TONS of codes - KCAN Line Faults. Car stalls and dies after a few minutes, won't restart without battery disconnect and reconnect

Love Gig Harbor. Home of the most incredible YMCA in the country and some of the best cycling anywhere.

The first thing to do is rule out battery and alternator before chasing CAN lines.

Need to verify voltage with both DC and AC (yes your read that right) voltmeters with engine running. Normally, the rectifiers in the alternator's voltage regulator convert AC to DC which is then smoothed with a capacitor/inductor circuit and then voltage controlled by a variable regulator commanded by the DME. When working normally, you should see a DC voltage within 13-14.5V or so with an AC component of under .15 VAC Peak-Peak.

When a regulator diode fails, one half cycle of the AC waveform is no longer rectified and a significantly larger AC ripple is then smoothed and regulated. On the output, you may see DC voltage in the normal range, but you will see a larger AC voltage component, over .5V.

Once the alternator is ruled out, have the battery rechecked. All of this should take a good tech under 10 minutes.

Now, given there are no significant number of voltage related faults, I doubt your alternator or battery are the problem, but they have to be ruled out.

Need to start off measuring the integrity of the K-CAN bus since so many components are griping on it. Using an ohmmeter, should read between 55-60Ω between the K-CAN H and K-CAN L lines. Easiest place to check this is in the trunk on the PDC module. One plug to unplug, use a fine tipped meter lead and go between the twisted pair black and yellow wires.

The K-CAN bus on the car is terminated with two 120Ω resistors on opposite ends of the car. None of the modules have termination resistors in them. The CAN transceivers in each module should have a very high resistance (over 10kΩ or so) so having them connected and in parallel should have no appreciable effect on lowering bus resistance. If you see a K-CAN bus resistance under 55Ω, the transceivers perceive this to be an error and shut down transmission.

If bus resistance is higher than 55-60Ω, you need to unplug modules and figure out why. It's possible one module is causing the failure, member @RedRum5 recently experienced this on the PT-CAN bus and the fault was the SZL CAN transceiver.

If all modules are unplugged and the resistance is still below 55Ω, there is a wiring harness fault. Not out of the realm of possibility if there is any evidence of rodent habitation anywhere in the vehicle.

If bus resistance is higher than 60Ω, there is your problem, one of the two terminating resistors isn't terminating anymore. It's a broken wire or failed resistor (very uncommon).


The KGM (karosserie gateway module..German for chassis) is the CAN gateway/translator for the entire car. It takes your OBD signals, passes can frames between K-CAN, PT-CAN, and D-CAN busses as necessary. This is the only unit that technically has to be plugged in to the bus for the bus to operate. If the rest of the K-CAN modules are disconnected (do not attempt to disconnect PT-CAN modules), you can probably run the engine with the KGM only and see if it helps (although I would not drive the car).

Rodent habitation

Rodent habitation: That is the first thing that came to mind reading this car sat for a couple of months. Still something to check for in the dash or other dark crevices.

Thanks JColley - you appear to be a seriously invaluable asset to this forum. I couldn't find herein such a concise explanation for CAN diagnosis. I know Accuracy connected a breakout box and scope to the system but this guy is so busy he had to send me on my way before thing$ got out of hand. I have an ebay laptop on the way to me with all the software preloaded... BMW Inpa 5.06- Coding Tool - NCS Dummy/Expert, Toolset 32 & BMW TIS. cheaper to buy the entire computer on ebay loaded up than for me to dig around for all this coding ammo.

What is your take on the stalls requiring a battery reset/disconnect to restart the motor? do you feel that the pressure sensor is the problem? I'm not as scared to pull over on the freeway and get behind the car to disconnect the battery cable for 30 seconds, as much as I'm scared I'm leaking down fuel due to a stuck injector from months of sitting while i worked on the car. (Reading about bent rods put the fear of God in me) New filter on the way anyway, so it'll be good to have that inspected and running clean... if it's clean i'll feel much better about the injectors. It ran fine for an entire weekend before going into the shop for KCAN diagnosis; now it won't run for more than 3 or 4 minutes after a battery disconnect without stalling out... even while driving at 3k rpm... not just at idle anymore. It got noticeable worse over 20 miles - i could feather it to 1500rpm to keep things lit but now no more.

