Chapter 13 of D-Jetronic articles by Dr-DJet describes systematic fault finding (troubleshooting) and repair on your Bosch fuel injection system.
Again and again I see components from other car manufacturers or other motors and the engine runs satisfyingly in some load states. However it will not run well at all loads and engine speeds and will likely consume 20 l / 100 km. Unfortunately this has created a nickname for D-Jetronic in German: Thirst Jetronic (D-urst is thirst in German). I do prefer the other nickname Thinker-Jetronic (D-enker is thinker in German).
|Manometer 2,5 Bar with hose fitting|
13.2 Check fuel pressure
Fuel pressure in loop line is essential for D-Jetronic. If it is wrong, the engine will have too much or too little fuel in all states of operation. Remember that ECU can only vary opening times of injectors to dose fuel. So pressure must always be within reference. Normal reference is 2.0 +0.1 Bar, which means from 2.0 to 2.1 Bar. Some cars (Volvo 6-cylinder) have a higher fuel pressure of 2.2 +0.1 Bar. You can measure fuel pressure anywhere in loop line. Easiest access often is through cold start valve. Either you just remove it and inject your manometer instead or you use a T-fitting to connect. I recommend a glycerine dampened manometer (only slightly higher cost) with a range of 2.5 Bar. Measure with running engine. When you are suspicious about fuel pressure at higher engine speeds because of a possibly low delivery rate of fuel pump or a blocked return line to tank, then you can install this manometer with a long fuel hose temporarily inside your car watching it while you drive. I once ran a 350SLC at 190 km/h on the highway, just to notice that there was no problem with fuel pressure even if the car felt like lean mixture.
13.3 Check component no.
There is no connector or software coding in D-Jetronic. So you can combine absolutely everything from various manufacturers, even you can connect an injector to the air temperature sensor connector and vice versa. There is only one exception: You cannot connect a 4 pin throttle switch to a 5 pin connector for those cars with full-load contact. So there is no closed loop control nor any diagnosis. Your ECU does not notice whether a sensor is missing, whether you wiring harness is broken with short-cuts or whether you have connected wrong components. Last two problems are the most common ones and more probable than a broken sensor or ECU. Workshops also like to install any injector they have, completely ignoring colour codes and flow rates and even mixing different colours. MAP sensors and ECUs anyhow all look the same to them. That is why you first have to verify all component no.s on a car of unknown source. Bosch gives you a good help with its equipment lists on Bosch Automotive Tradition. For Mercedes-Benz cars I have summarized all injection and ignition parts and spares in Appendix A. I am sorry that it is only Mercedes-Benz, but it is a hell of a job to do.
13.4 Systematic measuring on D-Jetronic
|Multimeter and manometer are enough|
If you want to measure deeply in your D-Jetronic, you will need a multimeter. While all Bosch workshop manuals recommend the Bosch EFAW 228, I can calm you down that you do not have to spend up to 1000€ for such a tool. The EFAW 228 is nothing but a multimeter for measuring voltage or resistor plus a switch to activate your fuel pump permanently. The real good thing about it is the adapter that you can hook between engine wiring harness and ECU. Besides that it is a real dumb device and it cannot tell you much about your ECU or MAP sensor. Be aware that you might need adapter EFAW 243 if you have the older version EFAW 193 (that looks very close to EFAW 228). Often workshops only use EFAW 228 as decoration to demonstrate their expertise in D-Jetronic. Such a workshop has once told me that they cannot use this tool on Mercdes-Benz cars as it would only be able to measure 4 cylinder engines. The workshop owner did not even know that injectors are grouped by 2 or 3. So don't get blinded by someone with an EFAW228 and buy a nice multimeter. I recommend one that can also measure dwell and engine revolutions like Voltcraft AT-400. It is fully sufficient and you save hundreds of Euros. Spend some on proper maintenance of your car instead.
Use your multimeter and remove any cover from your ECU. Now measure directly on ECU harness connector. I emphasize this, as this is the only place to see what is fed in your ECU by broken sensors or broken wiring harness. Remove the connector from your ECU and start to measure with multimeter tips directly on connector contacts. If you measure on sensor instead, you risk to oversee problems with your wiring harness.
