Loss of ODB2 Port network connection, I think it's a TIPM problem

[Ozymandias]

Oh boy... that truck is really laughing at you.

 

[MichaelOverfield]

Oh boy... that truck is really laughing at you.

It's the computers.. Seems all computers hate me! Could it be taking that computer out to the shooting range in College to get revenge is coming back to get me?

 

[Ozymandias]

It's the computers.. Seems all computers hate me! Could it be taking that computer out to the shooting range in College to get revenge is coming back to get me?

Very much so.

My wife is the same, as soon as she is around the stuff starts failing.
Printer out of order, WiFi down, Cell con. only one bar, ATM refuses to cash out and so on.. she says she's cursed.

 

[MichaelOverfield]

So update on the saga.. after about 6 drives it seemed to be working, but they were short drives. I did a one hour and return yesterday, and yep.. problem remains. About 20 minutes, then maybe 30, and then only 7 minutes. It appears that it is made worse when the ECU is warmed up, either by a very long power on engine not running, or a drive cycle. Since I had signed up for Just Answer, I thought I'd test them on this issue.... the first mechanic that helped with the WCM failure to program keys, well he decided to pass on this issue, another mechanic did come online, but at first it was clear he didn't fully understand the nature of the problem.. after an extensive dialogue, and his asking me about things I have already checked, and verified good.. I'm back to checking the ground to the ECU and power supply again, I did note the UG, while it was still on and in comms with ECU was reporting a slightly lower voltage than the other voltmeter, and had slight variation in voltage.. between 14.2 and 14.3, with the other meter reading a pretty consistent 14.5. Could there be a power issue to the ECU? Well I guess it's possible, so I'll retrace the power side, and ground side to the ECU, again.

The last comment from the mechanic is actually a bit ironic.. he says if I can't find it to just drive the truck until it fails completely! That is EXACTLY what I seek to avoid! I know if I just wait for that to happen, it will happen at a very bad and EXPENSIVE time while away from home and my full assortment of repair tools, and thus I'm stuck paying a crazy towing and/or repair bill that I seek to avoid... but it is clear, he is also stumped by this bizarre situation.

 

[MichaelOverfield]

One step forward... TWO steps back...

So, I got to thinking. It appears when the ECU goes out of communication with the DLC (wiTECH, or any OBD2 device), it is STILL in communcation with everthing else, like ABS and TIPM. Otherwise one would expect associated U codes.

So to test this theory, I needed to get the ECU to go out of comms, then unplug the engine harness (cutting out the ECU) and look for the U-codes. If the U-codes show, then I know for 100% sure that the communication with the ECU continues after the DLC communiction with the ECU stops. That leaves the issue at the TIPM, or ECU only (though it is odd ECU won't go through TIPM to DLC, but goes everywhere else?).

So, first try, over 4 hours IGN on, engine not running and connected to power supply.. no loss of comms. Next day, over 8 hours, same result. Decided I'd take it for a pair of 1 hour drives and see what happened. Less than 20 minutes in, ECU comms through the DLC are lost. Same result on the return drive.

So, once I arrived at destination, I cycled the igntion quickly to stop the engine, but keep the ECU in its non-communication state.. then connected wiTECH. No codes, no ECU comms as expected.

Disconnected the egnine harness/ECU and instant U codes from ABS and TIPM as they see the loss of the ECU. So verified... Even in the non-communication state, communication with other modules on the CAN C bus continues.

Quickly plugged the engine harness back in, and codes went from active to stored, meaning communcation returned, but NOT TO THE DLC.

Tried same basic proceedure upon return home, except this time keeping the ECU/engine harness diconnected longer to let the ECU reset. Then when plugged back in the commucation to the DLC returned. So, something with the ECU specifically appears to be drivng this condition, and it happens on 2 different ECUs.. so that is strange indeed.

Next step is to break out the CAN C bus directly and do O-Scope analysis of the signal there, Also I plan to add a direct connection to CAN C, basically bypass the TIPM, because if the commincation is continuing on the CAN C, but not getting through the TIPM, a way to "fix" this might be to wire a direct separate DLC that connects to CAN C. At this point I'm willing to entertian such measures to put an end to this crazy saga.

