Mechanics

OXYGEN SENSOR FAILING TO READ, 95 LE SC

1995 Toyota Previa • 165,000 miles

I have been chasing this problem for a couple months. I have had 5 Previas (two of them new and 3 of them SC's) so I am pretty familiar with their maintenance as I do all the servicing. Please read all of this. It gets really interesting. This problem is really elusive and defies explanation. The current status of my 95 SC is:
The engine idles very rough and the rough running continues into low throttle condition. The OX1 sensor (and OX2) is reading lean (0 - 0.05 volt typical continuous). It never indicates rich condition. Initially the P0171 showed up on a trip and the engine was not idling particularly bad yet. P0171 is set when the fuel trims are driven to their max values (computer trying to enrich perceived lean condition). MIL codes that I have been receiving include lean mixture P0171 and cylinder misfire P0300, 303, 304. After replacing the OX sensor and checking its wiring, I began testing, replacing and swapping things with known good components. I only use genuine toyota parts.
All fuses were checked and are good, battery is new and good.
Fuel filter replaced with new, no change.
Fuel pressure checked and is spec.
Fuel injectors replaced with new, no change.
Spark plugs replaced with new, no change.
Valve clearance checked and is spec.
Coolant temp sensor reads ~186 F when engine is warm (Is this warm enough for closed loop fuel control?)

Since I believed the OX sensor readings, I was falsely chasing a lean condition. I can easily look at a real time plot of the ox sensors and fuel trims thru the OBDII port with software that I have and compare this with my “good” 96 previa readings. When I reset the codes and the ECU, the ECU resets fuel trims to zero. Could not then get into closed loop fuel control and the engine began running rougher.

Swapped out ECU with two other known good computers, no change.
Cleaned the MAF sensor, no change.
Swapped out the MAF sensor with known good one, no change.
Cleaned and checked operation of the throttle body, ISC valve and TPS, and adjusted to spec, no change.
Swapped the throttle body/ISC/TPS with known good, no change.
Replaced distributor cap, rotor, spark plug wires with new, no change.
Checked distributor condition (was replaced new ~30k miles ago), no problems found.
Looked extensively for vacuum leak with none found.
Removed and checked intercooler for leaks, none found.
Suspecting the new OX sensor of also being bad, replaced it with another brand new sensor, no change.
Rechecked the OX1 sensor wire from the connector on the ECU (E6 pin5) to the sensor plug pin 3 and to the pin 4 on the D5 data link connector, good continuity, no shorts to ground.
Rechecked the OX1 sensor ground wiring from the sensor pin 4 to vehicle ground, all good.
Load tested the OX1 heater sensor wiring from ground to the ECU connector, all good.
Placing the vehicle in gear on jacks, I somehow managed to get the ECU to go into closed loop fuel control and both fuel trims marched up to their maximum values with the OX sensors continuing to measure constant lean condition.
The catalytic converters and the exhaust gas seemed very hot, so on a hunch, I disconnected the ISC connector and powered the ISC valve closed by directly connecting +12v to the ISC valve +B pin 2 and grounding the RSC pin 3, the engine immediately died from the rough idle. This indicated to me that there was no air leak, otherwise the engine would have sped up from the improved air fuel ratio. I confirmed that the engine was running very rich by opening the ISC valve (connecting +12v to the ISC valve +B pin 2 and grounding the RSO pin 1). The engine immediately sped up and began running smoothly at a high idle speed. The rich mixture was also confirmed by placing a cotton rag over the exhaust pipe during rough idle for 5 seconds and the hot exhaust gas (very hot from the extra gasoline delivered to the catalytic converters) ignited the rag.
I suspected bad gasoline may have poisoned the OX sensor, so drained and refilled tank with fresh gas, no change.
Verified the OX1 sensor is operating correctly by removing it from the 95 previa (where it always reads lean) and installing it in my 96 previa where it operates correctly (cycling rich – lean – rich – lean with the ECU controlling the fuel trim as it is supposed to).
Replaced the OX2 sensor for good measure, no change.

I can’t help but think that if I had an internal wiring diagram of the ECU that maybe I could find an errant supply or ground path related to the OX sensors.
Is there any condition that disables the ECU operation of the OX sensor drive circuits?
I am stumped! Any ideas on what to check next?
Avatar
Drocketman
January 24, 2012.



Rivermikerat,
I once had a head gasket blow direct between cylinder and outside in an old Rambler inline 6, and it made a loud snapping noise. No such noise here. Where could the leak be and how could I detect it?

All three ECU ground wires (EO1, EO2, EO3) are less than 0.1 ohm to vehicle ground. I checked the distributor sensor and crankshaft position sensor wiring and all good. The sensor coil resistances are 2070 ohm for the crankshaft sensor and 165 ohm for the distributor coil and are in spec. I will do a few more checks of voltages between the ECU ground wires EO1, EO2, EO3, E1 and the OX sensor ground wire to see if any bias exists with the engine running. I don’t expect to find any.

Other than that, I am out of ideas to check without a scope. I need a scope to check the injector timing and durations, the crankshaft position sensor signal, the distributor signal, and the igniter drives from the ECU.

I want to make sure I don’t tear down the engine unnecessarily.
I'm stumped and still open for suggestions.