13.4.1 Power supply
Your ECU has two power suply lines, one for the ECU itself (pin 16) and one for injector drivers (pin 24). If you omit 12V on pin 24, your ECU will operate as normal but just not drive the injectors. That is a good mode for measuring its wrong behaviours. On both pin 16 and 24 you should measure at least 11.5V against ground (pin 11). If you do not see this voltage, either your main fuel injection relay is broken or the wire from battery plus pole to relais is loose. You might of course also have a problem with your ignition lock, corroded mass contacts or your wires. When someone welds on your car, make sure that your ECU conector is completely removed, even if your battery is already removed. Voltage spikes from welding can destroy your ECU via mass contact alone!
Let me tell you a little story regarding power supply on ECU. Sometimes you can loose that voltage quicker than you might think. Only once my MB 450 SL stopped while driving, and that was on the middle of a one lane bridge on a main street! What had happened? I had fetched my car from its winter garage and only stuck pluspole on battery, but did not fix it properly. D-Jetronic receives power from a separate wire attached to your plus pole. So even if everything seems to work on your car including starter, it can be that ECU and fuel pump will not receive any power if this specific wire is loose. And that happened on the middle of the bridge. Someone helped me push my car off the bridge where it started again. Next day it would not at all start in the morning. I found missing voltage on fuel pump but could not see any loose wire there. So I checked relais, everything okay. Finally I read my wiring diagram and found the additional wire from battery there. Fixed it and my SL would start again like a charme. Fortunately that was the only time that my SL ever stopped running and hopefully it will be the only event ever!
13.4.2 Start signal
When you crank the engine via starter, ECU pin 18 receives +12V. This is only input signal. On BMW this is also used as output to after start relais.
13.4.3 Temperature sensors
All engines have an engine temperature sensor and most (only very few early engines like Citroen ad Opel don't) an air temperature sensor. If present, they are always connected to pins 23 (engine temperature) and pin 1 (air temperature). Engine temperature is either sensing water temperature or block temperature on air cooled engines. Both are so called NTC (negative temperature coefficient) type, meaning, the hotter the lower the resistor. Air temperature sensor has a nominla resitance of 250 Ω at 25°C, engine temperature sensor has 2 kΩ at 25°C. Their tolerance is rather high with 15% and depending on where you search, you will find slightly different reference values. What is not common is how their second pin is connected. It can be pin 12, pin 13 or even just ground. You have to look this up in your car speicific documentation. Typical values are:
|air temperature |
Pin 1 versus pin 11, 12 or 13
|resistor ± 15%|
|-30° C||2340 Ω|
|-20° C||1500 Ω|
|-10° C||950 Ω|
|0° C||640 Ω|
|10° C||430 Ω|
|20° C||300 Ω|
|25° C||250 Ω|
|30° C||210 Ω|
|40° C||150 Ω|
Pin 23 vs Pin 12 or 11
|Resistor ± 15%|
|-30° C||26,7 kΩ|
|-20° C||15,5 kΩ|
|-10° C||9,3 kΩ|
|0° C||5,8 kΩ|
|10° C||3,7 kΩ|
|20° C||2,4 kΩ|
|25° C||2,0 kΩ|
|30° C||1,7 kΩ|
|40° C||1,15 kΩ|
|50° C||800 Ω|
|60° C||600 Ω|
|70° C||430 Ω|
|80° C||325 Ω|
|90° C||250 Ω|
|100° C||190 Ω|
If you measure values deviating more than 15%, open or no resistor, then you either have a broken sensor or a broken cable. Repeat measuring directly at sensor pins.
These sensors have great influence on air fuel ratio. Air temperature sensor can enrich mixture by 20% if disconnected, engine temperature sensor can do even worse and enrich by up to 300%. Some mechanics consider it a good trick, to reomove air temperature sensor when idle-run is not perfect. Do not do this, you now know better.