 

[sag2]

One step forward... TWO steps back...

So, I got to thinking. It appears when the ECU goes out of communication with the DLC (wiTECH, or any OBD2 device), it is STILL in communcation with everthing else, like ABS and TIPM. Otherwise one would expect associated U codes.

So to test this theory, I needed to get the ECU to go out of comms, then unplug the engine harness (cutting out the ECU) and look for the U-codes. If the U-codes show, then I know for 100% sure that the communication with the ECU continues after the DLC communiction with the ECU stops. That leaves the issue at the TIPM, or ECU only (though it is odd ECU won't go through TIPM to DLC, but goes everywhere else?).

So, first try, over 4 hours IGN on, engine not running and connected to power supply.. no loss of comms. Next day, over 8 hours, same result. Decided I'd take it for a pair of 1 hour drives and see what happened. Less than 20 minutes in, ECU comms through the DLC are lost. Same result on the return drive.

So, once I arrived at destination, I cycled the igntion quickly to stop the engine, but keep the ECU in its non-communication state.. then connected wiTECH. No codes, no ECU comms as expected.

Disconnected the egnine harness/ECU and instant U codes from ABS and TIPM as they see the loss of the ECU. So verified... Even in the non-communication state, communication with other modules on the CAN C bus continues.

Quickly plugged the engine harness back in, and codes went from active to stored, meaning communcation returned, but NOT TO THE DLC.

Tried same basic proceedure upon return home, except this time keeping the ECU/engine harness diconnected longer to let the ECU reset. Then when plugged back in the commucation to the DLC returned. So, something with the ECU specifically appears to be drivng this condition, and it happens on 2 different ECUs.. so that is strange indeed.

Next step is to break out the CAN C bus directly and do O-Scope analysis of the signal there, Also I plan to add a direct connection to CAN C, basically bypass the TIPM, because if the commincation is continuing on the CAN C, but not getting through the TIPM, a way to "fix" this might be to wire a direct separate DLC that connects to CAN C. At this point I'm willing to entertian such measures to put an end to this crazy saga.

The CAN C should be a direct pass through in the TIPM, but in case it isn't you would need to wire around it with a jumper, but leave the OEM wiring still connected. The reason is the TIPM is responsible for a few things to make the truck run, so if you bypass it completely I don't believe it will work. Also remember if you completely bypass it, the interior BUS will not be able to communicate with CAN C as the TIPM is the translator for the two BUS systems.

 

[MichaelOverfield]

The CAN C should be a direct pass through in the TIPM, but in case it isn't you would need to wire around it with a jumper, but leave the OEM wiring still connected. The reason is the TIPM is responsible for a few things to make the truck run, so if you bypass it completely I don't believe it will work. Also remember if you completely bypass it, the interior BUS will not be able to communicate with CAN C as the TIPM is the translator for the two BUS systems.

I would not be changing any existing wiring or connections. The idea is to see if I can bring a CAN C signal direct to a separate OBD 2 port that does not go through the TIPM, basically make a second DLC with only a CAN C network connection. It appears the TIPM and ABS are in communication with the ECU, but the wiTECH and any DLC connected device is unable to communicate. I can't think of any reason why this should happen, but I know for a fact it is happening. It is also weird, the issue seems to originate in the ECU. If I unplug, and wait a short time, then plug back in, communication is restored! That part is bizarre. I think I need to see the CAN signals on the CAN C directly to see if there is something strange there, but then I'd expect some kind of U-codes, but I get nothing, except a random U0141 from the WCM, which happened once with the NEW WCM, forgot to mention that earlier. I always find the U0141 as a stored code, so it's a intermittent issue there.

 

[MichaelOverfield]

Well, I now have a dedicated CAN C DLC in addition to the original DLC. I started by testing the CAN C directly under hood, separate from the CAN D that is via the TIPM. I verified good off state resistance, and nominal voltages.. then connected the 2 channel Oscilloscope and observed perfectly normal CAN signals. From there I wired in the new DLC. Went for a short test drive.. and the results.... Problem persists. Same issue. When I got back from the drive, connected wiTECH to the original DLC, same situation and no error codes of any kind.

 

[sag2]

What pins did you connect the CAN C to at the DLC?