Tiny
Drocketman
Jan 29, 2012.
My dad has an old Rambler Marlin. Heh. What I was talking about with the head gasket sucking air is if the gasket is blown it can suck extra, unmeasured air into the combustion chamber(s).

You're hunting a lean condition. You've checked just about everything there is to check. Except injector pulse width.

Have you checked compression? That will tell you if the head gasket is blown on a cylinder. It can also tell if you if a valve that isn't opening and closing correctly. You could also pull the valve cover and watch the valve action from the top of the head.


Rivermikerat
Jan 29, 2012.
I have the same year Toy Van-no SC and still kicking like new

Have you checked all the items I've given you earlier and rule it out as the problem?

Check engine compression and injector for those cylinders that are misfiring-

*** Coolant temp sensor reads ~186 F when engine is warm (Is this warm enough for closed loop fuel control?) Should be 195-205degs

Read the sensor's resistances hot and cold- The CTS controls the loop while in open mode when it switches the O2 sensors takes control of the fuel demand-


Rasmataz
Jan 29, 2012.
2 adjacent cylinder misfirng could be a blown headgasket


Rasmataz
Jan 29, 2012.
Oh, my dad's Rambler Marlin is a 65.


Rivermikerat
Jan 29, 2012.
Who knows it could be leaking thru the injector's seal. I remember that car rode on it before so as the Gremlin, Nice cars though


Rasmataz
Jan 29, 2012.
Rivermikerat: In college (circa 1974) I bought a 57 Rambler SW that was a high school student’s labor of love (see pics). 196 cu inches of inline power, ha. Called it the green hornet. Green metal flake paint, black and white faux leather interior, bullet muffler, 3 spd manual on the column with a 2 spd automatic overdrive, chrome slots in the back and chrome reverse in the front. A memorable car. The head gasket blew out on the firewall side of the rear cylinder and made the snapping noise.


Tiny
Drocketman
Jan 30, 2012.
I checked the Previa’s compression with a cold engine and all four cylinders measured even at176 psi peak. I bought this Previa at 47k miles and it now has 165k. I’ve run only Mobil-1 in it and the valve clearances have never gone out of spec and I checked them after this problem started.

Also for good measure, I bought another spark plug and replaced the black #4 plug in the picture above. No difference.

Watching it run with the valve cover off is not an option in the Previa. The engine is laid over more than 70 degrees and it would be a real mess with oil going everywhere. See the picture here of the valve cover seen thru the passenger door.

***Have you checked all the items I've given you earlier and rule it out as the problem?

Yes, and I have gone over some again.

I just ordered a PC based oscilloscope and will check the injector timing and durations, the crankshaft position sensor signal, the distributor signal, and the igniter drives from the ECU as soon as it gets here. That should take about a week and I will post images of the scope traces here.

***Coolant temp sensor reads ~186 F when engine is warm (Is this warm enough for closed loop fuel control?) Should be 195-205degs

I had checked the cold temp resistance of the coolant temp sensor and it is in spec there. However I did not remove it and it may be lagging 10-20 degrees below the actual warm engine temp due to deposits. I will pull it and test in boiling water to see if it is biased out of spec there.

***Read the sensor's resistances hot and cold- The CTS controls the loop while in open mode when it switches the O2 sensors takes control of the fuel demand-

Yes that is what I have seen, however there seems to be a third mode that keeps the engine safe when the ECU is first initialized, and my 95 SC is stuck in that mode. When the ECU initializes from power loss (battery disconnect), BOTH the fuel trims are set to zero. And they will not read off of zero and the ECU will not engage the SC clutch until some conditions are reached. I’m not sure what those conditions are but think I have read that the OX1 sensor must read rich at least once. I can see how my good 96 Previa SC behaves different from my “stuck” 95 Previa SC. Here are some snapshots of the engine data of the 96 Previa on a cold start.

Ignition switch on, before startup: Note that all system monitoring is complete, fuel control is open loop, coolant temp is 98.6 DegF, and the long term fuel trim is at a value of 3.1%.


Tiny
Drocketman
Jan 30, 2012.
Here is the plot of the fuel trims and the OX sensor voltages also with ignition on, before start.
The long term fuel trim in blue is holding constant at 3.1%.
The short term trim, OX1 voltage and OX2 voltage all are steady at zero.
(The vertical lines are ~2 seconds apart so the charts show the last


Tiny
Drocketman
Jan 30, 2012.
This snapshot was captured about 20 seconds after engine start.
The OX sensors have begun to warm up and are still lean, but rising. OX1 is at 0.18 volts.
The red and blue fuel trim lines are hard to see, but the long term trim (blue) is holding constant at 3.1% and the short term (red) is holding at 0.0%. This is where the 95 Previa gets stuck. Its OX sensor values start out the same as the 96 Previa, but go to a solid 0.00 volts as it warms up and starts missing. If I rev the engine on the 95 in the first minute or so after a cold start, I can sometimes get OX1 to show a little blip above zero. But once it is warm, there is no getting it off of 0.00.
(The vertical lines are ~2 seconds apart so the charts show the last 10 seconds of data.)


Tiny
Drocketman
Jan 30, 2012.
Page 2 of 4