Mercedes-Benz Forum banner

Items responsible for adjusting duty cycle?

20K views 52 replies 11 participants last post by  Stretch 
#1 · (Edited)
Hi folks,

What are the components responsible for duty cycle?

I ask because my mechanic tried to adjust duty cycle on my 190E 2.6 (1989) but the needle stayed at 0.2 mA with no fluctuation. Rev up the engine and it would go to 0.4 mA then slowly go back to 0.2 mA, and stayed there like before.

Additional symptoms car is having: hesitant starting during cold starts, but when engine is warm it starts up right away. Idle is hunting when warm.

We checked oxygen sensor and compared against a new one, with no change in duty cycle readings. So we put back my old oxygen sensor.

Found an open vacuum line by the mass air flow, going to the check valve for emissions - solved the fuel smell. But did not have impact on duty cycle readings.

Checked fuel pressure regulator vacuum hose - is it supposed to hold vacuum? We applied vacuum towards the rubber hose coming from fuel pressure regulator, and it did not want to hold vacuum.

Pulled plug for idle control valve during idling, and RPM went to 1100 RPM. Plugged idle control valve and engine died. My previous readings on the forum state idle should go as high as 2000 RPM. Is my idle control valve suspect? Vacuum hoses going to and from this unit are in good condition (replaced with new in 2010). Current one is a spare I had lying around, cleaned the inside and installed with no change. Still have old idle control valve.

Fuel accumulator (Bosch), fuel pumps & check valves (Bosch), and fuel filter (Bosch) are all new (including main fuel hose to tank). OVP (KAE) is new. Fuel pump relay (KAE) is new. EHA (Bosch) is new. CIS-E temp sensor (4-prong) is new (Beru).

What else should I look at? :confused:
 
See less See more
#2 · (Edited)
I immediately thought OVP relay but you've done that (checked the fuse? ya never know...). If your problem is only a hard cold start then I'd stick with the basics first. How are things under your distributor cap, no condensation? Plugs and wires good? You've probably checked all those already, but I had to ask. Is your EGR valve stuck slightly open?
Any other symptoms at all? If not, the next time you cold start, run some propane into your intake while someone else cranks it over. If that makes it start faster, and smooth out sooner then we will have a direction in which to go.
Don't start the propane flow until the person is ready to crank the engine, and you might want to feed the propane directly into the intake area of the metering valve, not into the air cleaner. I find that an unignited hand held propane torch provides the right flow.
 
#3 ·
Quoted from Megawatt Man on MB World

The entire information below is not by me, it is from Megawatt Man, and I quote it verbatim. Perhaps something there may help you or lead you to the answer.

"KE Jetronic Diagnosis

Hello all on this forum. I hope to bring to some the benefits of my experience in effecting a remedy to a faulty KE Jetronic system, on a W124 Mercedes 300CE, Australian version. Not that the latter factor makes much difference, seemingly apart from the Aussie version not including a cold air input temperature sensor. It took me a long time to get there and I hope that the following information will be found of help to people with injection system problems.

It can be very daunting coming to terms with a faulty system. You have a great car but it runs disgustingly, if at all. And the solutions are costly and it seems that nobody in the world apart from a few competent service providers knows anything about it. But not only that, it seems that you need special diagnostic tools that cost the earth. When you look through the internet you find dozens of people who have paid large sums to have their systems serviced to no effect.

But here are solutions to several frequent problems that don't involve diagnostic tools much beyond a multimeter.

Before starting anything, please refer to a tutorial about how the system functions, Bosch have an excellent summary readily searched out on the internet. Search for “Bosch KE Jetronic system function” or similar.

It's likely that your symptoms include “won't start when cold or hot; backfiring; cough and splutter on acceleration”.

You will have found from your searching that a most important component is the “Over Voltage Protection Relay” and it is most important!

But there are some tricky things to check out first, before getting into the complicated parts.

First. Note that all the sensors have a return path to ground. Some have the benefit of a wire from the sensor, while others rely on current paths via the engine block and chassis. This latter path can be fraught! The engine block is of course mounted on compound engine mounts which are not electrically conductive. So there is a flexible copper conductor run from the block to the body. On my car this is located between the manifold and a mounting bracket for the ignition coil. It's flexible because there is relative motion between the engine and the body every time the engine speed changes. And because it's not ideally flexible, some reaction occurs on the attachment points at each end and they can become lose. Result? Poor hot or cold starting; coughing and spluttering.... you get the idea. To find these connections you need to take off the air cleaner, find the connecting cable and look. Check each end for tightness, broken cables etc. If they are loose, complete the job by removing the attachment, thoroughly clean the terminals and tighten them up. Then see how your vehicle behaves. In a few cases a miraculous transformation will have occurred. But the probability is not high, especially if the vehicle has a few km or miles on its clock – the electronics do not have an indefinite life!