The tech explained that when fully reset via battery disconnection, that both fuel pumps operate full on. Then, after a set time, they kick down. His take is at the time of kick down something goes awry and I have not enough fuel pressure to keep the engine on. Are there 2 fuel pumps on the car? I didn't think so.

big thanks man... 1 pizza credit to you anytime. (or 6pack of beer delivery, if they do that in your locale)

Neonguy there are 2 fuel pumps on the car. According to the complete vehicle guide, the DME can increase or decrease the first pump's pressure with speed of the pump and the second one kicks in high demand. Here is a video I did to show how the pump acts on a cold start up. Either the second pump is running or the first pump is told to run high during cold idle start. I assume it's to pump more fuel in to heat the cats up quicker. After a few minutes as you can see in the video, the pump runs at 45 psi when my rpm's drop to 700, hot idle. I use this gauge to check for leak down (stuck injector) after I shut the car off. Plus it would come in handy if I had your problem.

https://www.youtube.com/watch?v=yBL8VLXPJSI

With it stalling have you verified both fuel pump fuses are good? 30 in the glove box, 72 in the boot. Your tank might not be breathing as fuel burns and this has been known to collapse the plastic fuel tank and might cause fuel starvation after a few minutes.

Thanks B767capt. if the tank was pulling a vacuum, I would imagine the car would not immediately start up again after stalling via a simple disconnecting and reconnecting of my - battery strap. Yet this procedure always allows the car to restart, albeit I can smell the gasoline out of the exhaust until the revs stabilize.

I checked all the fuses front and back; all good. Car is 12.35 before starting, 13.6 to 14.5 after. No AC voltage read over .06vac. I pulled the black fuel relay and jumped it out but the car still dies, repetitively, at roughly 2 minutes after starting. Note: the stalling at 2 to 4 minutes after battery reset only started happening after the removal of a Lead Acid battery, and the registration of a new AGM battery. The car ran fine with the AGM and no registration for 30 miles.

KCAN resistive measurements next.

I would get ISTA/D and see if you can test all the reported modules, one by one. See if you can talk to them. Keep and eye on the system diagram as you do. I bet the network is disconnected somewhere and a whole bunch of nodes are not reachable. This before you start looking at termination resistances.

The stalling is likely a fuel pressure thing. Don't get so hung up on the battery disconnect reconnect restart manifestation. I would say: Measure the fuel pressure. From there you go. The error related to pressure is not a fluke. It is trying to tell you something.

while snooping for the KCAN fault, I found a little surprise... not the fault I was looking for... but I'll take it! LOL. not that i'll ever push this car to the far limit... but if I want to go faster on land than I have on my Ducati, looks like I can do it now?

http://www.m5board.com/vbulletin/attachment.php?attachmentid=668921&thumb=1

I've ohmed the kcan line and it's either full open or shorted. one of my meters says open on a 200 ohm sensitivity but the on the diode/conductivity setting says .5 ohms. The other meter says 0 ohms, leading me to believe I have a short. Time to buy a fluke. I'm getting ready to take the entire dash apart to look for more modules.

Is there a comprehensive photo guide to where all the modules on the car, and more specifically the KCAN modules, are located? All connections in the trunk and under the passenger air filter look good... solid. Also, all fuses in the compartment under the cabin air filter tested good, as well as behind glove box and in trunk. So now it looks like I rip out seats or the dash board. Not looking forward to it. Again, your .02 is valued here!

Neonguy

Looks like the previous owner was an electrician? Oh boy.

