On to the next quirky thing affecting my car.

I've been running an AEM wideband for a few years, and the AFR averages 14.7 with +/- 0.3 cycling.

Past month, as I drive on the freeway the AFR average slowly decreases from 14.7 to 14.3 and sometimes to 14.1 (cycles to 13.9 sometimes). I can see how the car would run rich with a thermostat that failed open (my recent problem), but I just fixed the thermostat issue (Temps are fine now, and ECU was reset) and the AFRs are still trending down.

If I release the throttle on the freeway briefly, and go back to cruising throttle, the AFR will go back to 14.7 and again slowly trend down over 30 seconds, staying low.

Anyone know what could cause this? Is this a possible failure mode with an AEM wideband? I don't have anywhere near to 50k on the wideband yet, maybe less than 15k.

Fuel system is unchanged from previous setup that averaged 14.7 AFRs on cruise (440s, SFMU with idle pressure at 18psi, intank + inline pumps, 10ms injection duty cycle crusing at 65mph with RPMs at 2700.)

TIA!

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

This is the weirdest phenomenon, just can't explain it. Anyone care to throw out guesses?

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

This is a tough one. I'm stumped. I'll throw a guess out there and say perhaps a very small vac leak somewhere or even a bad gasket? I had some loose bolts in my intake manifold/upper plenum connection that affected my AFR's in a similar manner.

I contributed to the Wiki.

Thanks for the thoughts - that's a good idea, I'll check for leaks this weekend. It seems counter-intuitive though, wouldn't a leak somewhere on the intake side of an engine result in MORE air coming into the cylinder (because the intake is at vacuum), thus causing a lean condition?

My AFRs now go to 13.9 while cruising on the freeway, after slowly falling from 14.7. I release the throttle briefly, AFRs go back to 14.7 and slowly drop to 13.9 and hold. The e-manage has been completely disconnected, there are no extra electronics affecting the sensors, injectors, ignition, or anything.

Long term fuel trims are at -22 to -25 (max) right now while it happens, while short term trims are at 0. This is unusual for my car, both LTFT and STFT typically always remained within +/- 5. My fuel setup is unchanged - I still have 440 injectors running at an idle pressure of 18 psi, or 20 psi at cruise.

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

Gets weirder. Today, I lost the speedometer signal (happens when I the battery cables get moved around, which I've been messing with recently), and when the speedometer reads zero while I'm at freeway speeds, the AFR stays around 14.7 as it should! So the ECU is looking for some specific variables when its at freeway speeds, and adjusting incorrectly. Anyone know what the ECU checks differently during closed loop at 0 MPH, versus closed loop at 65 MPH?

I noted that the cumulative fuel trims were < -25, in other words the LTFT were at -25 while the STFT were at -3 to -5. When I had speedometer signal, if the LTFTs were at -25 the STFT would stay at 0. If the LTFTs were at -22, the STFTs would be at -2 or -3, but still never go below -25.

I checked my vacuum lines, all seem to look ok.

Ooo and just threw a P0137 code 10 minutes ago. This is getting fun.

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

I think you are on the right track. I would start logging the critical parameters and run through to find one that flatlines intermittently. You can definitely chase your tail for days until you find that one parameter. I've had an O2 circuit within the ECU die, dead voltage regulator within ECU, and a coolant sensor connector corrode before. All caused very mysterious symptoms until I happened to see the dead input in the data I was logging.

Corbin
'95 ESI-T

Now with more power and fewer leaks



Humble servant of the Wiki

Corbin you are a sight for sore eyes. I was just thinking to myself, man I miss the days when you and others would regularly appear in threads and provide a helpful answer.

I do feel it's some kind of sensor issue - yet so far I have not been able to witness a temporary dead signal in any of them (unless you count the rear O2 sensor, but the best info I have is that it does not play any kind of role whatsoever in fuel trims). My problem is the refresh rate on my scangauge is too slow, and I can only really log via the e-01, but with my emanage getting repaired I have to now hardwire into each signal line off the ecu one at a time.

Back to MFI 101. ECU depends on these sensors to calculate fuel requirements:
front O2-Doubt it's this, b/c why would the ECU instantly go back to 14.7 if I momentarily release the throttle or hit WOT? It's still using the O2 sensor immediately after those tests! less than 20,000 on the new O2, but it could have been fouled up by past problems.
MAP-MAP fluctuates continously, but don't know what to look for there.
Cam sensor/crank sensor-I'd expect hesitation with issues in the crank and cam sensors.
knock sensor-???I suppose it could richen up the mixture, but I'd expect to see ignition timing adjustments too.
VSS-not a variable, b/c I can remove it and the car runs fine. Speedometer shows no fluctuation anyways.
IAT-a possibility
coolant temp-temps seems steady after fixing the thermostat issue, but it's worth a shot to change.

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

I'm mostly interested in why your LTFT is -25%, when it used to be near zero. To me, that says that one of the sensors used to calculate required fuel has changed drastically and the ECU is reading the narrowband O2 to bring AFR back to where it wants. If your wideband O2 says you are near 14.7 at cruise, then the ECU must be interpreting the narrowband O2 output correctly. I am assuming that you have three O2 sensors (wideband, front narrowband, and rear narrowband). I would look closest at the MAP and IAT. You could dig through the electrical manual and find the expected output of both sensors at two points (absolute pressure versus voltage for MAP and temperature versus resistance for IAT). You could then calculate what the raw output means and decide if that is interesting. Of course, you'll look like a nerd, driving down the road writing notes and punching numbers into the calculator. It would be safer and less embarrassing to buy a proper OBDII logger cable that works with a laptop. I might only use mine once every few years, but I never once regretted the purchase.


