• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

help with limp mode issue


New Member
a month after taking my 05 Acura TSX to Dealership for a PCM corrosion recall my car went into limp mode (while it was pouring down rain). the DTC was U0107 (lost communication with TAC module). i followed the factory service manual flow chart & it led me to replace TAC module.. did that..no fix. all the voltage & continuity checks were normal. the only thing left that could be the problem (as far as my knowledge takes me anyway) is the PCM is to blame.

i had removed the interior trim on passenger side to access the PCM connectors but eventually I got frustrated and decided to just pull more trim out.. took out the drivers side to reveal the other side of the PCM & found it insanely corroded!! IMG_20191105_071840~2.jpg
when testing some of the PCM pins I had tried to remove all 5 connectors from the PCM but 2 wouldn't budge!
for the recall they were supposed to put a "waterproof" cover on. i found the cheap plastic piece just thrown in the spot where the PCM sits but not covering it at all. after seeing this drivers side angle & the all this corrosion i did a little more research and found out that the PCM was supposed to be removed & the cover bolted back down with a bolt that holds the PCM. I'm guessing the reason it wasn't done in my case could have been due to those connectors being stuck. the only way the mechanics were instructed to actually replace a PCM was if ONE particular DTC was present (that stands for no communication with PCM.. which if a car was driven in there's no way this DTC could have possibly been active!)

anyway, into my questions...
1) even tho the TAC modules ground circuit had continuity I'm curious what you guys think would be a normal resistance value for this particular circuit?

i feel pretty certain the problem is the PCM but would still like to be positive the TAC module ground wiring is within specs.

2) Is it just me or would you agree that it's ludacris to assume a corroded ECU/PCM would always throw a code for the PCM not communicating?! I'd think that the issue could result in numerous different problems depending on which components on the PCB were affected?! am i wrong to assume such a thing?!


Active Member
You need to remove the pcm.
Inspect the harness connectors for corrosion and repair them if found. If its there, the terminals inside the harness are likely heavily oxidized and may need replacing. They tend to fail mostly where the wire crimps to the terminal, cases like this will lead to confusion a lot of times because when you test at the back of the wire it appears good. When you test at the terminal, its easy to push it enough for the bad joint to connect and appear good so keep that in mind when testing them.

Next open the pcm and see if the corrosion has made its way inside. The unit should be sealed and the pcb covered in conformal coating but honda/acura where a little sloppy about this on older models. If its inside clean it up with some isopropyl alcohol, rework any oxidized components and maybe you will get lucky.

Corrosion/Oxidation has its way of doing odd things. Anything from one problem to many problems. intermittent or consistent. So it's not ludicrous at all.

It is indeed possible to drive the car and have a dtc for no communication with pcm. Those dtc's are set in other modules if they are having issue's talking with pcm. It rarely means the pcm is dead.

If this code is active (U0107 - lost communication with TAC module) and your car is starting ok then we know this much is true:
-The pcm is working ok on the data line. (the immobilizer needs to talk with pcm to allow starting & your scan tool connected to read the dtc).
-The TAC and pcm are not talking to each other
So my suspicion would be the TAC has lost power/ground or the data line from the TAC to pcm is shorted, open or heavily oxidized making it a poor connection.

If its a history code then it could have been some moisture that caused a short. Clear the code and the car will not be in limp mode any longer and you would only need to worry if it happened again. (still clean up all the connectors and check inside pcm)

This is where you need a good scan tool. You need to scan all modules on the car to see who lost comms with who and if you can even connect into the TAC module itself. This would help gather much better data and narrow the possibilities down a lot.
Last edited:


New Member
THANK YOU for your input!
I actually did not back probe the wiring but unhooked the connector to do these tests. I'm going to redo all the wiring checks by backprobing and see if i get diff values..I hadn't even thought of that so I really appreciate the tip!

each wire the service manual said to test (relating to this DTC) were thankfully all located in harnesses that I was able to remove and I did visually inspect the terminals on all 3 harnesses (& PCM plugs) & nothing was corroded. HOWEVER as I mentioned in my previous post there are 2 PCM harnesses that will NOT budge so it's been impossible to remove them!