First off, dump that silly EVC module. If you want to remove the governor, do it properly with a DME tune. I'm amazed a single one of those ever sold given how easy it is to remove the limiter these days. See installation manual for this module in this post. This module intercepts the signal from the right rear wheel speed sensor (the one the DME uses for speed measurement) by cutting this line. The unit then passes the signal through until it exceeds a threshold, then just passes an acceptable signal to the DME. It's a kluge and silly way to go about things since the DME uses that signal for a lot of other functions, such as alteration of air filling parameters (intake charge). These were popular before the days of owner flash tuning.

Secondly, I would lean towards a failed module somewhere. Your multimeter I'm assuming is an analog version, good for you. You've potentially diagnosed more in a few seconds than you could have in days with a digital.

The analog meters use different voltage values on the output depending on the resistance range selected to prevent over-ranging the meter. On a very low resistance setting such as 200Ω, it only puts out a very small voltage, I'm going to guess less than 0.7V (breakover voltage to forward bias a Silicon based semiconductor PN junction). This could indicate a problem with the wiring and terminating resistors as the reading should be consistent at any applied voltage. When used in a higher resistance range, a higher voltage is used on the meter leads to bring the needle back in the center of the meter face, apparently over 0.7V. More specifically, you have a diode test setting which most definitely applies greater than 0.7V to ensure sufficient voltage is applied to forward or reverse bias a PN junction.

Now, it would appear some module somewhere is faulty as there shouldn't be anything in parallel with the K-CAN Hi and Low lines representing as a PN junction. No telling what this may be as I'm unfamiliar with the CAN transceivers, but good electronic designs use transistors or opto-isolators on their signal IOs, so perhaps one of these have failed.

I'll look up a k-can schematic and post it when I get a few minutes today.

SSP for the K-CAN system.

View attachment K-CAN SSP.pdf

Eradicate Unit By Selective Removal

It looks to me like there is no comms possible on the K-CAN line at all. This is likely because every message put out on k-CAN is causing a CAN error frame when the faulty unit is present. Since every message sent out is received by 'everybody', this can get very chaotic quickly.

An approach that can be taken is to unplug 1 out of the all the K-CAN members at a time and see if the error messages stop flooding the network. If you have a CAN tracer device, this can be seen on the fly when you start removing one device at a time.

But if you don't have such a tracer, try this:

1. Remove one of the k-CAN devices.
2. Clear all fault codes.
3. Turn on the car and let all CAN faults log for a few minutes.

I bet that when you disconnect the faulty module, all of a sudden there will be a LOT less faults stored. There will still be fault stored for the one and only missing module, but not for all of them. This is because this one faulty module impairs traffic to the rest of the network system.

Good luck!

The new Fluke meter says the KCAN is open, at least on the PDC wires in the trunk. Voltage is 4.3 black to yellow, 4.69 yellow to ground, .24 black to ground. Black must be high? these voltages do not mate with CAN diagnosis protocol (voltage doesn't matter if engine started or ignition off). HOWEVER, regardless if car is running or or not, the KCAN line voltage yellow to black always goes to 10.25v volts after 23 to 25 seconds. Also, have I ever mentioned my brake lights are flashing in a periodic sequence? I'll post a video... it's an odd rhythm. I can open a door, or push the lock or unlock on the remote... and KCAN voltage goes right back to 4.3v... for only 25 seconds, then it ramps up again. Does this sound right?

I've read material stating there is a KCAN line to the front bumper where there should be a termination resistor... but I cannot find any wires other than the PDC harness wires, air temperature wires, wires to the radiator and hood latch and fluid pump; no black/yellow or twisted pair. I am sure the KCAN is likely a twisted wire pair throughout the car? And likely consistent Blk/Yellow? I also read opposing advice that the 120 ohm KCAN line termination resistors for the E60 M5 are located one in the steering wheel assembly, and one under the passenger seat. Does anyone have a photo of what i'm looking for?


BTW - The balance of my post is here just as a detailed formality in case someone wants to read up on CAN diagnosis.

Diagnosis on CAN-bus
The CAN (Controller Area Network) bus system is a linear bus system that is characterized by the following features:

signals are broadcast in both directions a message is received by all bus nodes. Each node decides for itself whether to process the message or not. new nodes can be added by simple parallel connection the bus system forms a multi-master structure each node can be master or slave depending on whether it is connected as transmitter or receiver the transmission medium is a two-core lead; the twisted pair cores are designated CAN Low and CAN High.