Corbin
'95 ESI-T

Now with more power and fewer leaks



Humble servant of the Wiki

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

Just noticed something...

On startup in open loop, the AFRs hovered steady around 11, before going to 14.7 when the ECU entered close loop. I've noticed this a couple times now. However, if I've been subconsciously paying attention to this AFR in open loop all these years, warm up in open loop was in the 13s (rich but not crazy rich like now).

Open loop fuel calculations are done off of ECU maps. Now, I know some sensors need to have signals for the car to even start and idle in open loop - like the MAP. But are they being used to calculate fuel requirements? This thread (http://forums.2gnt.com/dcboard.php?az=set_threaded_mode&forum=8&page=&topic_id=77827&prev_page=show_topic) suggests that only the MAP and CLT sensors are used to adjust pulse width during open and closed loops.

If not one of those sensors... perhaps this is a problem with my injectors...

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

I've always used a cable from: http://www.obd-2.com/

I can't say if that company is still in business as the website looks the same as it did in 1998. I have used that cable with the software available on the RRE site, but I prefer the OBD-2 software. There are probably lots of new things out there that I don't know about. My car is from the '90s and so is most of my knowledge. You really only need three things to calculate steady state required fuel: air pressure, air temperature, and engine RPM. That means MAP, IAT, and CPK sensors. I think you can rule out CPK since that should result in much more noticeable symptoms than just some AFR drift. MAP and IAT should be your first stop. If one of those is reading incorrectly, the ECU will calculate required fuel incorrectly, and have to compensate with LTFT and STFT based on output from the front narrowband O2. I think you will find that the CLT sensor is most critical at start-up on a cold engine and when the IAT is not available (disconnected or considered failed by ECU). At some point with a normal engine warming up, the ECU will stop using the CLT to calculate fuel and start using the IAT. Not being privy to the code, I can't say exactly when and how that occurs. Short story, I would look closely at MAP and IAT output.

Corbin
'95 ESI-T

Now with more power and fewer leaks



Humble servant of the Wiki

The problem has been fixed, but you aren’t going to like the findings.

I never once saw signal loss or fluctuations in any of my sensors through the OBDII port (MAP, IAT, CLT). Which means I had to start throwing $ at this. I was methodical with diagnosis, changing one variable at a time, to see if that one variable affected the rich condition on the freeway.

Items changed, in order:

VSS – I’ve had intermittent speedometer issues for a few years. Thought it was the wiring, because when messing with battery cables/other wires the VSS would switch between functioning and not functioning. Yet the VSS wires/supply voltage/GND always tested to specs. VSS is involved in calculating pulse width, and shares GND and 8+V circuits with many other critical sensors, so I decided to change it in case a bad sensor was causing interference/bad grounding on those shared circuits. Results: No change to issue. But speedometer finally works 100%!

CLT- I’ve been suspicious of my old CLT for a while. Results: No change to issue. And the new sensor is quite laggy compared to the old temp sensor.

EGR vacuum hard line – The old line had a pin hole leak that was patched, but the patch may (?) have been leaking. EGR failure modes seemed like they could be consistent with the freeway AFR problems. Results: No change to main issue. However, initially the car idled and ran much leaner (15.0 AFR). But the ECU soon compensated, and averaged 14.7 on surface streets.

Stock spark plugs – Old sparkies looked like they were subjected to very rich conditions. Results: No change to issue.

EGR – tested with vacuum pump, works appropriately. Moved some coolant lines that were butting up against the egr hose and potentially pinching them. Results: No change to issue.

Took off and cleaned CAM sensor. Results: No change to issue.

MAP sensor – Results: No change to issue from the brief testing time I had – it looked like AFRs were falling to low 14s again. Living in Orange County, it’s almost impossible to get 20 consecutive seconds of 65 mph on local freeways until late at night. However, the car stops cracking and popping on decel, and runs smoother.

New cat converter and gaskets. I have not had a bad cat code for 2 months after washing it in detergent, but decided to give it a shot. Results: Issue resolved. Nice super steady 14.7 crusing AFRs all day long.

I also emptied the catch can and placed a new PCV valve.

So, the most inexplicable issue I’ve seen with this car yet has the most surprising solution. How does the ECU decide to inject more fuel in if a cat/cat gasket is bad? One of the cat gaskets was burned up, but even an exhaust leak this far back has no impact on the major sensors used for fuel requirements. The front O2 sensor should have kept things at 14.7 independent of what’s happening so far back in the exhaust system. It’s gospel in these parts that the rear O2 sensor has ZERO input into the ECU’s fuel algorithm. This 99 RS suggests differently. The sum of the LTFT and STFT now is allowed to go below -25 in total. It would never do that before.

The cumulative result of all the work above is that the car feels like stock again. No rumbling/pops/hiccups/race-car sounds. LTFTs are still quite negative though, and I’m pretty sure it wasn’t like this before.

2gnt: '99 RS-T, killed by a toyota, pending rebuild...
Daily: Volt
Daily #2: '99 EVG ebike- STOLEN by PEDRO

It sounds like you performed some tasks that needed to be done anyway (other improvements despite main issue remaining). You might find that if the ECU suspects that the cat or downstream O2 is failing, it may play with the fueling during low load conditions as a test. Especially on the later models, there are probably lots of little tests and functions that we are not privy to. I am still surprised that a new cat and gasket fixed it.


Corbin
'95 ESI-T

Now with more power and fewer leaks



Humble servant of the Wiki