Now precisely the same effect would occur if at the other end of the chain of supply was affected – that is the battery supply. This goes to the injection control unit, or computer if you like. This is located behind a lovely plastic moulding immediately in front of the battery. Carefully remove it by inserting a finger in between the bodywork and the moulding and you'll work it out from there. Now great care must be used in following the next check. You'll note that there is a large cable entering a plug that in turn is mounted on top of the computer and it's about 150 mm long. To remove the plug it is necessary to deflect the two horns immediately beneath the cable entering the plug towards the computer body. This will release the plug and allow it to be withdrawn. Very carefully lift the plug from its cable end, precisely in line with its attached position. The end of the plug furthest from the cable has a rectangular tongue located in a slot in the computer's body and this is the last part to come free. Of course it is the first part to position when replacing the plug. And it's here that we must observe absolute care. You'll see that the computer has 25 flat pins that each have to enter a connector slot in the plug and these connector slots are fragile. The first connector is say 25 mm from the tongue and as I said above, the plug is about 150 mm long. So in replacing the plug there is a mechanical advantage of about 6 at the point where the first connector pin enters its fragile socket. On my car the first slot had been damaged due to the plug not having been first positioned so the tongue fully engaged its slot and then possibly the plug was pressed into place in such a way that it was not completely parallel with the socket. It took a fair bit of care to realign the connector so that a good contact was effected, using a tiny jewellers screwdriver, but it was done. It did make a difference.

But those things are almost outside the realm of the KE Jetronic system itself and I'll go on to that now.

Fuel Pressures
When the key is turned on, the fuel pump operates for one second, bringing fuel in the system up to operating levels. Upon engine start, the fuel pump is turned on again. Serious running problems can be experienced if fuel pressures are too low. If you are fortunate enough to have access to fuel pressure measuring devices, it's great to prove the pressures right, this can save a lot of time.

If your vehicle will not start at all, a faulty fuel pump or electrical supply can be suspected. It's not much use turning on the key and then crawling under the rear of the vehicle to check whether the pump is operating, because after the first one second, it should have turned off anyway. That is a two person operation, one turning on the key, one listening for pump operation - for one second only!

There's a fuel filter in the line, if it is blocked you'll have poor fuel supply and erratic running, so it will need to be changed. There's another component in the fuel line called the Accumulator. This is just a spring loaded reservoir that maintains fuel pressure during times of low fuel volume requirements, when the pump may be turned off, in the same way as at starting. A symptom of a faulty Accumulator is stalling when idling, say at traffic lights. The acid test for this is to take off the air cleaner to expose the air flow sensor plate. Start the vehicle and run it just for a short time, then shut it down. After about one minute, depress the air flow sensor plate. It should offer resistance to your effort, indicating that residual fuel pressure is being held. If the plate moves downward without resistance, you need a new accumulator.

Diagnosis by Codes
The system is provided with a diagnostic socket from which electrical readings are made that show the condition of the various components. On my car there is an 8 pin socket, concealed by a screw cap, located just above the ignition coil on the inner mudguard. Readings for the injection system are taken from pins 2 and 3 of the socket. Other cars have other styles and sizes of sockets, but there seems to be some commonality in that pins 2 and 3 are used for this purpose. You'll need to determine this from on-line searches, specifically for your model.

But before these codes can be read reliably, the condition of the Over Voltage Protection (OVP) Relay needs to be checked. This relay performs the valuable task of making sure that any short term voltage spikes do not burn out sensitive electronic parts throughout the vehicle, particularly in the computers. Reports abound of OVP relays with failed soldered joint on their printed circuit boards and if these are dodgy, the supply voltage to the injection system components can be intermittently interrupted, producing the same symptoms as the faulty ground or battery connections mentioned above. The OVP relay, while protecting the electronics, turns on the fuel pump relay and other components that prepare the injection system for operation. So a faulty supply from this relay can give symptoms of horrible running, while masking other faults that may be present. You might be able to roughly gauge whether the OVP relay is OK by measuring the voltage at the fuel pump relay while giving the OVP relay a bump. A voltage loss is a sure indication of a problem OVP, but such a method is not reliable. If you don't want to start out with a new or known to be good used unit, you can follow the diagnostic steps, but if you receive variable results, change it out.

These electrical readings are codes, each of which indicates something. There are two forms of these codes. They consist of a set of pulses output by the computer, rising in voltage from zero to the injection system voltage (this can be just a bit less that battery voltage, , depending upon engine operating speed) and the number pulses indicates the nature of the fault. If you don't have access to the proper gizmo to measure these pulses, you can make a pulse counter from a light emitting diode with appropriate resistor and a three connecting leads. Making the test leads is covered in many places on the internet, as is using them to get the computer to output the diagnostic codes. Once connected, you instruct the computer to output one code at a time, then count the number of pulses, then look up a table from Mr Benz to identify the faulty part. My vehicle is a 1988 model and I was not able to find the list, but I reckon that almost everybody else's models are covered by information on many sites. There's some good news though. Alternative checks can be made just using a multimeter on pins 2 and 3 of the diagnostic socket. In one form of multimeter test, the average value of the pulse train voltage is measured and in another form, the on/off ratio of the pulse train is measured. In the following, the average voltage method is described.