Switching off defective control units. In general, every node can communicate with all other nodes via the bus. An access mechanism controls data exchange on the bus. The main differences between the K-CAN (body CAN) bus and the PT-CAN (powertrain CAN) bus are detailed below: K-CAN: Data transfer rate around 100 kBits/s. Single-wire operation possible. PT-CAN: Data transfer rate around 500 kBits/s. No single-wire operation possible. Master: the master is the active communicating node, i.e. the one that initiates communication. The master is in control of the bus and controls communication. It can send messages to the passive communicating nodes (slaves) on the bus network and on request receive messages from them. Slave: the slave is the passive communicating node. It is instructed to receive and send data. Multimaster system: a multimaster system is one in which all communication nodes can take on the role of master or slave at a particular time.

The failure of communication on the CAN-bus may be caused by the following: - Circuit breaks or short circuits on the CAN Low or CAN High communication leads - Defective plug connections (contact damage, soiling, corrosion) - Interference voltages in the vehicle electrical system (caused for instance by defective ignition coils or ground connections) - Failure of the communication modules in the individual control units - Failure of the voltage supply of individual control units (a battery voltage decreasing gradually when the battery is almost discharged can also lead to fault code entries as not all control units switch off simultaneously due to the voltages supply being too low).

Inspection procedure for impedance measurement PT-CAN, F-CAN, Local-CAN When measuring impedance, it is generally necessary to disconnect the circuit being tested from the power supply beforehand. The vehicle's battery should therefore be disconnected. Wait about three minutes to allow all capacitors in the system to discharge. Inspection procedure for resistance test: the CAN bus must be disconnected from the power supply No other testing equipment must be in use (connected in parallel) The measurement is taken between the CAN Low and CAN High leads the measurements recorded may differ from the specified levels by a few Ohms Impedance measurement with matching resistor PT-CAN, F-CAN, Local-CAN On the Tester, switch to Measurement system -> Multimeter Measurement function: Resistance Measuring range: automatic In order to prevent signal reflection, a 120 Ω resistor is fitted to two CAN bus nodes (at the extremities of the PT-CAN network). The two terminal resistors are connected in parallel and form a shunt impedance of 60 Ω. When the power supply is switched off, that shunt impedance can be measured across the communication leads. In addition, the individual resistors can be tested independently of one another. (Tip for 60 Ω measurement: disconnect an easily accessible control unit from the bus and then measure the impedance between the CAN Low and CAN High leads on the connector.)

DC voltage measurement PT-CAN, F-CAN, Local-CAN
Precondition for the measurement: battery connected and ignition on! On the Tester, switch to Measurement system -> Multimeter Measurement function: Voltage Measurement type: = Measuring range: automatic In order to establish whether the CAN Low or CAN High lead is defective, you can measure the CAN Low (CAN High) voltage to earth. CAN Low to earth: voltage approx. 2.4 V. CAN High to earth: voltage approx. 2.6 V. These values are approximate values and can vary by a few hundred mV depending on the bus load.

Oscilloscope measurement PT-CAN, F-CAN, Local-CAN Precondition for the measurement: battery connected and ignition on! On the Tester, switch to Measurement system -> Oscilloscope setting Measurement type: = Measuring range: +/- 5 V Frequency range: 1 kHz In order to obtain a clear idea of whether the CAN bus is functioning properly, it is very useful to be able to observe activity on the bus. What is important here is not to analyse the actual data being transmitted but simply to be able to see that the CAN bus is operating. The oscilloscope test can state that, ”the CAN bus is probably operating without faults”. If the oscilloscope is used to measure the voltage differential between the CAN Low and CAN High ground leads, a square wave signal with the voltage limits U(min) = 1.5 V and U(max) = 2.5V is obtained. If the oscilloscope is used to measure the voltage differential between the CAN High and CAN Low ground leads, a square wave signal with the voltage limits U(min) = 2.5 V and U(max) = 3.5V is obtained. These values are approximate values and can vary by a few hundred mV depending on the bus load. Resistance test K-CAN No defined resistance test can be carried out at the K-CAN data bus, as the resistance varies depending on the internal switching logic of the control units!