13.4.4 Throttle switch
Besides idle-run contact throttle switch has two or three more functions (only very early VW models till fall 1969 did not have these additional functions). All other ones have 2 acceleration contacts (plus a drag contact), signalling that the driver pushes down the gas pedal. So when you SLOWLY push the pedal down, you should see 10 times 0 Ω and in between open contact, both on pin 9 and vice versa on pin 20. You have to measure against pin 12, pin 14 or ground (depending on how your throttle switch is connected in your car). When lifting the gas pedal, you should only see open contacts. If not there is something wrong with the drag contact. Some cars have an additional full-load contact (those with black caps on MAP sensor), that will signal full-load condition to ECU just before throttle is fully opened. This is typically pin 2 or pin 14. If acceleration contacts or full-load contacts falsely close, they will give wrong signals to ECU, causing it to enrich air fuel mixture. Especially false acceleration contacts are known to cause jerking while cruising.
13.4.5 Absolute pressure switch
Very early cars did not have a full-load transition in their MAP sensor nor the late full-load contact in throttle switch. Instead they had an absolute pressure switch to recognize full-load. It measures intake maifold pressure that will be close to air pressure under full-load and then increases the more you close the throttle until it reaches its maximum in idle-run or over-run. Absolute pressure switch can be found on early VW type 3 or Citroen DS 21 and is connected to pin 9 and 14 on VW and Pin 2 and ground on Citroen. Absolute pressure switch closes if intake manifold pressure reaches 50 (VW) or 60 mBar (Citroen) below sea level air pressure and switches off when it reaches 130 (VW) or 90 (Citroen) mBar. You already recognize that these points will move to different load situations when you are in different heights than sea level and might not even be reached at all if you are in high mountains. For testing you have to evacuate absolute pressure sensor and measure switching points.
13.4.6 Test trigger points / contacts
|Trigger contacts in oscilloscope (8-Zylinder)|
Trigger points can be worn out and are often changed first as suspicious part. As it can cost up to 600€ (or 400€ repair) for 8-cylinders, it is a good idea to measure first.It consists of either 2 (4 and 6-cylinder) or 4 contacts (8.cylinder). We use a dwell meter and measure on cranking or running engine on ECU connector (while it is connected to ECU). on pins 21 (group 1), 22 (group 2) and on 8-cylinder pins 13 (group 3) and 14 (group 4) against pin 12. We should see a value around 135% to 165%. We do not really care exact value as ECU only triggers on falling edge. It is more important not to see a 0 and that all groups show similiar values. If that is the case, your trigger points are okay. Grease them and they will continue to work. If it fails, there are 2 options:
Either you ask Günter (user althase) for new cams and replace them. Or you use Norbert's (user nordfisch) measuring gauge to readjust your contacts. Whatever you prefer, it is a good idea to have that measruing gauge. I have used my lathe to build a slghtly different version than Norbert where I can both test opening and closing situation. The gauge is inserted into trigger contacts and simulates the opening and closing cam from your ignition distributor. If it does not open your trigger points any more, readjust contacts and it will continue to work for a while.
You can also measure the nose on your trigger point's cams. When new, they have 2,5 to 2,6 mm. When they reach 2 mm they will wear out.
If you have an oscillsocope at hand, you can also monitor all contacts. It should look like attached picture.
13.4.7 Testing injectors
We can test injectors for leakage, electrically and of course their spray. First we use a lamp and verify that none of the fuel hoses is wet from fuel. Next we do a short test and touch them while the engine is running. You should feel a small vibration on all of them. If not, test trigger contacts first. Now we do an electrical test. It is important to know how many injectors are connected to each ECU driving pin. That varies on each car. Remember that Jaguar has an external amplifier.
|Spray and leakage test plus flow comparison|
|# cylinders||#injectors /|
resistor pin 3
|4||1 / 2,4Ω||1 / 2,4Ω||1 / 2,4Ω||1 / 2,4Ω|
|6||2 / 1,2Ω||1 / 2,4Ω||2 / 1,2Ω||1 / 2,4Ω|
|8||2 / 1,2Ω||2 / 1,2Ω||2 / 1,2Ω||2 / 1,2Ω|
|12||3 / 0,8Ω||3 / 0,8Ω||3 / 0,8Ω||3 / 0,8Ω|
Every injector has a resistor of 2.4Ω, that is why you will measure either 2.4 Ω or with 2 injectors half, meaning 1.2 Ω or 0.8 Ω on Jaguar. Warning: Never operate injectors at 12V and never continuously open. The coil inside could burn! Measurement is taken on pins 3, 4, 5, 6 versus ground on all cars except Jaguar.