I realize I really need to get to the PCB inside & inspect it for corrosion.. it's just been something I've put off since those harnesses are stuck & the space where the PCM is located is small & hard to get to (esp with my neck issue).

when I was asking about the code for PCM communication, or lack thereof actually what I meant was.. ok so the dealership instructions say that the ONLY way their mechanics are supposed to replace a PCM is if corrosion is present AND THEY PULL THAT ONE CERTAIN CODE that means the PCM isn't communicating.. what I was trying to ask is would a corroded PCM definitely throw that one particular code, or is it possible that it'd throw any number of false codes, or no codes at all even? basically is a PCM smart enough to know when corrosion is causing issues within it's PCB?!


New Member
oh yes and I have tried to clear the codes multiple times but it never goes away. I'm working on finding a cheap way to run bidirectional diagnostics. I just bought a cord that would work with ford but supposedly isn't what I'd need for Honda.


Active Member
what I was trying to ask is would a corroded PCM definitely throw that one particular code, or is it possible that it'd throw any number of false codes, or no codes at all even? basically is a PCM smart enough to know when corrosion is causing issues within it's PCB?!
It could do any one, none or all of what you listed when damaged.
What is needed for the code you listed is only if the pcm and the tcm can not send messages between each other. You could have all kinds of failure inside the pcm and it not throw this code.

You can try cleaning the connectors with some qd electronics cleaner.


Well-Known Member
Most Helpful Member
+1 to what narkeleptk said above.
basically is a PCM smart enough to know when corrosion is causing issues within it's PCB?!
The PCM is not specifically able to detect corrosion, but if there is an anomaly in any part of the circuit (whether that be outside, or inside of the PCM) which returns a value outside of an expected range, the PCM will trigger a fault code.
Corrosion, which tends to result in high resistance in a single circuit, can play havoc with expected values, especially if that corrosion spans multiple pins in a connector, where voltages can appear on circuits that are not expected to return a voltage.
Uxxxx codes are set by the control unit which is sending requests, but not receiving a response.
Basically a "Hey, xyz is not talking to me."


New Member
i am extremely grateful for the responses so far..i was asking about the corrosion-code correlation to get answers from knowledgeable people to hopefully help prove my point to Acura that their recall did not adequately solve the issue with their ECM corrosion problem... just in case I decide to seek recourse.

anyway, onto my particular situation..i was finally able to remove ALL the PCM connectors.. now I'm stuck at getting the bolts removed to get the PCM out to be able to open it up and see if the corrosion has made it to the PCB. the connector terminals both male and female appear to be clean. but I'm interested in the resistance values.. which if you read my original question is what i was mostly trying to ask in the first place.. can anyone tell me what the resistance values should be for circuits related to the TAC module.. like the serial line, both input and output at the PCM, any sensors related to the ETC maybe.. grounds & anything else you think is relevant.. the resistance values because the voltages are all normal and there IS continuity at all the circuits the chart had me test.

oh yea and to follow up on what someone said about the different modules.. my vehicle has the ISO-9141-2 protocol so apparently the only control module is the PCM? there is no BCM or TCM.. the body control stuff is handled by the multiplex control which is built into the relay/fuse box area beside the drivers door. it's not a CAN bus system but K-line. i wish i knew the order in which the drive by wire system communicates with each component (as in sensors, controllers etc) bc i feel like it'd be helpful but I haven't had luck finding that info.


Active Member
Since yours is a all in one unit then it makes sense they do not want to handle it under recall. The recall is probably specific to a certain design flaw that always results in failure of the transmission controller. Something like poor soldering of the mcu or a small hole from factory used to inject conformal coatings in the ecu not being sealed properly. So while you may be experiencing issues similar, it still may not be related to the official recall. Keep in mind they will want to cover as little as possible.

The MUX (fuse box) honda and acura use is the same thing as a bcm, just a different name.

You can get schematics here: https://techinfo.honda.com/rjanisis/logon.aspx
I'm cheap so I always do the $10 (1 day) "express" system its messy and hard to search.
You may want to do "Service Information System" which is $20 (1 day) looks like they have pinpoint testing as well which may give you the resistance value's your looking for. Be warned, honda/acura's system is one of the worst dealer networks I have used but still useful. Since the sub is only for one day, make sure to save all information to your pc you feel is useful for later referencing.

Latest threads

EE World Online Articles