Hi everybody,

I've a 94 Cougar (4.6L, 38k) that purred like a kitten until a few months ago. After being parked for a week in damp weather (Newfoundland), it started having intermittent flats (as if the spark was missing on certain cylinders), most of them being rhythmic (it looked like one cylinder went off for a few seconds, then started running again). This evolved quickly into a rough running at any rpm, random (no pattern) and intermittent (the engine would run just fine for 2 minutes, then would start sputtering - with obvious loss of power - for 5 minutes, then come back to normal run for a bit.

The ECM never put out any codes except for a misfire - during one more severe episode of sputtering. I've been driving the car with a ScanGauge attached at all times, to keep an eye on what's going on. Here's what I did so far, to no avail:

- replaced spark plugs, ignition coils, ignition control module. I did not replace the cables because there was no repetitive sputter after the first week of rough running, which would have happened if one or more spark plugs would have had no spark due to a bad cable

- disconnected injectors (one at a time) to identify one that's not working. Found a dead one (probably plugged, it wasn't clicking), replaced them all with freshly cleaned remans, no change

- checked fuel pressure while the car was parked - no problem

- engine coolant temp and air intake temp are read correctly

- TPS, RPM reported OK

- disconnected MAF sensor and drove the car w/o it, the stutter improved a bit,but never went away. ECM put out the right code, and it cleared it once I reconnected the MAF.

- the oil pressure is a bit low (it even went to the lowest point of the "normal" range on the gauge for a few minutes during a hot summer day, but then bounced back to about 1/3 of the normal range.

Now here comes the weird part: during stutter episodes, the ScanGauge reports some incredible fuel economy (2.5 - 3 l/100 km while driving at 40-60 km/h) which is unheard of from a 4.6L engine. The instant fuel economy jumps back to what I'm used to (8-10 l/100 km) when the stutter stops. When I had the MAF sensor off, the reported fuel economy was shown under 1 l/100 km (sometimes it went straight to zero while driving on a highway...) The engine also knocks louder than I ever heard it - when everything was fine I hardly ever heard it knocking, unlike now. All this makes me believe that the ECM is either getting wrong info from the sensors (CPS, O2) or cuts off the fuel way too early, leaving the engine running with a very lean mixture. However, nobody could confirm this is the case, and I don't want to start replacing parts on the engine before I have a good idea what can be wrong.

I wanted to replace the MAF, but if this sensor were the problem, running without it should have eliminated the flats - which is not the case. The stutter continued, randomly, so apparently caused by something else.

I was also about to replace the o2 sensors on the assumption that a false reading of too little oxygen in the exhaust would cause ECM to reduce the amount of fuel injected, to balance the mixture. Nobody could confirm this assumption though ...

Please let me know what you think. This drives me crazy, I love my car dearly and I hate to see it suffering like this. Any ideas are more than welcome..

Thank you kindly,
cougar_fan

 

Is the fuel pressure consistent at the specified pressure while driving, under a load, or even just parked and idling? How long does it take to clear up once it happens?

 

Thank you for the reply, greasemark! I did not drive the car with the fuel gauge connected, but will do that later today to verify what you mentioned. While parked, the pressure is right on the the mark, and stays there once I shut the engine off.

On the second part, lately, the rough run only clears when the engine had been running for less than 10-15 minutes ... once I drive the car longer than that, the rattling/stutter/roughness is there to stay.

Here's a fresh finding, maybe this would help us figure out what my problem can be: I programmed the ScanGauge with four XGauges (O2 sensors output, both banks, upstream and downstream from the catalytic converters). I visually confirmed that my car has four sensors, to make sure I'm not reading crap from the scanner. To my surprise, the LH side (bank 2) front sensor (1) would only go between 0 and 5 (open or close loop). The corresponding sensor on bank 1 would go between 0 and 90 (as it should) in closed loop, and stay at 85-90 while in open loop mode. I thought the front LH sensor was toasted so I replaced it today with a brand new one - NO CHANGE! I then left it disconnected to confirm I'm reading what I should, and the ScanGauge shown zero output - so the ECU sees too much oxygen coming out of the four left cylinders.

Weirdly enough, the ECU did not put out a code for the disconnected sensor in some 3 minutes of running w/o it (maybe I should have waited a bit longer).

At this point I suspect that my ECU is acting up. It either misreads the O2 sensor, or fails to properly drive the LH injectors. A good ECU should have tried to enrich the fuel mix (because it sees it lean) which doesn't happen (my car actually increased its fuel mileage with about 30% since this problem started, so it's not only what ScanGauge claims).

I found an identical ECU on eBay (tested good, according to the seller), it should be here in about two weeks. I'll give that a try, and I'll also report back here with anything new.

Again, please let me know what you think. Any idea, no matter how off the topic may seem, could be the way out of this, and we'll all learn a bit more about these things we love (and hate) so much, our cars.

