Cold Starts Lean, Warm Up Enrichment Issue

[suprajim54]

I'm having an issue with my cold starts being too lean. I have 550's and Lexus AFM running on a JDM ECU. I do not have any codes showing. On a cold start while idling, Vf shows to be 0 volts and the AEM wideband is showing high 17's or blanks out. Vacuum gauge shows about -14 and the car stumbles quite a bit. Fuel pressure off vacuum is 38 and about 30 with vacuum attached. Once the car heats up to normal range, it doesn't come out of warm up enrichment. If I turn the car off and right back on, Vf is 2.21 volts, the AFR's quickly go to 14.7-15.1, vacuum hovers between -18 or -17 and fuel pressure is about 28 with vacuum and the engine runs smooth.

I've tried to drive the car when still in warm up enrichment mode and the Vf stays at 0 volts and the engine misfires. Part throttle at 30mph and 1500 RPM, the AFR's are in the 16-17 range. I didn't push it too much to avoid breaking anything. Turn off and back on and everything is back to running well. 30mph @ 1500RPM my AFR's are 14-15. Vf is typically at 2.21 but ranges to 0 volts during normal driving. When I go WOT, AFR's drop to 10.

Sometimes if I turn off and turn back on, it will go back into the warm up enrichment mode or at least has the same poor running conditions and gauge readings. Immediately turn off and back on again and all is well. I have driven about 75-100 miles since the battery was last disconnected/ECU Reset. The ISCV clicks when I turn the car off and I had disassembled and cleaned it when the injectors went in. I've done lots of searching but can only find warm restart problems or cold start issues with MAFTPro or AEM. I have an SAFC, but it is not connected right now. Even so, I can't add fuel only to my cold starts with my current setup. Could the ECU temp sensor be failing without it throwing a code? Any suggestions of something else to check?

 

[Backlash2032]

I'm willing to bet the STA signal is not getting to the ECU. Check that circuit.

 

[suprajim54]

I'm not getting a code 43 which is related to the STA signal, but I'll try checking later this evening. Does the ECU use the STA signal in calculating the injector duration?

 

[Backlash2032]

No, the ecu energizes the cold start injection system with the STA signal.

 

[Nick M]

Does the ECU use the STA signal in calculating the injector duration?

Only during cranking. At least according to Toyota Motor Sales. The input signals are too low at starter cranking speed. Put your ohm meter on your ECT sensor before a cold start and see that resistance is according to the manual. If the resistance is low, TCCS thinks the car is warm and it needs to be rich to light off when cold.

 

[suprajim54]

I checked the ECT cold and got 1.968kohms. It's about 85 degrees so it seems like it's within spec.

I left it unplugged and tried starting, just to see what happens. It didn't want to start or if it did it only stayed running for a second at most. If i tried to give it gas, it would die immediately. I plugged it back in and started the car. AFR started at about 15 but quickly went to about 17.5. While it was running, I unplugged the sensor again. AFR's went off the scale and engine stumbled worse and wanted to die if I moved the throttle. I plugged it back in (still running) and idle picked up slightly, but still stumbled. AFR's went to high 16's. I went ahead and let it get up to temp and idle for at least 10 minutes. Vf was showing 0 volts the whole time. It never came out of warm up enrichment.

So, I turned off the car, unplugged the sensor and turned it right back on. Vf went to 2.21 volts and AFR's were about 14.9. Engine ran smooth. I could freely open the throttle and no issues. I plugged the sensor in again and still no change. VF and AFR stayed the same. Checked codes and only got a 22.

Any other ideas?

 

[morganson]

I have a similar problem since i got my car. I would like to know here as well.

 

[3p141592654]

Unplugging ECT sensor causes the the ECU to default to 80C. That is why it performs poorly when cold and you unplug it.

I have a suspicion your O2 sensor is not cross counting when you are having drivability issues. Try monitoring the OX1 port with a good quality DVM. You should see it cross counting after 2 minutes from cold start.

After cold start VF1 stays at 0V until the O2 sensor flips rich, then it goes closed loop and VF1 should start reading normal values. If the O2 sensor never goes rich, it will stay in open loop forever.

 

[super51fan]

Good idea of a lazy o2 sensor.

 

[suprajim54]

Ok, so first thing I did was test the heater circuit of the O2 sensor and it was within TSRM specs. I also checked for continuity between the signal wire of the O2 sensor and the Ox terminal, good there. It seems the wiring was good, so I moved on to the sensor itself.