Test 1
Average voltage method. Turn on the key, but don't start the engine. Measure the DC voltage between pins 2 and 3. This should be 30% of the injection system voltage as measured usually at pin 6 of the diagnostic socket – it's a little less than battery voltage under the conditions of the test, ie whether the engine is running or not) and correspond to Mercedes' specification of a 70% duty cycle or “on/off ratio”. In fact Mercedes uses the total proportion of its off period as its indicator number, but the multimeter respond to the proportion of time the pulses are on, at injection system voltage. Hence the 30% mention above – 100% minus 70% equals 30%. This initial reading doesn't tell you much, only that the computer is capable of putting out diagnostic codes – but that's something! If you don't receive any readings, maybe your computer is past its prime.

Test 2
Same test conditions
Deflect the air flow sensor plate. The duty cycle should decrease to 10%, that is, the multimeter should read about 11 volts. If no change occurs, that is the reading still is about 3 volts, test the throttle valve switch (Job 07.3-121 in Mercedes shop manuals). Please note that the 10% is nominal, 9% to 11% is seemingly OK.

Test 3
Same test conditions
Deflect the throttle fully, the duty cycle should read 20% or about 10 volts, indicating that the Full Load Contact switch is OK, unless it decreases to only 40% or about 7 volts. In this case test the air flow sensor potentiometer (Job 07.3-121 in the Mercedes shop manual). Again 20% or close to it is OK.
[cont'd]
 
#4 ·
[cont'd from above post]
Test 4
Engine operating at 80 degrees C. Check the Lambda Ratio. The Lambda Ratio is the ideal ratio of air/fuel that enables complete combustion of the fuel. It is 14.7/1. The controlling element is the Oxygen sensor in the exhaust gas stream. This sensor operates only at high temperature, so is provided with a heater supplied by the battery, via the fuel pump relay. In operation, it determines the proportion of oxygen present in the exhaust gas and signals the computer to provide more or less fuel to maintain an average value corresponding to the ideal Lambda ratio.
Measure the voltage between pins 2 and 3. It should fluctuate, showing that the oxygen sensor is causing the computer to adjust the mixture on a continuous basis. The duty cycle readings should be between 45% and 55 % . Higher than this indicates a lean mixture and lower indicates rich. The voltage levels corresponding to these values are, depending on the battery voltage while running are 7.4 to 6.1. (These values assume a system voltage of 13.5.)

If there is no fluctuation at all, just a steady reading, something is wrong with the oxygen sensor or its supply. It is located not far along the exhaust pipe and replacing it involves a big spanner and a lot of effort. Its wiring is located under the floor mats on the side of the vehicle above the exhaust system.

My vehicle presented readings indicating 55% to 65%, so running was lean. The symptoms were sluggish pickup, coughing and sputtering as well as stalling. The remedy involved adjusting the mixture screw in the injection distributor, but we'll discuss this later.

Air Flow Sensor Potentiometer
This is one of the very important pieces of the system. It is operated by the air flow sensor, the circular metal plate that deflects under the influence of the air flowing into the engine, which in turn is determined by the throttle opening. It is a fair bit more complex than a volume control on your stereo, but it does the same thing. As the air flow sensor is deflected by the air stream, the shaft upon which it rotates moves sliding contacts over a track of resistive material and this feeds the computer with a voltage reading which allows it to control the precise amount of fuel required by the measured air flow. Now every time the throttle is operated, the air flow sensor deflects and moves the sliding contacts across the resistance material and that can happen a few million times, I guess, in the lifetime of a car. So the track wears and the computer can then be fed with a certain voltage at one throttle opening and with just a very small movement the sliding contacts encounter an area where no resistance material remains, so the voltage indicating air flow drops to zero. The computer doesn't know what to do, so erratic operation is the result.

Testing this component is carried out with ignition on and engine off. The potentiometer is connected by a three pin plug, with locking tags at top and bottom. Ease these out and slightly withdraw the plug, so that the socket pins can be accessed by test leads from your multimeter. The top and bottom pins should have nominally 5 volts across them, as supplied by the computer. The centre pin feeds back to the computer a voltage less than 5 volts for each position of the air flow meter. Connect the multimeter between the centre and the top pin and watch the voltage reading as you very gradually deflect the air flow meter plate. The voltage should increase continually, never dropping out to zero. If it does, you need a new potentiometer. Mercedes only sells these already mounted on a replacement air flow meter and you may think that its cost is comparable to the value of your vehicle. There is information on the web about obtaining a replacement potentiometer at a much lower cost. If you go this way, you'll have to remove the old pot and replace it with a new one, then calibrate it, but that's a simple matter really. One reason for which Mercedes provides an all new solution may be that the sliders that run over the resistance tracks remain in the housing and they may also be worn, so installing a new pot may not produce any improvement. Removing the old potentiometer involves using a very fine screwdriver to insert in the gap between the two roughly circular plastic covers on the front at either end of the potentiometer housing and easing them out, exposing four Torx screws that mount the potentiometer to the air flow sensor body. At this stage it's a really good idea to put a pencil mark around the pot on the air flow sensor body, as an aid in getting the replacement one roughly in the right position. Undo the screws and gently withdraw the pot from its shaft. Look at the tracks and marvel at how the car operated at all! You'll note that the four mounting screws go through slotted holes in the pot housing that allow it to be rotated relative to its operating shaft and this is how the pot is calibrated. Once you have replaced the pot with the screws just loose enough to allow its rotation around the shaft, you reconnect the three pin plug, again with the socket pins exposed for measurement, then measure the voltage at the centre pin. Turn the new pot around the operating shaft until you read about 1 volt. At the 1 volt level the engine will be running rich, but if you try a lower initial setting, you may not be able to start the engine. Now start the engine and get it up to operating temperature, then adjust the pot position until you read 0.7 volts between ground and the centre pin. The pot is now calibrated.