Did you take a photo or recording of C when the problem was present?

 

[MichaelOverfield]

What pins did you connect the CAN C to at the DLC?

Did you take a photo or recording of C when the problem was present?

I used the wiring diagrams to locate the white/light green CAN H at the TIPM and the white/light blue CAN L also at the TIPM (it was an easy and place to access and splice under the TIPM, brown plug). I verified these were correct via O-scope, and also break out on the engine connector to verify continuity from that connector to my new splice and new CAN C DLC that I routed into the cab. For the rest of the DLC connections I used the existing DLC connections, which is just 2 grounds and power. The CAN D (from TIPM) was just replaced with the direct CAN C line from under the hood. This was easy to do using a splitter I had on hand, the last connector I simply cut the CAN H and CAN L wires and connected those under hood to the CAN C, and the rest of the connectors are already connected via the existing CAN D plug.

It clearly is connected to the ECU, since it works for a time after a start, the issue is at some point the ECU stops communication with the connected device, AND that includes wiTECH. What makes no sense is that it appears to communicate to other modules, since they do not set codes. Oh, and I also verified the proper off state resistance for the CAN C, and it was 59.8 Ohms which is right where it needs to be. If you check on the original DLC (CAN D), you are just getting the 60 Ohm resister within the TIPM, it's not a direct connection to the CAN C, and thus my new plug.

Also, I verified proper and normal CAN signals were being transmitted by direct CAN C connection, both when the ECU was in communication, and when it was out of communication, there was not difference that was notable on the O-scope.

Just when I thought it could not be more bizarre, it goes and gets even weirder.

Though I will keep the direct CAN C DLC, since it makes for much easier trouble shooting that trying to find a place to probe the CAN C under the hood, that was not a design feature the makes it easier to solve network issues.

Oh, to one of your questions, I did a few freeze frames from the O-scope, but once I had all the settings correct, the signals were textbook normal.. I could show you a picture of a proper can signal and overlay what I was getting and there was no difference whatsoever. One easy way to verify is set up the add function of the 2 channels, since they are supposed to be symmetrical with the H going volts up, and the L going volts down by the same amount, the ADD of the 2 should always be zero amplitude, that line was flat and stable, thus the signals are very clean and proper. One interesting thing however, the signals were much cleaner than the previously measured CAN D from the original DLC that routes via the TIPM, that tells me the noise on the CAN bus is coming from either the TIPM or the CAN B bus, since the CAN D carries both signals.

I have a couple of thoughts.. when I disconnect the ECU via the engine plug, and DTCs are set by TIPM and ABS, one other problem has been created that I don't have a solution for, yet.. that being the EOL 120 Ohm resister is also removed when I do that disconnect, that alone creates a network issue.. so I might try to see if I can find a way to add a 120 Ohm resistor "switch" where I can interrupt the ECU signal path, but replace the 120 Ohm EOL resistor needed to ensure the CAN signals continue to transmit.

It might be possible that the ABS and TIPM set the loss of ECU due to the loss of the 120 Ohm resistor.. but that is a bit of a longshot, and the reason is this.. the TIPM and ABS are not setting loss of comms DTCs with each other when the ECU is unplugged, so it seems communication is happening, but I can unplug the ABS and the TIPM sets a loss of ABS code immediately, so if the ECU unplug messed up the network, then the TIPM should also be setting a code for loss of ABS, but that is not happening.... So, that might be a dead end, like so many others before.

So I think my next effort is going to be another ECU swap I'll go with the other and see what happens, again. Right now the original is in there, as the replacement was doing the same thing. In the process of replacement I'm going to do some pin continuity checks on the power supply to the ECU. It's a longshot, but if the input power is unstable from a high resistance connection, it might be causing ECU CAN transmission errors (though not obvious on the O-scope). There is an error counter that will cause an ECU on the CAN bus to enter an off state, if 2 ECUs are doing the same thing, the thing that is now in question would be the wiring harness.. and it would not be my first Dodge that had wiring issues from the factory!, see my earlier "sea story" for the details on that one.. that still seems to be a possibility, though it seems if it were the issue it should also be setting loss of communication DTCs, but it is not setting any DTCs... and that makes absolutely no sense whatsoever.