Now you remove injectors from its holder, but leave them on the loop fuel line. Put a small glass underneath each of them and remove pin 15 from ignition coil (to prevent engine from starting). Now ask someone to start the car and let starter run for some time. Watch spray (it should be nicely distributed in a cone). Injectors must not leak after cranking has stopped. Leaking would fllod a cylinder and let fuel pressure drop in loop line. When you have enough fuel in glasses, put them on a board and compare levels in each glass. If they are all the same, your are fine. If one or more deviate too much, injectors need to be cleaned. Cleaning works similiar, only that you use a cleaning fluid and let them spray into an ultrasonic cleaner. Before you reinstall injectors: Check and replace rubber sealings under injectors in intake manifold if necessary. They are a weak point. Rubbers get a lot of heat and tend to harden. That finally causes false air. This sealing was finally improved on L-Jetronic.
Now I hear you ask in which sequence do injectors inject? Well that is easy. It is the same sequence as ignition sequence and is often written on valve head.
13.4.8 Testing MAP sensor
MAP sensor is the most important sensor of D-Jetronic. Unfortunately you can only do a rough test on it. One is whether it holds vacuum and the other is resistor measurment of primary and secondary winding. When you pull a vacuum of 0.5 Bar and it falls to 0.45 Bar in less than 10 seconds or you cannot aplly vacuum at all, than your sensor is broken. If it is a type 2 that is a very common failure of a torn diaphragm. Bosch Classic in Germany can do an overhaul but it costs 400€ (2014). But good news is that I can also do that repair as of Easter 2015.
|Measure where?||Measure what?||Resistor|
|Pin 7 - Pin 15||Primary winding||90 Ω, tolerance: 80 - 120 Ω|
|Pin 7 - Ground||Ground fault||Open|
|Pin 8 - Pin 10||Secondary winding||360 Ω, tolerance: 300 - 400 Ω|
|Pin 8 - Ground||Ground fault||Open|
|Vacuum||Leakage rate at 0,5 Bar||Measure with hand pump|
Unfortunately measuring vaccum and resistor is no guarantee for a properly working MAP sensor. They could have been opened and mis-adjusted. For years I have had many, many MAP sensors in my hand and measured them to build up references for slope and full-load transition. When I compare a MAP sensor with these references on my test stand, I can be sure whether it is working properly at all vacuum pressures and adjust or repair it. Only then can I be sure that no one has misadjusted it, that aneroid cells and springs are still okay and that full-load transition starts at right vacuum pressure. If you want to support my work, drop me a note with your MAP sensor's Bosch number and I can verify whether I need it for building up further references. And if you just want verification, I can do that as well.
What you should never do with your MAP sensor:
- Blow into or apply pressure from outside
- Shake it or beat it
Both will lead to destroying your very sensible sensor. Therefore it should always be mounted on rubber feet and in 90° to drinving direction. Braking and accelerating enforces too much stress on diaphragms and aneroid cells. Last check the vacuum hoase from intake manifold. If it is worn or porous, replace it immediately.
13.4.9 Fuel pump control
Now measure resistor of fuel pump relais between pin 19 and pin 16. It should be only a few Ohms. Failure to start fuel pump is a common problem on ECUs. When you cannot hear a 2 second humming from your fuel pump after switchwing on ignition, measure whether pin 19 against pin 16 goes to 0V for 2 seconds after switching on ignition. If that is the case, your ECU is okay and you should check fuel pump relais, cabling and the fuel pump itself. Cables on fuel pump can corrode or fuel pump can be worn. Other than on K-Jetronic, fuel pump relais normally is not a common source of failure.
13.4.10 Engine wiring harness
Check your wiring harness for damages, broken isolations and rubber bushings. Replace torn rubber bushings immediately. But do not get shocked by their original OEM prices of up to 35€ per bushing. I do know cheaper soruces as you can read in chapter 10.