Cheers,
cougar_fan

 

I took the car for a ride with the fuel gauge connected to the rail. Here are the readings:

- ignition on, engine off: 40 PSI after the pump primes the system;
- idling: 32 PSI
- driving: 32-38 PSI (38 PSI is reached during hard acceleration , close to WOT, 3500-4000 RPM)

It looks like the fuel pressure goes up as the throttle opens and the vacuum drops, which is how it should be, right?

Cheers,
cougar_fan

 

Ok, so it's not fuel related. That rules that out. And of course, you're getting the incredible fuel economy because it's running really, really lean. I call this fuel injection system EEC IV and a half, because it's got the very first year of EEC V hardware, but has mostly EEC IV programming (most of the eec v monitors for o2 and o2 heater aren't there. I'm wondering if you've checked all the connectors for corrosion and fit. Also, wiggle the wiring looms and look for a change (ANY change when wiggiling is a clue that something is going on; if it's running badly, it might clear up for a moment and vice-versa). And you mentioned one other thing that caught my attention. These processors are programmed to set a code after 2 consecutive key cycles; which explains why the code(s) didn't set. Have you checked with either a noid light or a good old fashioned test light if the injectors are getting a good pcm driver signal? Just connect the test light to the not red with whatever (that is the power side)wire and connect the light to B+ and look for a steady flashing; be careful to look for a pattern that starts to shift when the problem occurs. From the sound of it, it could be loosing a bank or any number of injector drivers; they can come and go and be very difficult to pinpoint. Also, the same test (light connected to ground) can be used to check for constant power to the red wires going to the injectors, anytime the car is running. If you see missing blinks for it just quits flashing, that's a injector driver/wiring problem. One more thing; did you check the EGR vavle for sticking? It would be a reach to see this problem, but stranger things can happen. Just disconnect the vaccuum line and drive it to see if it goes away. And since I'm rambling on here, how about a processor with split personality problem? If you're concerned about inputs, which is a good place to start, then force the O2 lean (create a vacuum leak) and monitor the O2 directly with a (digital) voltmeter and watch for a low (< .5 volts, though they'll go to near zero) signal. Watch the scanner at the same time for whatever o2 you're monitoring to respond with a very similar reading. You also can see if the system can be forced rich (spray carb cleaner into the intake snorkel) and look for a high (> .5 volts) reading. If these readings are not obtained, then check the wiring to and from the pcm. If ok, it may be a defective pcm. GOOD LUCK! and let us know.

 

Thanks again for all the info. My weekdays are darn busy at work, but will try to do what you suggested over the weekend and in the evenings.

I suspected that, especially since I had my car rustproofed with Krown twice, and I previously had another electrical problem (the airbag module blew its internal fuse - code 13 I think it was, side impact sensor shorted - which cleared out to code 51 a week or so after. i since replaced the airbag control unit and had no other problems). I didn;t wiggle the harnesses, but I checked most connectors and found them clean, dry and with no rust problems. The only notable exception was the big square firewall connector (linking the engine bay harness with the cabin harness) - this one had traces of rustproofing oil in it, and a few pins were covered with a light greenish, greasy crap (yeah, I know what that means ...). I cleaned both mating parts, scraped each pin individually with a fine file and cleaned the whole thing with alcohol. No change at all, so there was no problem there ... darn.

Not yet, but I plan to hardwire 8 brightness-matched LEDs, one in parallel with each of the injectors, and run the cables back into the cockpit so I can compare the injectors' individual duty cycle by simply looking at the LEDs brightness while I drive. That should show if some injectors stay open a lot less than others, and the real-time status can be corroborated with the hiccups.

Nope, not yet, but will do and post back.

All valuable ideas, THANK YOU !! Will do all this, and post back what I found.

Thank you kindly,
cougar_fan

 

Greasemark, I was lucky enough to find a EEC V break-out box on eBay for about $350, it's in the mail and should be here in a week or two. I'll wait with wiring check until the box arrives, it'll let me do a much better job at measuring signals from sensors and to injectors, as well as identify any cabling problems. I just wanted to post this so you won't be thinking I got my problems solved and forgot to share my experience with everybody else here in a timely way.

Cheers,
cougar_fan

 

No problem. Have fun with the BOB. They're really fun to hook up. But I'm sure you can fix it!

 

A new development: I configured ScanGauge to show loop status while driving the car around.

On yesterday' colder weather (10-16 degrees Celsius), the engine ran almost perfectly while in open loop (the first minute or two), and ScanGauge reported something close to the expected fuel economy.

As soon as it switched to closed loop, the instant fuel consumption dropped and the rough running came back.

The LH O2 sensor reading stayed at the bottom (0-5) while the rest of them were showing normal values (regardless of loop status). So I'm pretty sure this is either a wiring problem (low resistance between the sensing wires from the O2 sensor or a ground fault on that harness) or the ECU's input is toasted. The BOB will tell.

Cheers,
cougar_fan

 

When you are ready to connect the BOB, carefully check the pins and connections on the plug and ECM. At least once hooked up you'll know those connections are clean and good. Keep us tuned in , thanks.