I'm still in a back and forth with a Dodge Mechanic on Just Answer, and he too is completely stumped on this issue. No one, I mean no one appears to have seen anything like this, no matter where I look.

 

[MichaelOverfield]

Given I have valve lash adjustment and crankcase filter replacement coming due soon, and plan to install head studs as well.. and some inspect and new intake heater work coming up, I decided to order a new engine wiring harness. Can't be sure it's an issue, but I have had a harness issue on a prior dodge, and given the the time an effort involved, the $218 for the engine harness was not too devastating, so I'll replace it when in there doing the other work, and see if that changes anything.. one theory is a voltage stability to the ECU might be be an issue, though you would not know it from the O-scope CAN signals that looked perfect.

 

[MichaelOverfield]

So I was digging deep in to the wiring diagrams. The ECU main 12V power comes via a 30Amp fuse (42), with a 10 Ga red wire, this goes to that big engine connector (C130, pin 4) near the brake master cylinder. From there it gets interesting. One can look at the ECU connectors and pins and clearly see no pin there is going to be rated to carry 30Amps... so there is a splice, S165 in the harness just above the ECU, that splice takes the single 10Ga Red wire and splits it to three 16GA wires to go to 3 pins on the ECU, the C2 connector is the right side connector, it's Pins 49, 50, 60. If there is a bad splice in that harness, and say one or more of those pins is not connecting power, I could see that causing unstable power to the ECU.. but it's just a guess if that is actually an issue, when I do the ECU swap I'll be looking for that splice and testing the continuity of these wires. Last night I went ahead and did and open and inspect on the spare ECU, this one had a bit of abuse showing on the case and a small dent. I wanted to verify that was not an issue, and also that it was well sealed. The unit is sealed with RTV. Everything seemed OK, and it was well sealed, and I re-sealed with fresh RTV and put it all back together, here are some pictures of what the inside looks like:

 

[DavidC]

The days of electrical troubleshooting are pretty much over when you get to that point.

 

[sag2]

I could be thinking of the wrong ECU, but I think I saw a guy online that repairs the ECU's for Mopars, and he said that ribbon connector is only good for a couple openings. That ECU appears to have been opened before. You might do a search on Youtube for Mopar ECU repairs.

 

[MichaelOverfield]

I could be thinking of the wrong ECU, but I think I saw a guy online that repairs the ECU's for Mopars, and he said that ribbon connector is only good for a couple openings. That ECU appears to have been opened before. You might do a search on Youtube for Mopar ECU repairs.

This one is a spare. I would doubt the ribbon connector would be what would fail for multiple openings, but I think the case itself where it bends WOULD only take a couple of openings. The metal, aluminum in this case work hardens when you bend it, and I would expect if I opened this ECU one more time, it would likely be the last time for that "hinge" before it cracked..

In any case, when I swap it I'll know pretty quickly if it is an issue, and I always have the original as a spare.

 

[MichaelOverfield]

Saga continues… Yesterday I swapped ECUs, again… my son was making a joke that I now swap ECUs like a spare tire, and literally do carry a spare ECU!

Anyhow, the primary focus was to do some pin to pin checks are the wiring harness to the ECU, focusing on the power and grounds, and to take another look at the ECU mount.

There are three 12V power red 16GA wires on C2 (right/aft) connector on the ECU. There are also three 16GA black ground wires to the ECU. My thought was if there was a problem in the harness, a bad splice would seem a possible harness problem (the three red 16GA splice to a single 10GA that runs to the TIPM and 30A fuse 42, if I recall the number correctly).

Well, I got good continuity on all power and all grounds. The power ran at 0.4 Ohms each, the ground at 0.3 Ohms. When I took the value at the engine connector on the power side I got 0.3Ohms for power, so from connector to TIPM there is another 0.1Ohms. I expected lower resistance actually, so maybe there is something to this, but I can’t see how it’s a major issue.