If that was the main problem, your vehicle should now be much improved.

Mixture Adjustment
There's a lot of instruction on the web cautioning any playing around with the mixture control. The screw that allows adjustment is protected by various types of guards against ready access, so Mercedes takes it seriously. But it isn't too much of a problem, the biggest part of the task is removing the guards. Consult the web for details of your particular model. The adjustment consists of using a 3 mm Allen key in a spring loaded coupling to turn an adjustment right inside the injector housing. The screw raises or lowers a plate that changes the aperture through which fuel is dispensed to the injectors. It is essential that you don't leave the Allen key in place after the adjustment is made, because it can cause serious damage to the injector system next time you start.

That having been said, here's how. Bring the engine up to normal operating temperature, about 80 degrees C. Then connect the multimeter to terminals 2 and 3 and adjust the screw until the voltage readings fluctuate between 6 and 7.4. Make adjustments almost microscopically, the tiniest turn will have a significant effect. Allow at least 10 seconds between adjustment and reading, to allow the system to settle into its new operating condition. Those voltage readings correspond to a Lambda ratio of between 45% and 55%, just what you want.

The Cold Start Valve
There is a cold start valve, in the form of another injector that supplies additional fuel when starting from cold. If it or its wiring are faulty, cold starting can be awful. The injector is turned on by a solenoid operated valve that receives its supply from the computer. To check the electrical operation, unclip the stainless steel retainer on the electrical connector and withdraw it. Measure the coil for continuity, you should read about 10 ohms. If that's OK, you need to know if you are receiving an adequate supply to the terminals. The most reliable way of doing that is to measure the current flow through the solenoid, so you have to jury rig a pair of test leads and set your multimeter to read DC amps. Expect 1.2 amps. If all that seems OK but the symptoms persist, you'll need to remove the cold start valve and watch it when simulating a cold start. Put the valve in a jar while connected to its cable and turn on the ignition. If you are not rewarded by a nice spray into the jar, the valve is shot. If you are so rewarded, you'll know why it is necessary to put the valve in a jar. A big one preferably.

Another problem that can arise with these valves is continuous operation, so that the vehicle runs rich, fuel costs rise, spark plugs are sooty and the engine oil is being washed off the bores! Get a new one.

The Coolant Temperature Sensor
The function of the coolant temperature sensor is to tell the controller the current engine temperature. When the engine is cold, the resistance of the sensor is around 2,500 ohms and an appropriate signal is given to the controller to cause mixture enrichment for starting. As the engine warms up to operating temperature, the resistance of this sensor falls to about 350 ohms, telling the controller that no enrichment is necessary. Between cold and hot engine temperatures, the degree of enrichment is progressively reduced.

This component seems fairly forgiving. The main problem for my engine was difficult hot starting. But my coolant temperature sensor was non-functional, so I temporarily wired in a 350 ohm resistor, thus providing advice to the controller that the engine was permanently hot. I had no trouble starting the engine when it was cold in warmer weather, but when ambient temperatures fell to below 15 degrees C, cold starting became a little troublesome. Maybe it would stall immediately upon starting and sometimes it took me three attempts before all was well. But if it does not meet specifications fit a new one.

The Electro-Hydraulic Actuator (EHA)
This is an important (and expensive!) part of the system. It is attached to the rear of the fuel distributor and has the function of opening and closing the fuel metering valves inside the distributor in accordance with the demands of the engine and the instructions from the computer. It is basically an electromagnet and the degree of opening and closing of the fuel metering valves is determined by the strength of the electro-magnetic force it develops, in turn determined by the current flow through its coil.

Consultation of the specifications for current flows through the coil shows that for various models, quite different currents are needed. But commonly, the maximum current flow is 15 milliamps in response to a throttle blip, while at idling, a steady current of about 1 milliamp is normal. It is stressed here that the currents are completely dependent upon which model your are working with, so look carefully at the specifications.

The coil has a resistance of about 20 ohms and you can check this by removing the connector and applying your multimeter. In order to persuade a current of 15 milliamps to flow in the coil, the applied voltage has to be about 0.3 volts. Now if you use a digital type multimeter to measure the voltage output from the computer with the EHA disconnected you may read as much as 4 volts. Ignore that. It's the result of using a very sensitive instrument while the computer is not in its normal operating state. When it is in that state, it will be putting out voltages only up to a maximum of 0.3 volts. What I'm getting at here is that the only valid readings to take relevant to the EHA are those of current. So you need to make up a set of test leads that will let you connect your multimeter in the EHA circuit. Then measure the current flows. No flow at all means that you have an open circuit coil, wiring or connector.

