The Top Mitsubishi Galant VR-4 Resource

Join the best E39A 1991-1992 Mitsubishi Galant VR-4 community and document your GVR4 journey.

  • Software Upgraded - Reset Your Password to Login
    In order to log in after the forum software change, you need to reset your password. If you don't have access to the email address you used to register your GVR4.org account, you won't be able to reset your password. In that case, follow the instructions here to regain access to the forum.

Wanted to share a solution to a recurring problem I've had

broxma

Well-known member
Joined
Nov 16, 2009
Messages
911
Location
San Antonio Tx
It's technical in nature so I am putting it here. First off, I wanted to say there should be more problem resolution posts around here considering the age of these vehicles and the issues we run into. That being said, here is what happened. Hope it helps someone later on.

The setup is somewhat unimportant but my car is running an Evo VIII ECU so OBDII protocol. The first time this happened, I presented with 3 CEL codes, P0340, P0335 and P0122. The car sounded like it was running on two cylinders. The solution the first time was a pinched wire on the COP plate on the head.

Few days ago, I'm strapped to the dyno and I present with the exact same codes, but now, the car is running fine. It does however have full blown knock above 3000 RPM. The only recent change to the car has been the DNP manifold, the new wastegate and a new exhaust. I log knock and see at 3000 RPM is shoots to 9 counts of knock and stays there. All other functions of the car are totally normal. AF ratio's are kind of odd, bouncing a lot more than normal. So what is causing this?

Same thing as before but different. Essentially, either the starter wires or the alternator wires are arcing to the block or a frayed wire at a connector or thin shielding is allowing positive juice to get to the block. What this does is it throws any sensor with a 5 volt reference signal (Cam, Crank, TPS) into a fault. it also sets the O2 sensor into havoc mode. Since the car was still running on all 4 cylinders I could be fairly sure it wasn't part of the ignition causing this. Since I had pulled the same codes this time as before, I worked under the premise that the problem was going to be another positive wire arcing on the engine itself. In fact upon inspection, there were several places it could have come from. The starter solenoid trigger wire was exposed at the connector. The alternator wires had rubbed thin against the alternator due to my wiring route. The secondary igniter, which I subsequently removed, had several wires with cracks in the shielding.

Essentially, at one of those points, I was shorting onto the motor. In addition to throwing the three codes I mentioned, after I fixed it, the idle totally smoothed out, even without any idle control. So, if you are doing a restoration or hunting down electrical gremlins, remember to double check the little things. All the things I moved or repaired today were things I had gone over previously and assumed I had done right the first time. Anyone who knows me or has looked at the work I put into my VR4 knows I'm not one to cut corners, so even the best of us sometimes screw up. Hopefully, if anyone has similar problems in the future they will find this post, using search hopefully, and get in touch with me. That being said, tonight was a win for deductive reasoning and electrical troubleshooting. Nothing irks me more than trying to find the source of an electrical problem. Hopefully, none of you will have to spend the 5 days it actually took me to figure out what was wrong if this happens to you.

/brox
 

curtis

Well-known member
Joined
May 4, 2003
Messages
11,892
Location
Clarksville TN
Sweet man good info right there. Glad you figuired it out, chasing sparks sucks.
 

524of1000

Well-known member
Joined
Oct 15, 2008
Messages
574
Location
San Antonio, Tx
Yeah... we're back to chasing the problem again /ubbthreads/images/graemlins/banghead.gif
 

broxma

Well-known member
Joined
Nov 16, 2009
Messages
911
Location
San Antonio Tx
The car went about it's normal day for about 30 miles then suddenly about 2 miles from the house started the exact same thing it was doing the day before. I pulled the codes again and the same ones came up. I know the knock detection is false at this point so driving is not an issue but 9 counts of knock is 3 degrees of timing and the O2's are all over the place again. Steve and I took the main harness, the one that runs behind the radiator, off tonight and I am going to be doing work to it tomorrow. If all else fails and I cannot find the problem, I will simply rewire the entire car.

I also figured out why the car knocks at exactly 3000 RPM. The map I use for the Evo ECU has a setting for timing knockdown below 3000 RPM. I assume for increase spool on large frame turbo's, but who really knows. Anyway, it seems this also has the effect of eliminating knock detection at these levels, so it ignores the false knock signal I currently have, until this RPM is hit. I further did some testing and realized I could in fact get the car above 3000 RPM without any false knock if I stayed essentially in closed loop. At any point above 3K the transition from closed to open loop fueling triggers the knock sensor. So now I am trying to think of what electrical parts become active only during open loop operation. We also noted the knock sensor appears to be frayed a bit right at the point the wire enters the sensor. Nothing is touching this area but it could be causing some inductive interference. How it would be affecting the TPS, cam and crank signal I am not sure.

A while ago, I saw a post showing a rewired engine with some fancy braided line. Anyone know what that stuff was, where to get it in large amounts?

/brox
 

Andy_S

Well-known member
Joined
Jan 30, 2007
Messages
982
Location
Shithole Wisconsin
I have been experiencing similar issues with my custom wiring harness. My knock sensor has been consistently receiving interference from the CAS. I am using shielded wire, however I'm getting an almost perfect square wave from the hall-effect sensor. It counts up in steps as the knock total reaches the max reading of 43. I went to my local electrical supply store and bought some braided metal sheathing. I'm hoping the additional shielding will eliminate my problems. Another thought I have had is the factory knock sensor doesn't have any shielding on the sensor side of the harness. If I'm still having issues, I may even shield that portion of wire as well.