In any case, having the ECU out, I took the mount off again to take a closer look. At first I thought I had something… the back of the mount shows signs of rubbing against the engine. The mount has 4 silicone bushings that allow it to absorb some vibration. It would fit my EMI theory if there was another point of ground on that mounting plate other than the ground wire.. so if the plate was rubbing against the engine block that could induce EMI… but looking at the block under the mounting plate, one finds it has a plastic based insulation material there, assuming it is for noise and possible thermal insulation so the point of rubbing was not a ground, though it could cause more ECU vibration if the mounts were not as free to move as designed, so I opened that gap up a bit using some washers under the bushings to give clearance. Also coated the contact points with liquid electrical tape, mainly so I can verify it’s no longer making contact. Pictures show the before and after of the ECU mounting plate.

One of the reasons to swap ECUs was to verify my inspection of the spare did not render it inoperable. After I sealed it up, I cleaned the outside, masked off the connectors and mounting bolt holes to preserve a good ground connection, and then applied some clear POR-15. The original ECU has some kind of clear coating, so really looking to do the same on the spare, as well as be certain it’s well sealed.

Spare ECU appears to be working fine, I also wanted to verify my prior strange idle/high alternator load issue did not return, as it was present with that ECU when last installed. It did not come back and I think that issue was a bad connection to the alternator sense pins that provides signal to the ECU. I found both alternators had a loose connector and that might have been the issue all along.

In any case, spare ECU working fine… so I did not break it. Though…. About 10 minutes of running.. and original problem returns. ECU stops communication with the OBD2 UG, and when connected wiTECH can’t communicate with ECU. Cycle ignition, ECU comes back, no DTCs of any kind.. so 2 ECUs do the EXACT SAME THING. It is unlikely the TIPM has any part to play in this issue, since my new CAN C DLC is making a connection direct to the ECU on the CAN C. So I’m left with the original issue, which I kind of expected.

The new wiring harness is on its way, and will get installed when I do the crankcase vent filter, replace the intake horn/grid heater, and install head studs and adjust valve lash that is coming up this winter. That harness will be easier to replace with all those other things out of the way.

 

[MichaelOverfield]

Ok, got the new harness today, did a quick inspection, it appears to be the right one. I will do resistance checks on the brand new harness and see if that 0.3 Ohm reading from the existing harness is repeated on the new harness. Either way, during the next major maintenance period that new harness is going in.

I've done a couple of drives since last ECU swap, on the way to work, both times.. no problems with comms, about 30 minute drive.. oddly on the way home yesterday, would not stay up for more than about 10 minutes. Did 2 resets, same thing. It will be interesting to see what happens today on the way home.

I also did another test for a possible "solution" as a last resort if I can never figure this thing out... I could install a "reset" switch. The test was to pull the 10A ignition fuse when I got home and engine was at idle.. expected result, engine shutdown, result verified. I'm thinking I could have a "kill/reset" switch do do just that and reset the ECU when it goes out of comms. It's totally something that should not be needed, but it would have an added side benefit if I were to hide this switch as an additional anti-theft device (of course unnecessary, the manual transmission is perhaps the best anti-theft device possible on a modern vehicle!).

It would not be difficult to wire a switch in that ignition power supply to the ECU.

 

[MichaelOverfield]

Ok, the existing and new harness checked out, 0.3 ohms is basically what I get when I short out the test probes, so power and ground is apparently not the issue.
Got to thinking, could the ABS module be creating some CAN interference that would cause the ECU to drop low priority communication? Well, thought up a way to test that theory... hoping I could confirm it.. since I'm still skeptical the wiring harness is the issue.

So I had a long commute last eve, on the way out it worked for close to an hour before going to it's now infamous off state. On the return, after looking up the operation of the ABS module and noting that it bascally does nothing unless wheels are about to stop rolling, thus if it was not even plugged in, brakes remain fully functional as they would be in any non-ABS vehicle.. so I decided to do the return drive with the ABS module unplugged. This test, if the problem did not happen would tend to indicate that the ABS was suspect.

Well, it did the usual at about 40-50 minutes (sometimes less), so it is looking like ABS module is not the issue.

I did learn some interesting things about the ABS module that I thought I'd share. With it unpluged, there is no vehicle speed signal, so no speedometer, and the ODO does not change, and MPG is horrible (burning fuel, with no distance measured!). Oddly the reverse lights are somehow disabled... can't really figure out how that has anything to do with ABS, but found it when my back-up camera would not engage. Now on an automatic transmission vehicle that loss of speed signal might cause other problems, but on a manual it really does not seem to matter.