Final Remarks
All the best with your fault finding!

Megawatt Man"[/I]

The entire quote above is copied and pasted from MBWorld. None of it is mine.
 
#5 ·
At your cold start after overnight ?
Hesitate starting ? : ( Firing off and on prior full running while key is press)) or takes longer to start ? If it is then prime suspect is fuel regulator not hold pressure during long layoff and it supposes to hold vacuum .
Change this one and cold start valve and having say that : I had the same problem and it solved . Good luck bud...:cool::thumbsup:
 
#15 ·
Where did you purchase the cold start valve? How do you replace it? Yes my FPR (fuel pressure regulator, not fuel pump relay) is not holding vacuum.

What symptoms did you have when you had the problem? Mine starts without accelerator input intermittently. This morning it started up "okay" almost immediately with no right foot input. On my way home it didn't start until after 5 seconds, still with no right foot input.

As I mentioned earlier in the previous post I tested it and shut it off. Restart was immediate, like normal. :confused:

This from Identifix:
Vehicle:
1987 - 1993 Mercedes-Benz 190E 2.6
1985 - 1993 Mercedes-Benz 190E 2.3
More...
Symptom: Engine will have excessive crank time and is hard to start without manipulating the throttle. It may also have a hesitation on acceleration. Fuel pressure and volume is to specifications and overvoltage protection relay if applicable has been tested ok. Mixture adjustment has been reset correctly but symptom persists.
System: Emissions/PCM/Fuel
Problem:
The rest position of the airflow sensor plate is incorrect.
Test & Fix:
With the air cleaner removed, verify that fuel system rest pressure exists in the fuel system. Visually inspect the rest position of the airflow sensor plate engine not running. There should be no gap visible. The airflow sensor housing is tapered from the air cleaner seat down to the area where the sensor plate sets, then underneath the sensor plate the cone flares out again. A narrow band where the sensor plate is supposed to rest is vertical, and the edge of the sensor plate furthest from the fuel distributor should align with the upper corner of that band. If it is too high, (visible gap) the adjustment pin is located in the top of the airflow sensor plate housing to the side of the hole for the mixture adjustment screw or to the side of the tower for mixture adjustment screw. Gently tap on the pin with a pin punch and hammer to adjust the sensor plate rest height to specification. If the pin is hammered too far (sensor plate too low) the airflow sensor housing will have to be disassembled to tap the pin back up from underneath. The mixture screw basic setting will need to be readjusted after changing the airflow sensor plate height.

The entire above quote is from Identifix. None of it is mine.
This seems similar except I don't have hesitation at speed. In fact it's running so well at speed and once started. It's just getting there (starting the car) without any hesitation that's the problem.
 
#6 ·
This from Identifix:
Vehicle:
1987 - 1993 Mercedes-Benz 190E 2.6
1985 - 1993 Mercedes-Benz 190E 2.3
More...
Symptom: Engine will have excessive crank time and is hard to start without manipulating the throttle. It may also have a hesitation on acceleration. Fuel pressure and volume is to specifications and overvoltage protection relay if applicable has been tested ok. Mixture adjustment has been reset correctly but symptom persists.
System: Emissions/PCM/Fuel
Problem:
The rest position of the airflow sensor plate is incorrect.
Test & Fix:
With the air cleaner removed, verify that fuel system rest pressure exists in the fuel system. Visually inspect the rest position of the airflow sensor plate engine not running. There should be no gap visible. The airflow sensor housing is tapered from the air cleaner seat down to the area where the sensor plate sets, then underneath the sensor plate the cone flares out again. A narrow band where the sensor plate is supposed to rest is vertical, and the edge of the sensor plate furthest from the fuel distributor should align with the upper corner of that band. If it is too high, (visible gap) the adjustment pin is located in the top of the airflow sensor plate housing to the side of the hole for the mixture adjustment screw or to the side of the tower for mixture adjustment screw. Gently tap on the pin with a pin punch and hammer to adjust the sensor plate rest height to specification. If the pin is hammered too far (sensor plate too low) the airflow sensor housing will have to be disassembled to tap the pin back up from underneath. The mixture screw basic setting will need to be readjusted after changing the airflow sensor plate height.

The entire above quote is from Identifix. None of it is mine.
 
#8 ·
Not sure if we've helped MBeige at all, his question specifically referenced duty cycle contributors. I know what the term "duty cycle" means (the percentage of any given time frame in which a device receives an 'on' signal), but am unsure of which particular duty cycle he was referring to. I have a feeling we may refer to it by another name such as switching, but I'm guessing.
Not knowing the answer to his question I decided that going directly after the symptoms was a good work-around.:surrender:
 
#9 ·
Let me get back to you guys in a bit, I'll read the responses and report back on your questions. In the mean time:

Duty cycle - sorry it was a rushed post - read many threads and probably got the verbiage mixed up. What I meant was adjusting the duty cycle using the tower on the air flow meter (maybe 3mm allen you push down to adjust in small increments at a time after pressing LED button to set diagnostic mode). It used to fluctuate between 0.4 and 0.5 mA but now it's running very rich at 0.2 mA and the needle stays there.

