How to disable TCS from affecting ECU

[NSX2398]

As some of you may have read, I have been fighting an intermittant issue for about a month now where my TCS light came on .. followed soon after by the CEL. When checking for codes, the most common one has been a 3-1 which points to a problem with one of the signals between the TCS and the ECU and the troubleshooting guide focuses on either a TCS module problem; an ECU problem; or an inter-module wiring issue .. the first two of which are going to cost me big $$ and if I guess wrong, there are no returns.

For unknown reasons, some times when I start the car, the lights will both be off and then without the car moving, the lights will come back on. The latest variation was that I was able to drive about 4 blocks before they came on but this was followed by a 'click' behind the driver seat followed by the CEL going out. Which led me to wonder if there was some way to disable the TCS from affecting the ECU. And no .. pressing the Disable TCS button doesn't do anything if the TCS light is on.

I don't really care if my TCS is disabled, but I do care about the impact it has on the ECU because when the CEL comes on, it won't idle lower than 1200 RPM (like the choke is on) and I'm afraid that driving any extended distance might cause my cats to burn out if it is truly running rich. At the moment, this is keeping me from attending NSXPO this year.

Any suggestions?
Tks ...Ian

 

[Briank]

Cel code is 31 is an automatic trans signal issue? Are you sure its not a 1 and a 3, 1 being front o2 sensor and 3 being a map sensor. What are the TCS codes?

Since you drive your car year round I would be looking at connection and grounding issues.

 

[NSX2398]

Cel code is 31 is an automatic trans signal issue? Are you sure its not a 1 and a 3, 1 being front o2 sensor and 3 being a map sensor. What are the TCS codes?
.

Sorry Brian .. I meant the TCS not CEL code is a 3-1. The CEL is on but doesn't ever blink out a code like the TCS does when you short out the service connnector.

 

[TANTO]

You can simply pull the TCS fuse and use the outcome to help you see whats going on.

 

[Old Guy]

Ian:

If you are still suffering the TCS blues, I thought I might add some suggestions and comments. You may have checked these already. Also, my car is OBDII so some of my suggestions may be off for your 91 car.

The 3-1 trouble code for the traction control system indicates problems on the serial data line between the ECM and TCS. From the shop manual, it appears that this is a two way communication port and the TCS and ECM do a handshake on start-up. With the ignition on - TCS starts signaling the ECM, once engine is started the ECM receives the TCS signal and signals the TCS back. If you have a serial data line fault, I would not expect the MIL light to come on until the engine is started.

On the OBDII cars, absence of a signal on the serial data line will cause the ECM to light the MIL and generate a DTC (P1676 or P1677 on OBDII). As such, my guess is that you cannot simply disconnect the TCS from the ECM without having the ECM generate a DTC.


According to the shop manual, the failure options for the 3-1 trouble code (as you have noted) are:

1. The serial data line has an open circuit or is shorted to ground.
2. The ECM is faulty.
3. The TCS is faulty.

Your description of the problem makes me think that you have an intermittent open or ground on the serial data line (which would be a best case scenario). This could be a failure in the actual wire (between terminal 24 on the ECM connector F and terminal 14 on the TCS 20 pin connector) or a problem with the terminals themselves. I don't know about your '91 car; but, on later cars the serial data line is shielded. This shield is grounded at the TCS end only through terminal 5 of the TCS 20 pin connector. As such, it is possible (probably a low probability) that you could have a ground located in the middle of the serial data line without having damage to the external jacket on the cable between the ECM and the TCS. From this thread and your previous thread on this problem, it wasn't clear to me whether you had checked for problems on the serial data line. I suspect that you have a shop manual; but in case you don't and for others that may not have access to a manual, the shop manual test procedure for the serial data line is as follows:

Ground or open check

Check for ground- pull ECM F connector (26 pin) and TCS 20 pin connector. Using an ohm meter, check for continuity between terminal 24 on ECM connector F and the body ground. Continuity indicates an insulation failure on the serial data line. The shop manual doesn't include this but I would also check for continuity between terminal 5 and terminal 14 on the TCS 20 pin connector. Continuity here would indicate a failure between the serial data line and its external shield (you could have this failure without a body ground on the serial data line).

Check for open - with both connectors removed, you could use your ohm meter to check for continuity between terminal 24 on the ECM F connector and terminal 14 on the TCS 20 pin connector. This test is not in the shop manual.

When doing the ground and open checks, having somebody wiggle the cable which may help to show an intermittent failure in the cable.
​

The manual also has checks for the ECM and TCS ports on the serial data line (to determine a TCS or ECM failure).

ECM check - reconnect the 26 pin ECM F connector. Turn ignition swich on (II) and check for voltage at the back of the ECM F connector on terminal 24. It should be approximately 5 volts to ground. Absence of voltage indicates a faulty ECM. You can also check for voltage at terminal 14 of the still disconnected TCS 20 pin connector. This gives you a double check on the continuity of the serial data line.