13.4.11 Cold start valve
Cold start valve is normally operated by a logic outside of ECU under control of a thermo time switch and optionally a relais, but only during starting. BMW uses cold start valve additionally for after start enrichment via ECU. For this variant, please check your workshop manual. I will describe the thermo time switch operated solution here. The switch is a bimetall with a heating that measures engine temperature. Depending on engine temperature it will close contact "W" to ground during starting. At -20°C that will be for roughly 20 seconds, above 35° it will not close any more. Exact values can be read from the side of your thermo time switch. Contact "G" is the heating contact. It avoids that intake manifold is flooded with fuel if you start several times. This means that you can measure opening of thermo time switch and thus operation of cold start valve only once and only below 35°C engine temperature. If you do not manage to measure then, you have to wait for thermo time switch to cool down first. Cold start valve is connected to loop fuel line and only opens by thermo time switch. It must not leak at any other time. If you assume problems with it, there are 2 possible tests before dismounting it. If you want to test whether it leaks, close the fuel line (if it sits at its end like in my MB 450SL). Then you should see a decrease in idle-run CO. Or if you believe that it does not open and you have bad starting with cold engine, apply 12V to its contacts to force it open. Other tests can only be done when you dismount it.
13.4.12 Auxiliary air valve / regulator
Auxiliary air valve (also called regulator, slider) is only responsible for increasing engine revolutions during cold-run. It is also a part of known wear. Expansion elements gets weaker over time, might leak and together with a hanging piston insde it can get stuck or operate only very slowly. It can cause famous sawing of an engine (only if your ECU has over-run shutoff). But in any case, either you have too high revolutions with warm engine or too low with cold engine. If you assume an open and air leaking auxiliary air valve, you can easily close the air hose to verify if it has a change on your engine revolutions. Actutally their should be no influence on hot engine and a big one on cold engine. I have repaired several ones on water cooled engines by replacing expansion element and holder. But it is a tricky thing to do.
13.4.13 Engine control unit (ECU)
Since the start of D-Jetronic, mechanics assumed that every engine they could not fix had a broken ECU. But that also means that they have not followed the checklist of all other possible faults. I check and repair broken ECUs and I can tell you that 2 out of 3 tested are okay. If you have really checked everything else: It is not easy to diagnose a broken ECU. There are of course faults like a broken fuel pump control or an extremely lean or rich mixtures where it helps to just swap in another ECU and the fault is gone. One has to verify function block by function block before they can be declared okay. You can just do the basic test of verifying their power supply, their sensors and vibration of all injectors. Everything else can only be diagnosed by a specialist.
Before you are looking for someone to repair your ECU, be sure, that (s)he has reference values for your up to 14 function blocks of exactly your ECU (Bosch no.) and that (s)he will hand you a protocol of that verification afterwards. There are too many hobby electronic technicians that mix an analogue electronics with a digital one. Even if they find a broken component and replace it and your engine seems to run properly afterwards, this is not good enough! Every semiconductor has tolerances and nowadays one needs to use alternatives anyhow. Replacing a semiconductor means changing signal threshold and amplification in an analogue electronics. That is why the ECU mostly has to be retuned at least in the function block where this replacement happened, sometimes even in all blocks. Without this you might either use too much fuel or run your engine too lean. Just calculate how big such a difference with only 5000 km and 10% richer consumption will be. A cheap repair does not pay off, a professional repair is better. I can help with all ECUs where I took references, it is now all Mercedes-Benz, BMW, several Opel, Citroen and Volvo, both Europe and US versions. Just ask if you need help. Of course Bosch (Koller & Schwemmer) can be of professional help. I heard that they charge 500€ for such a repair and they test it with their Bosch KD-JE 7400 before they send it back. I have my own test stand for that. Always remember to test everything else before you send your ECU for repair.
D-Jetronic is not nuclear science nor magic! Other than ECU and MAP sensor it is easy to test and its construction is simple. Remember to look for faults in D-Jetronic only after prerequisite tests. It is a reliabel system and I always find it surpsrising how reliable its electronics are after more than 40 years. All you need to test it is a manometer, a multimeter, a hand vacuum pump and a systematic approach. That will let you find 90% of all possible faults. And for the rest you have Bosch Classic Center or this forum.
WIshing you succesful fault finding, Your Dr-DJet (Volker)