I'll update when I get it all redone. Maybe some of my issues are similar to your own?
 

broxma

Well-known member
Joined
Nov 16, 2009
Messages
911
Location
San Antonio Tx
Well I'm using an Evo knock sensor since the 1G sensor is deaf to the Evo ECU. Since the knock sensor isn't exactly magic or anything, I'm thinking about getting some steel braid and wrapping it. I am still trying to figure out why I get codes for the other sensors though if it is an interference problem. Tomorrow I am putting everything back together and I suspect it will run normally again, for a while.


/brox
 

broxma

Well-known member
Joined
Nov 16, 2009
Messages
911
Location
San Antonio Tx
Right so here is the big finally update to end all problems with the car.

So it has been about 10 days. In that time, I have rewired the entire engine harness. I am now so familiar with which sensor goes to which pin on the ECU I am ready to choke myself. I have thus learned a great many things from this experience in the field of automotive electrical problems, so much so now, if you had the money and the need, I could make a totally custom harness in about a day. So here are a few things I picked up along the way.

First and foremost, if anyone is looking at the circuit diagram and trying to figure out where body ground point 7 is, it's the one near the kick panel in the cabin. The 12v power for almost every sensor is run through a single factory spliced harness. All 12v wires lead off this one wire. Almost all the grounds for the engine sensors will trace back to body ground 7 which if you can be sure, should be a pretty good ground. The TPS sensor signal wire branches off many places but it should definitely be uninterrupted to the pin on the ECU. If you have resistance between the TPS connector(Pin 2) and the ECU(Evo 84/DSM 19) you need to trace the signal wire back as far as you can through the harness. It will branch off and go through the cabin sub harness. At that branch, run a line to the pin.

Upon doing an Evo ECU swap, use the EDM(9653xxxx) ROM if your state does not do OBDII testing, even if you have the black top hall effect CAS. The USDM ROM for the Evo ECU seems to have a problem interpreting the DSM cam/crank angle positions. This will present as a recurring P0335 and P0340 CEL with no solution.

The knock sensor has a braided steel sheath surrounding the wire leading to the sensor. This braided steel, in addition to the braided steel surrounding the O2 sensor wires, is in fact grounded and grounds at body ground 7 in the cabin. If this wire is frayed and pulled back at the knock sensor, it may default to "Unplugged" and on the Evo ECU, will default to 9 counts of knock at any detection level.

The injectors only have 4 wires running into the ECU even though there are 8 wires total leading to them. The resistor box on the firewall has one large power feed and 4 smaller wires which lead to the individual injectors. These wires are the correct resistance power feed for the injectors and the wire leading back to the ECU from the injector is the return. The ECU trips the circuit by allowing the return path from the resistor box through the injector. Even if you cut the wire bundle at the resistor box, and randomly ran the 4 wires back to any of the injectors, they would work. The injector timing is all done with the return lines by the ECU at the 4 individual injector pin locations.

The ECU only uses about 40 wires total to run the engine and some of those may be able to be removed still. The round two pronged plug at the T-stat housing is for the AC and if you removed the AC, you can remove those wires entirely, although one is a ground at guess where? Body ground 7.

Pin 40 on the Evo ECU is the sensor ground return, not 42. Oddly, the car will operate just fine with the ground going to 42.

If you override control of the cooling fans manually, use a body ground and run power from a switch in the cabin, the large power connector near the hard lines on the passenger rail will only have 2 functional wires running through it. The majority of the wires at that connector seem to be power wires to control the fans.

You can remove in excess of 200 feet of wire from the car with no change in function. The main engine harness, with the TPS, cam and crank position, knock sensor, ignition and the injectors, is about 4-6 feet longer than it needs to be if you just run the wires directly to the sensors from the firewall. You have to be careful about the injectors though since in factory form, they come around the back of the motor. This means if you have the harness completely apart, all wires exposed, the longest injector clip will the the one closest to the throttle body and likewise have the most extra wire to remove.

The igniter pack can mount anywhere and the firewall right behind the throttle body, towards the bottom has a nice place for it to fit. In total, I have only 3 wires which go all the way across the back of the motor, and those are the three wires for the ignition coils, of which only 2 of those come from the ECU. One wire is a ground shared with the igniter pack and the timing ground to adjust the CPS. All other wires go no further than the knock sensor. And speaking of the knock sensor, the Evo knock sensor connector will pull almost all the way back in the cabin. Is it totally bundled up in the harness and in fact the wire leading to it is about 3 feet too long.

I've got a ton more things I could post up but hopefully, given the right keywords, someone will find this mess and shoot me a PM if they are dealing with wiring issues. Needless to say, I now consider myself, probably in error, pretty damned educated about the engine electrical circuits.

/brox
 

Brunoboy

Well-known member
Joined
Apr 25, 2008
Messages
2,880
Location
San Bruno,CA Home of SFO
good notes, i accidentally cut the knock sensor instead of the o2 sensor for an LC-1 narrowband install. I had to trace the Knock sensor back to the metal braided line without these notes aha. Also I noticed a lot of my wiring has been cut and deleted so that is true. Nonetheless its good to know what wires do what and what can be deleted.
 
Support Vendors who Support the GVR-4 Community
Boosted Fabrication ECM Tuning ExtremePSI Fuel Injector Clinic Jacks Transmissions JNZ Tuning Kiggly Racing Morrison Fabrications RixRacing RockAuto RTM Racing STM Tuned
Top