My car has no EGR valve (1989).

I wanted to also clarify, I am having two problems:

1. Cannot adjust duty cycle (inquiring on which parts to look at to check)
2. Hard cold start

For #2 I'm leaning towards fuel enrichment at start up but I may be chasing ghosts so I'll do more readings per your well-appreciated responses.

Thanks!
 
#11 · (Edited)
Let me get back to you guys in a bit, I'll read the responses and report back on your questions. In the mean time:

Duty cycle - sorry it was a rushed post - read many threads and probably got the verbiage mixed up. What I meant was adjusting the duty cycle using the tower on the air flow meter (maybe 3mm allen you push down to adjust in small increments at a time after pressing LED button to set diagnostic mode). It used to fluctuate between 0.4 and 0.5 mA but now it's running very rich at 0.2 mA and the needle stays there.
[snip]
Please confirm you meant rotate, and not "push down". Also, I have some fancy equipment which means I don't have to always think and so I have to ask about the mA measurements. You don't measure frequency or voltage, just current?
 
#10 ·
MB, here's some technical information most of which you may be well familiar with. A reading of pages 25-30, and then whatever interests you up to page 42 might help you resolve your two problems. My gut feel is that both problems have the same source and fixing one will fix the other.
Bosch K Jetronic Fuel Injection Manual
Do you have a heated O2 sensor? Sometimes O2 sensors short internally, even non-heated types and trick the module receiving the signal into thinking a bad signal is a good one. The module then makes the wrong decision about what to do and commands
fuel injection excessively rich or lean. Since it thinks it's getting a good signal no "Check Engine" light is set.
OBD2 has feedback loops which prevent a lot of similar bad decisions, but we have (very) basic OBD.
 
#12 · (Edited)
We tested three (3) oxygen sensors. The current unit in my car, an old oxygen sensor from my car and a new Denso unit - all 3-prong which is the correct type for my car. None of them had a change in the readings. This is the puzzling part, because my mechanic suspected the oxygen sensor, and yet when we replaced it with a new unit, nothing changed.

Been reading about oxygen sensor power supply - seems to come from the fuel pump relay and is heated as long as the engine is on. The fuel pump relay on my car is new, KAE installed last year.

Please confirm you meant rotate, and not "push down". Also, I have some fancy equipment which means I don't have to always think and so I have to ask about the mA measurements. You don't measure frequency or voltage, just current?
You're right, I forgot to indicate that it is rotated after pushing down. We only measured current, yes.

This is the tool my mechanic uses to adjust duty cycle. Plugs directly onto the X11 port.

 
#13 ·
Even if the heated portion of the sensor fails, the O2 should come alive within 5 minutes once it heats up from normal exhaust flow. It starts working at about 600F.
I believe your mechanic was on the right track suspecting the O2 sensor but got fooled when he tried known good units. Remember, the sensors output has to be received and then acted upon by another device. Once that device decides what to do it has to send a signal to a slave device to vary its operation.
Take a look at page 32 in Bosch K Jetronic Fuel Injection Manual You'll see that the "controller" has to make the right decision based on information from the O2 (Lambda) sensor. In the diagram the controller signal goes to a frequency valve, but in our case I believe it is the EHA valve. If that valves misbehaves -because it is told to by the controller- you can adjust until the cows come home but you'll never adjust enough.
 
#14 ·
So which component receives input from the O2 sensor? The EHA is new.

In fact, when I drive the car now after fixing the vacuum leak it is no longer hunting at idle. However, on my way home I had the same symptoms.

1. Get in and crank engine, keeps cranking maybe 5 seconds then pathetically sputters to life. Once it's running, it's smooth.

2. As a test, I shut it off and attempted to restart. The restart was instantaneous - I would call it similar to a normal start.

My guess is the cold start issues is an effect of the non-adjustable duty cycle. So I need to figure out how to fix the problem first, then adjust the mixture.

Do you think it's the O2 sensor?
 
#16 ·
#17 ·
Okay, I replaced the fuel pump relay (2-year old KAE) with a good used one the used to run on. I did this because:

1. Fuel pump relay controls the oxygen sensor
2. Fuel pump relay is responsible for priming the system at start up.

My mechanic reported back to say the fuel pressure regulator on both his cars holds vacuum! Now I need a good used one - spending $200 on a new FPR and $266 on an EHA is emptying my pockets faster than I can fill them in. :eek:
 
#18 ·
So it's been at least 3 hours of the car sitting with engine off. Went out to try it out and with ONE attempt, the car started NORMALLY.

I shut it off and re-started. SAME.

Will try it again in the morning. Seems pretty conclusive to me it was the fuel pump relay. The good used one made in Hungary (probably from MB) seems to have cured the problem.