So, I started the car with meter connected to the OX & E2 terminals at the ECU. It started at 0 volts and as usual engine was stumbling. Vf was also at 0 volts. I let it get up to normal temps, but never any change in the voltage readings. I turned the car off and back on, Vf went to the 2.21 volts and OX was reading about 450 millivolts. I'm using a digital meter and counted 4 or 5 fluctuations. It would go from like 450 to 460 and back to 450, so I counted that as 1 fluctuation. Sometimes, it would go lower than 450 (like 430), but still counted as 1 fluctuation. So according to the TSRM, it should fluctuate at least 8 times. Still no code for O2 sensor, but I'm assuming it could be "lazy" as super51fan put it? I tried this multiple times with the car warm and could only count 4 or 5 changes in 10 seconds. However, it would sometimes fluctuate around 470 or 480 millivolts instead of the 450 when I turned off and back on.

Then, I removed the sensor and connected the meter's positive lead to the signal wire and the negative was connected to the sensor body with an alligator clip. It started at 0, but went up with heat applied from a propane torch. The voltage went up over a volt when heated over a minute. I moved it to and from the flame and it would fluctuate, but didn't seem to respond quickly. While over the flame, it would still show 0-200 for a fraction of a second. I'm not sure how quickly that should change, though. The "on car" test seems to indicate that the O2 sensor may be failing. I don't have a problem buying a new one, but I don't want to unnecessarily replace a good part. I've done too much of that over the years. Could there be anything else besides a failing sensor that could make it fail the on car test?

Edit: Not sure if it makes a difference, but all of my tests have been without the diag block jumpered

 

[3p141592654]

The propane torch test sounds okay. The response test should normally be hard to judge, i.e. pretty fast to the human eye. A half second delay seems slow, but could be your meter not the sensor.

Since your on car test was a bust, try using the conditioned Ox output from the ECU.

1) ECU in diagnostic mode with jumper
2) Warm engine, hold at 2500 rpm and monitor VF1 output. Should toggle between 0 and 5 volts rapidly (> 8 times in 10s).

If your meter is slow, this test won't be any better than the first attempt though. Does your meter have a bar graph mode where it can respond faster than the digital mode? Or, do you have an old analog voltmeter somewhere?]

 

[suprajim54]

I dug out my old analog meter. It's pretty basic, but there doesn't seem to be much of a delay with it. Hoping it's good enough.

I jumpered the diag block and probed the Vf signal at the ECU. Started the car cold and let it idle. Same poor running condition as usual. I let it idle for about 15 minutes, just to be sure I gave the car plenty of time to warm up on it's own. Vf never changed from 0 volts. I ran it up to 2500RPM and the needle stayed dead at 0. When I let off the RPM dropped significantly and the engine stumbled but didn't die. Vacuum went down to -12 before it caught itself and leveled out at about -14. Still no movement off 0 volts.

I turned the car off and back on and I still had 0 volts on the meter, but AFR's were about 15.1, vacuum -17 and no stumbling at idle. I ran it up to 2500RPM and still 0 volts. Just to see what happens, I unplugged the jumper and Vf went to just over 2 (analog scale marks hard to get exact). No fluctuation either way at idle. Brought RPM up to 2500RPM and it never moved from ~2 volts. I put the jumper back in, Vf went back to 0 and stayed there at 2500RPM with no fluctuation.

Following the troubleshooting steps in the TSRM, it points to replacing the sensor. The only thing I didn't do was disconnect the PVC hose. Would I just need to disconnect any one of the hoses between the TB and accordian and plug both ends? I guess I'll go ahead and order a sensor so it can be ruled out. It'll take a couple days to get to me anyway and will just return it or keep as a spare if I find some other solution.

 

[suprajim54]

New O2 sensor came in today. I installed it and absolutely no change. Still lean AFR when cold and no issues when warm. When warm, there's no cross counting of the new sensor at 2500rpm still stays dead at just over 2 volts. When I let off the throttle RPM drops quickly to 0 and recovers to the ~2 volts. When in warm up enrichment it stays at 0 at all times. I'm using a JDM ECU since I have no EGR. I'm going to put a US ECU in this evening to see if warm up enrichment behavior changes.