Then the crazy... after a short time I get the P0501, that is for the speed sensor being erratic, yep, it was not connected so no surprise, but that is not the crazy part. After confirming the issue was not ABS, I stopped to plug it back in and restart the truck. As I pulled into the parking lot.. as soon as I engaged the parking brake.. the overhead module starts going nuts, saying DPF full, and SEE DEALER NOW messages. Odd, really odd, but I'm going ahead with my shutdown, plug ABS in and restart no matter these strange messages.

When I plug in the ABS and restart, it still has the messages at first, but soon all goes back to to normal (I wanted to reset the MPG and at first all I got were the stupid see dealership messages), MIL was still lit, but it was just the P0501 code per my UG scantool (of course it can only see P-codes based on OBD2 protocols), because I did not do a code clear, and figured I would do a wiTECH scan and see what else might be stored there.. and that turned out to be quite interesting.

So I had the following stored codes when I got home and did the wiTECH scan: P0501 (speed sensor, expected), U0121 (ECM and TIPM loss of comms with ABS module, expected), and the interesting ones: U0100 (ABS and TIPM, loss of comms with ECM). It gets really weird when looking at the data for each code. The ODO readings show something unexpected. the U0100 codes were present and set BEFORE I unplugged the ABS when I was on my way to my destination, further, they appear to have been set before the loss of communication to my UG, since it went dark just before I arrived, and the ODO for the code would indicate both were set about 7 miles after starting my drive.. and at that time no MIL or warning of any kind.. so if it was not bizarre enough already, it just got more strange. The other codes were expected from the test of unplugging the ABS, so nothing strange there but the U0100s on the way there with no other indication is strange indeed, and the accumulaton timer for both is 0, where for the expected ABS induced codes there is the expected 47 and 48 minutes which matched the drive time with the ABS unplugged. The P0501 was 28 minutes, which also makes sense, since it did not come in right away, it was part of the way on the drive before that one popped up.

So in summary, while I did show that the ABS seems to not be part of the problem, making the only thing left to be the wiring harness.. I guess it's good I already have it. The new U0100 codes again point to an issue between the ECM and the rest of the network, which is where that harness covers at least half the distance, of course with my luck it will end up being the other side of the harness between the TIPM and that engine connector, though part of that is the conection to the ABS from TIPM, and thus far that appears to not be a failure point.

 

[MichaelOverfield]

Just because it's parts shotgun time.. I found the ABS module for $124, and decided to go ahead and replace it. No change; as expected.. but I now have spares for every vital to run computer on this thing.... so I guess I can store the spares in an EMP-proof faraday cage, just because. In any case, next step is going to take some set up time.. the engine harness to the TIPM via the main connector is going to be replaced when I do major maintenance including the valve lash adjustments and new crankcase filter, install head studs, and revise the intake/grid heaters. So it may be some time before I have another update on this issue. If the harness doesn't do it, it will be a reset switch install time.. Though it's unclear if it will work from testing, but there will really be only the other side harness from the big connector to the TIPM, and since ABS has had few issues, though not zero issues... it's less likely to find the problem there, but given the nature of this saga, it's not wise to rule out less likely.. given my luck, the issue will be the very last thing I replace (which may well be the last possible thing to replace)!

 

[MichaelOverfield]

So, the saga continues.. after being in many parts all winter, and doing a replacement of the engine to body wiring harness.. the truck is back.. and the results are inconclusive. Out of 3 drive cycles, 1 had the same issue pop-up, the other 2 seemed to hold communication with the ECU. So I would say, not fixed, and whatever strange thing this is.. It's looking like it's here to stay. Only thing not yet changed out now is the harness from TIPM to the big connector on the firewall that connects to the ECU harness. Odds it is the issue are low, really low, because the situation defies logic.. it's not that the ECU is shutting down ALL communications, just communications with any scan tool.. that seems to me that is has to be a software thing in the ECU itself, but same issue on 2 different ECUs and no one, a I mean NO ONE seems to be seeing this issue, that I can find, and I've definitely been looking.