Let you guys know... then if it starts instantly, I will run by my mechanic to have the duty cycle checked, hopefully this time the oxygen sensor is now showing on the needle fluctuations as it did before.
 
#19 ·
Glad my post about the FPR worked for you. I remember opening up an old MB relay and a KAE relay and comparing them side by side. The MB relay looked a lot better built.
When I ordered a replacement one, I got it for $150 from PelicanParts. The new one I have is made in Hungary as well.
 
#20 ·
Well when I started the car this morning, it started with a *very* slight hesitation, but overall it was closer to a normal start than before.

I think this fuel pump relay replacement might have resolved this 50%. The other might be caused by the fuel pressure regulator not holding vacuum. But at $200 for a new I think I'll ask my mechanic to find me a good used one instead.

I've yet to check if the oxygen sensor is now working due to the fuel pump relay replacement, but I'll observe the car for a few days with the spare fuel pump relay I just installed and see what happens.
 
#21 ·
So I made a response to the wrong thread :eek:

Below should have been included in this thread:

Car started having RPMs fluctuate again. But it started "normally" this morning - albeit the first time in several attempts over the past few days - some were okay starts, hesitant but I can tell it wants to start.

Apparently the oxygen sensor we tested might not be the correct type. We tested using a DENSO unit with blue sheath - also 3-wire with correct plug. This item is similar to the Bosch counterpart for a '93 2.6 - even with the same plug. Not sure if there are huge/significant changes between the years despite having the same plugs - any insight?

Even if we tested with the used one with no change, it's possible the used one is also bad. So I ordered the specific one for my car (Bosch 13915) to see if this resolves the problem. Bosch 13925 has the wrong plug.

Your Parts Search Returned 0 Part(s) System&gid=5132@Oxygen Sensor

Lately it's been warmer, and the car has been starting "normally" :confused:

My mechanic was busy this morning with my sister's car so I wasn't able to drive over and have it checked. Got the new O2 sensor waiting to be installed.

But it intermittently dips below 500RPM when coming to a stop, making me fear it will stall, but it doesn't. If I give it a quick rev it will "wake up" and realize it's snoozing off and will idle normally again. This car is really testing my patience.

At least I can just jump into my 300D and go with no problems.
 
#22 ·
Probably not the Oxy Sensor

You've got a bit of a puzzler there MBeige, but I don't think the problem lies with your O2 sensor. When your car first starts all the mechanisms go off pre-sets. The info feed from the O2 sensor doesn't come into play for a few minutes, not until the tip of the sensor is at 600F or more.
If you have a 3 wire sensor then it has a heated tip and will come into play within a few minutes, unlike non-heated sensors which can take 10 minutes or more. My point is that the initial start will be independent of the O2 sensor. You can prove it to yourself by disconnecting the sensor and then doing a cold start. Any different?
Now that it's disconnected, leave it disconnected and keep running the engine until it's warm and then do a test drive. Don't worry, you will not harm your engine or your cat conv in any way. How did it drive?
When the sensor is disconnected the engine will run on preprogrammed values. How it runs will give me a hint as to where your problems lie. Frankly, I think the fault is in the idle control system, but where exactly I haven't a clue. Yet.

In regard to the sensor specific questions you had. I stock about 40 different sensors, both O2 sensors and air:fuel sensors. People tend to use the generic term oxygen sensor for both though the info from an a:f sensor is much more finite. An O2 sensor basically screams RICH! or LEAN! and ping pongs the fuel management system. An a:f sensor reports precisely how rich or lean the exhaust is. Anyway, I digress. All O2 sensors work precisely the same way and report precisely the same range of information. How they are designed and how they are wired varies and this is why there are so many different part numbers.

I have cut off mismatched connectors from one OEM's sensors and spliced it to another OEM's sensor just to get a car on the road. Scans of the data streams have shown the sensors to be functioning properly. So, will the 'wrong' sensor work? Probably, as long as the wires aren't crossed!

If you have a three wire sensor then one is delivering high amperage 12V to get the thermocouple cranking out heat. The thermocouple grounds through the sensor shell. Nowadays all sensors are 4 wire and ground through the computer. The other two wires are involved with carrying the voltage signal generated by the sensor back to the computer. Lots can go wrong.

Wires short out and voltage goes where it isn't supposed to, or grounds short and don't complete circuits, or prevent signals from continuing. The sensor is sending a signal of mostly 200 to 800 millivolts, 500 millivolts being the centerpoint. One side of 500 mv denotes rich, the other is lean.

Assume for a second that your engine is 'lean'. Your computer should bias fuel injection towards 'rich' and back and forth it goes. What if the sensor tells the computer that a rich engine is lean? Then the computer commands fuel injection of an already lean engine to become even more lean. You can imagine that that does not enhance engine smoothness, to say the least.

Fortunately there are limits to how far a computer is allowed to adjust the engine, otherwise it could cause damage or shut it down. However, even within the limits wrong commands can be obvious to the observant driver.
 