TCS unit check - turn ignition off, reconnect the TCS 20 pin connector. start the engine and measure the voltage between terminal 14 at the back of the TCS 20 pin connector and body ground. The best way to do this would be with an oscilloscope as there should be a square wave toggeling between 0 and 5 volts on the data line (the shop manual does not specify the frequency). In the absence of a scope, an analog voltmeter should read between 2.5 - 3.5 volts. The shop manual states that absence of voltage indicates a TCS fault.
​

Both the TCS and ECM use a resistor connected to 5 volts to pull the serial data line up to 5 volts and a transistor to pull the data line to ground. This is how they create the data bits, although I have no clue as to what communication protocol Honda is using. If the ECM and TCS check out in an initial test, I don't expect that there would be an intermittent internal fault in either unit associated with the serial data line.

Given the intermittent nature of your fault, if it continues, it would be nice if you had access to a hand held portable scope such as made by Fluke. You could connect it to the serial data line and watch the signal to see if it goes away on you as you drove the car around.

I have a scope. Its relatively small but requires AC power. It could be used to check the operation of the TCS with the car parked (in desparation it could be powered off a small inverter to do a mobile test if you had access to one). Since we both live in Regina, if you are still having problems with your TCS and want to pursue checking the serial data line operation with a scope, PM me and maybe we can make some arrangements.

 

[NSX2398]

Holy crap! What a detailed post! Thanks very much for taking the time to offer suggestions. I don't know where you got all the details but I'm guessing you're either a Honda tech or have an electronics background.

As an update on my situation, I should add the following:
1. Over the last couple of months, the problem (TCS code 3-1) has gone from somewhat intermittant to nearly permanent.
2. I do have a shop manual and as you pointed out, the 3 causes are TCS failure; ECU failure; or wiring. When the code went more permanent, the shop manual seemed to point more to a TCS problem
3. However, prior to doing that, I did follow the diagnostic routines and although I don't recall the exact pin at the moment, I did follow the guidance to use an analog meter and look for a fluctuating voltage which would be indicative of a pulsing voltage. There was no voltage at all, which again pointed to the TCS.
4. So, I took the TCS out and having experienced catalytic capacitor failures on the stereo, I decided to open it up and sure enough there were 4 caps with bulges and BrianK confirmed (via picture) that they were suspect.
5. At the time, I was still hoping to go to NSXPO and with an Acura TCS replacement running about $2000 and 6 week wait, I decided to grab a used one off eBay which supposedly came out of a working car. Popped it in but same problem
6. A few years ago, I got a replacement ECU chip from Prospeed and kept the original so my next step was to try swapping chips although that wouldn't necessarily eliminate other parts failures in the ECU module. Unfortunately, my son had moved out of his condo and was temporarily storing his stuff in my garage and I couldn't get to where I thought I had put the chip. We moved him on the weekend and today I think I found the chip so I might get a chance tomorrow to open up the ECU and swap chips.
7. Most of the ECU continuity tests in the shop manual required me to have the special ECU test harness so I kind of gave up going down that route but I will definitely give some of your ideas a try.

Hopefully I will have more to report by the end of the weekend.
Tks again for offering to help.
...Ian

 

[Old Guy]

I hope your weekend tests work out.

As a final question, when you did the check for the square wave voltage on the serial data line with the engine running, had you already confirmed that the serial data line was not shorted to ground before doing the voltage measurement? If there is a short to ground on the serial data line, this will also result in a 0 volts measurement which could be interpreted as a TCS failure (could explain why your Ebay TCS unit also fails to operate).

A bulging or leaky electrolytic capacitor is not a good sign; however, it doesn't necessarily mean that it has already failed although it is likely well down the path to failure. If it turns out that your serial data line is OK and you want to experiment with capacitor replacement, send me a note. With a solder vac (yes I have one!) simple capacitor replacement is about a 2 minute/each job. The uncertainty wil be whether the local electronics supplier has the caps with the appropriate ratings in stock. At the low voltages in the TCS, electrolytics are cheap, probably less than $1 each unless they are big which I would not expect on a digital board.

I know zip about the Prospeed replacement chips and don't know a whole lot about OBDI and OBDII compliant EMS units; however, my take is that when they rechip a stock EMS, it primarily consists of changing the values in the fuel map look up table and perhaps fiddlig with some of the other settings (VTEC and rev limit?). As such, if the Prospeed chip worked with the TCS when it was first installed, I would not expect some problem with the code on the chip. If Prospeed did something more than just optimize the values in the look up table, it might be possible that it is interacting with some of the code that manages the error codes or other operations (ghost in the machine sort of thing). That sort of thing might not have shown up until the TCS started creating problems. I had a friend who rechipped his car (not an NSX) which caused the EMS to generate all sorts of invalid error codes. That said, if you have the original chip, it is certainly worthwhile putting it back in to rule it out as a source of the problem.

For my own interest, never having pulled apart my EMS and not having any immediate intention to, did Prospeed just supply a new EEPROM or EPROM or was there something more in the replacement?