eha and CTS for continuity, all were fine. I then checked the air flow pot as saumil recommended, at rest it is around 1.4kohm, rising quickly and then more slowly until about 9kohms before falling again so no problem there.
Now onto the EHA valve, this passed the resistance test however I never noticed any enrichment whatsoever, it was always at 0ma, KOEO was 0ma too but I believe this is correct as mine is a standard non-cat car. When I had my fuel distributor rebuilt the gentleman checked the EHA and told me it is fine and working as it should be. I can only imagine I have an ecu fault. I have a spare EHA which I performed the quarter turn clockwise trick on, I swapped them over and this was also drawing no current at all.
I noticed one other thing, with my aircleaner assembly removed when I revved the engine white smoke, smelling of oil came out of the hose which goes from the valve cover to the air cleaner. I can't remember seeing this before, but the car still runs okay/no hesitation with the CTS and thus KE-Jetronic system disconnected. Where from here???

[saumil]

The air sensor pot's resistance rising sharply is not a good sign. Infact, the jumps should be less than 2K if you are moving the plate really slowly. Did you get a larger jump than that ? And was the jump close to the zero position, because that could explain your hesitation around 2K rpm.

And you dont have a catalytic converter, but still there should be enrichment current on the EHA. Do you get 10-20 or so ma when you just turn the ignition switch to run, before starting ? Let me check the expected test results on EHA for your car again.

[saumil]

I am also assuming that measuring currents is not a problem, that you are using a series circuit to do that rather than just connecting the meter's terminals across the EHA pins.

[saumil]

I checked, every car must have greater than 15 ma when the throttle is blipped at engine temp of 20 degC. So if you dont get any change in current under this condition, the computer must be replaced if there is not an obvious wire disconnection from the computer to the EHA. What is the exact model that you have ?

[brabus3.6]

I have engine number 103.983 ruf (non-cat) car. I will buy the merc harness and retest with that. I put the ammeter in series with the circuit, read 0ma.
I had the same problem exactly with two different air flow pots so I highly doubt it is that. As you mentioned it is probably that the ECU output to the EHA valve is blown. I was driving around for while without the throttle sensor switch disconnected (unknown to me, realised when I couldn't read the duty cycle), perhaps this caused it.
The file comb1 at the top of page 12 describes my problem, minus the stalling, perhaps it is that? If not I guess the ECU as you say.

[saumil]

Terminals 10 and 12 on the computer connector are connected to the two ends of the EHA connector (do you have the diagram for that connector ?), check for continuity for each wire before buying a new harness. If these two wires show continuity, for sure the computer is not sending the current correctly. The model you described, should either have 10 or 20 ma (depending on the year) as soon as the ignition switch is put on run (position right before the start). I dont know of anything else that can cause the EHA current to be absent, maybe someone else with more experience can pitch in OTHERWISE just replace the computer. Oh, one more thing, try disconnecting the battery for 10 or so minutes, just incase the computer needs a reset.

It is unlikely, but possible that both pots are bad, you should not have a sudden jump (i.e. like an open circuit) at all for any plate position, I can check this on my car and let you know. I dont know if the pot is linear or logarithmic, would expect it to be linear. But in either case, I am not sure if this will cause the EHA current to be absent.

[saumil]

One more thing, check the continuity of the coolant temp sensor terminal (one of the two terminals) to the computer connector (terminal 21). An open circuit in this circuit can put the car in a limp mode.

[brabus3.6]

I swapped ecu's with one in a breakers yard, no difference. I then realised that I wasn't using the ammeter function correctly, I plugged the wire into the wrong port . My EHA valve is enriching the mixture on blipping the throttle. I didn't see the +15ma change but I tested the car hot.
I decided to test my engine compressions and I got:

13 bar
13 bar
13 bar
12.5 bar
13.5 bar
12 bar

Cylinder 5 & 6 are just within the mercedes 1.5 bar tolerence in their literature. The compressions are slightly higher than 300E compressions due to Brabus raising the compression ratio. Cylinder 6 is leaking slightly from the rings. I introduced a teaspoon of oil and then read 13 bars from it's compression test. However I don't think my compressions should be giving me the problems I am seeing at the moment?
The crankshaft position could also apparently be causing this? I have never seen this mentioned on any forums as being such a cure.

[saumil]

O.K., so the ECU is ruled out. It is not the compression causing this either. You can also check the air sensor pot while the car is running, this may allow you to examine if there is erratic change in sensor voltage when you get the hesitation. The same pins (middle pin and one of the two neighbouring pins, I dont remember which one but you can figure this out) you checked resistance across, now measure the voltage. Note, to do this test, the sensor connector must be conected to the pot, so you can use alligator clips to firt make all the connections. At idle, you should get between 0.5 and 1V, ideally 0.7V. Then as you smoothly press the gas pedal, the voltage should increase gradually. Notice what happens around the hesitation point. Do you see a voltage that dips down.