Measuring K-CAN DC Voltage Precondition for the measurement: battery connected and ignition on! On the Tester, switch to Measurement system -> Multimeter Measurement function: Voltage Measurement type: = Measuring range: +/- 10 V In order to establish whether the CAN Low or CAN High lead is defective, you can measure the CAN Low (CAN High) voltage to earth. CAN Low to earth: voltage approx. 4.8 V. CAN High to earth: voltage approx. 0.2 V. These values are approximate values and can vary by a few hundred mV depending on the bus load.

Oscilloscope measurement K-CAN Precondition for the measurement: battery connected and ignition on! On the Tester, switch to Measurement system -> Oscilloscope setting Measurement type: = Measuring range: automatic Frequency range: 1 kHz In order to obtain a clear idea of whether the CAN bus is functioning properly, it is very useful to be able to observe activity on the bus. What is important here is not to analyse the actual data being transmitted but simply to be able to see that the CAN bus is operating. The oscilloscope test can state that, ”the CAN bus is probably operating without faults”. If the oscilloscope is used to measure the voltage differential between the CAN Low and CAN High ground leads, a
square wave signal with the voltage limits U(min) = 1 V and U(max) = 5V is obtained. If the oscilloscope is used to measure the voltage differential between the CAN High and CAN Low ground leads, a square wave signal with the voltage limits U(min) = 0 V and U(max) = 4V is obtained. These values are approximate values and can vary by a few hundred mV depending on the bus load.

CAN bus without function If the K-CAN or PT-CAN data bus indicate no function, it is likely that a short circuit has occurred on the CAN Low and/or CAN High lead, or that a control unit is defective. In order to localise the cause of the fault, it makes sense to use the following procedure: Disconnect one CAN-bus node after the other until the unit causing the fault (= control unit x) is found Check the lines to control unit x for short circuits If possible, check control unit x However, this procedure only leads to success if a tap line from a control unit to the CAN bus has a short circuit. If a CAN bus lead itself has a short circuit, the wiring harness must be checked. Diagnosis: Two different bus faults can be entered in the CAN bus control units: CAN communication fault, CAN wire fault

The communication fault provides an overview of the control units that have failed on the CAN bus, i.e. were no longer able to communicate. The ”CAN communication fault” can only be read out if the fault is currently not present. If the fault is currently present, it is no longer possible to communicate with the control unit. This means that the fault code memory cannot be read, either! Physical wire fault can be detected by the use of fault-tolerant CAN transceivers. However, at the moment there are only fault-tolerant transceivers for the K-CAN data bus. This means that only control units that are connected to the KCAN data bus can have made the fault code memory entry ”CAN wire fault” The CAN transceiver is also unable to distinguish between the individual fault categories listed below.

If the bus fault ”CAN wire fault” is entered in a control unit, this can mean: Break in wiring CAN High, Break in wiring CAN Low, Short circuit CAN High to earth or positive, Short circuit CAN Low to earth or positive, Short circuit CAN High to CAN Low, Break in wiring (single-wire operation): each control unit contains a separate bus termination. This means that, even in the case of a break in wiring, the voltage level can be maintained across the entire K-CAN network. The consequence of this is that a transmitting control unit does not detect this fault and continues to work in two-wire operation. However, if a control unit transmits a message across the break, the receiving control unit only detects activity on the undamaged bus line. The receiving control unit thus detects single-wire operation and sets the fault ”CAN wire fault”. If different control units receive messages across the break, a number of control units can have made this fault code memory entry in single-wire operation! Short circuit: if there is a short circuit in the system, all K-CAN control units must have entered the fault ”CAN wire fault”. In order to localise the short circuit, follow the procedure for ”CAN bus without function”.