#23 ·
I don't think the oxygen sensor is related to the odd starting issue, but I threw it in here because my car is experiencing two problems:

1. Inability to adjust duty cycle at idle. I forgot to mention that if the car is revved, the needle fluctuates back and forth like it should. But if engine rev is lowered back to idle, fluctuation STOPS.

2. Odd starts and idle dips when coming to a stop. Recently it has been warmer, and the car has started "normally" like it previously did. This is why I'm thinking of a fuel enrichment issue when it's cold - meaning when the ambient temps are colder than usual.
 
#25 · (Edited)
To me that pretty much eliminates air:fuel ratio problems, virtually blessing the entire injection system. I'm inclined to take another look at idle management.

You rolled up the carpet? Um, aren't we talking about the 190E? Isn't the O2 connector on the passenger side (right side) mini-bulkhead? Edit, no, you're right, I'm thinking of the 2.3. Much easier to access and change.
 
#26 ·
Shouldn't it be the other way around?

If the O2 sensor was unplugged and there was no change, then essentially the system appears to not be needing the O2 sensor. To me it sounds like the O2 sensor is suspect.

If I replaced the O2 sensor with another used one like I did and there was also no change, like what happened, then it's either that replacement O2 sensor is in the same condition as the original, and that the original may also be bad, no?

It seems to me that the system should respond to a disconnected O2 sensor if the O2 sensor was working well. I'd like to see your comments on that.

Additionally, what should I expect if the O2 sensor was disconnected, if it was the O2 sensor that was really the culprit?
 
#28 ·
Shouldn't it be the other way around?

If the O2 sensor was unplugged and there was no change, then essentially the system appears to not be needing the O2 sensor. To me it sounds like the O2 sensor is suspect.

If I replaced the O2 sensor with another used one like I did and there was also no change, like what happened, then it's either that replacement O2 sensor is in the same condition as the original, and that the original may also be bad, no?

It seems to me that the system should respond to a disconnected O2 sensor if the O2 sensor was working well. I'd like to see your comments on that.

Additionally, what should I expect if the O2 sensor was disconnected, if it was the O2 sensor that was really the culprit
?
I'm fairly certain that both you and I have KE Jetronic, the main improvement over K Jet being that we have a closed loop Lambda system (using an O2 sensor) which via the EHA precisely adjusts fuel quantity to the injectors. [For simplicity at this time I am overlooking input from coolant temp sensor, rpm, baro pressure, and throttle position.] By disconnecting the O2 sensor you forced the KE to revert to its fail-safe position; it became a K jet system! How? By replacing the input to the ECU, and thence to the EHA, with preprogrammed values which ran the fuel injection system as if it were a K Jet (non closed loop, non O2, non EHA system).

If the KE runs the same when it's forced to be a K, then you can be certain of one of two things. Either the problem is definitely in the KE portion of the system. Or, the problem is definitely NOT in the KE portion. :confused: But because your car runs well other than this relatively minor problem, and because of your description of the problem, I went with the 'NOT' response. Your O2 sensor and everything that depends on it would not normally mess up idle without causing significantly obvious other problems in the things they are meant to control.

One thing is for certain. If we had OBD II our engines would be so much easier to diagnose.

Much to discuss related to the above, but let me get back to your problem. Has your mechanic checked operating fuel pressure, is he able to graph it? Has he checked the boot below the regulator for proper fitment and lack of cracks? How? How old are the injector O rings? Did he try running propane around the injectors while the engine idled? Have both the warm up valve and the thermo time switch been tested? Have you checked actual coolant temperature (not gauge reading) by using an infrared scanner or tapping into the coolant temperature sensor?

Sorry if you've answered some of my questions earlier, I've forgotten.
 
#27 ·
Well I tried it again on my way home.

The engine RPM hunted wilder than before, to the point it actually stalled as I was about to park by a gas station. Reconnected the plugs and it went back to how it was, albeit with less dips below 500RPM during my drive home.
 
#32 ·
A new oxygen sensor, Bosch p/n 0-258-003915, allowed the needle to finally fluctuate and duty cycle adjusted. It was indeed running way too rich, took a couple of adjustments to get the needle within correct range. According to AutohausAZ I should be using "13925" but the female plug is incorrect. Alternate was 13915 which is correct plug for my 190E and same one I bought.

Apparently, 13144 is for a '93 2.6 but this is the same one currently used. The old oxygen sensor was Bosch p/n 0-258-003144. This is why I was asking earlier whether the incorrect O2 sensor will cause this kind of issue.

On my drive home, the tach needle did not dip at all (like it was going to stall).

My conclusion here was that we could have been adjusting the duty cycle based on a near-faulty or incorrect O2 sensor, thereby getting it way out of spec. This might also be causing bad cold starts, but tomorrow morning's test will confirm this.
 
#33 ·
Same observations today - good normal start. No dipping RPM especially when decelerating. Seems very smooth and very stable!

Seems like my conclusion is correct - the car was not adjusted to spec leading to poor running condition (dipping/hunting RPM) and hard cold starts - but all this was a secondary problem to a defective/possibly incorrect O2 sensor. With the new sensor in place adjustability was back.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top