UPDATE:
Changed to a USDM ECU. Started cold and no change. Warmed up for 15 minutes and VF never went off 0 volts. I tried adjusting fuel pressure up to 40psi and down to 20 psi no change to VF. I turned it back to about 28 psi. I ran it up to 2500RPM and AFR's were in the high 16's vacuum. Turned car off and back on and engine ran smooth. VF now at about 4.91 volts, though. 2500RPM, the AFR is 13.3 and about -13 vacuum, idle is 15.6 and -17 vacuum. VF fluctuates more than 8 times in 10 seconds. I turned off the car and swapped back to the JDM ECU. Started the car and VF still at 4.9, 15.2 AFR and -17 vacuum. I checked the ECT again and got less than 1000 ohms, so still within spec.

What other sensors or components are involved in telling the ECU to come out of warm up enrichment? Something is just making the car not want to make the "switch".

 

[Nick M]

What other sensors or components are involved in telling the ECU to come out of warm up enrichment? Something is just making the car not want to make the "switch".

Do you have a cold lean problem or it stays rich when warm?

 

[suprajim54]

Its a cold lean problem. 0 volts means its pulling fuel out, right? Cold start is showing lean on the wideband 17+. When VF is at 2.5 (2.2 on mine) the TCCS is neither adding or subtracting fuel and the car runs fine with near stoich AFR. I am doing absolutely no changes only turning car off and on. I'm going to double check the THW signal on both ends and make sure I'm getting continuity. Swapping the ECU had no real effect so I'm thinking its gotta be some other sensor or just wiring problems.

 

[3p141592654]

Your propane test was pretty suggestive that the sensor was likely good.

0V can also mean it is open loop. You have verified 12V across the O2 sensor heater?

Things that keep the car in open loop are coolant temp, throttle position, no O2 cross count, high rpm and high load. The last two are unlikely to be the issue.

 

[super51fan]

Hey OP I think you need to raise fuel pressure. My car runs 2.7 bar or about 38 psi with vac hose installed at cold idle. Just a thought.

Also are sure ignition timing is correct? Your engine vacuum seems low. Cold start I am at 17" then 20" once warm.

Would not be a bad idea to back probe THW at ecu. This way you can see what the ecu sees. THW input to ecu is second behind the Airflow meter as far as fueling corrections.

I am assuming normal engine code.

Your wideband shows a steady AFR at 2500rpm.

Was the car run properly before the airflow meter and injector swap? You do have the 7MGTE sensor in lexus airflow meter housing and the "o"rings are in position.

 

[Nick M]

I'm going to double check the THW signal on both ends and make sure I'm getting continuity.

Check the actual voltage at the computer between THW and E2. I think you might see something way to close to 5V reference for a cold start. If THW is equal to it then you can start looking for a break in E2.

 

[suprajim54]

Key in ON position but engine off, diag block is jumpered. THW is 4.97 volts at the ECU and sensor side of the harness. I have continuity back to the ECU on both wires. I checked connections of E2 at AFM & TPS and all have continuity to each other and body ground. I can't find any breaks in the connection. I was getting 12 volts to one of the O2 heater wires. The other one was open. I have continuity from this wire back to the HT terminal at the ECU, but no voltage/continuity between HT and E2 or B+. The sensor is not getting hot when plugged in and key ON.

I started the car and same lean cold idle conditions. I monitored the THW and it fluctuated a bit between 1-3 volts but went down to around 1 volt as the temps came up. I still never got any voltage on the HT connection. Once the car was up to temp, off and back on and VF went to 2.2, HT voltage at ECU was still nothing. I looked through the TSRM, but can't find any information about testing the HT signal at the ECU. It seems the ECU is not grounding the HT pin as it should. I found another thread where the OP had a similar problem, but he had a Code 21. The thread never really had a resolution, though. http://tinyurl.com/lp3zmej Post #30 says that certain parameters need to be met for the ECU to ground the pin. So, I need to figure out why the ECU isn't seeing what it wants to trigger the heater signal to ground, but why would it not throw a code?

 

[Nick M]

Key in ON position but engine off, diag block is jumpered. THW is 4.97 volts at the ECU and sensor side of the harness. I have continuity back to the ECU on both wires.

It is a 5V pull up resistor. And open has infinite resistance, so all the voltage drop happens at the open spot. If you do not have voltage drop in the wire, it has to be at the sensor. High voltage causes a rich condition, not a lean condition. Unless it is open then this happens.

Unplugging ECT sensor causes the the ECU to default to 80C.

And you will be lean.