Hello. I'm wondering if someone with a properly working '90 Civic DX could attach a wire to their O2 sensor and read it with a digital volt meter while driving around. ...And then report those findings here. I'd like to know what it reads at idle once fully warmed up (at least 15 minutes of driving), what it reads at part-throttle while cruising around, and while maintaining speed down fairly steep long hills, preferably at speeds between 50 and 60 MPH--that is, maintaining speed down a hill with the throttle almost completely closed for a length of time longer than 40 seconds. Here in Milwaukee I'd drive over the Hoan bridge to accomplish that.
I'd like to know details like whether it swings lean-rich under certain conditions, or maintains a steady reading, and if steady, what that reading is (average). Not as important, but I'd also like to know what it reads when the car is first warming up.
On flat land under most loads I get a rich-lean reading. Going downhill I get about .8 to .9 volts. If I completely close the throttle, however, it goes to 0, indicating that the injectors are closing properly.
The engine idles at about .8 to .9 volts, too. I actually think that it is idling slightly rich. It lopes when started cold, and when warm, if I attach a vacuum pump to the MAP sensor and increase the vacuum to above what would be found at idle, the engine speed increases slightly and smooths out, although it's already pretty smooth at idle. It's super smooth, though, when I increase vacuum at the MAP sensor. Of course there is a point at which this leans out the engine too much, and then it idles poorly. Seems to be running rich, though.
I seem to have improved things slightly. First I took the guts from the EX TPS that I had installed half-assed and installed the guts inside of the original DX sensor so that it could be bolted down properly. I then purchased some stainless metric hex-drive bolts and mounted the sensor securely It is adjusted for .49 volts with the throttle completely closed.
As a reminder to anyone who hasn't read any of my other threads, the CEL problem I have been having actually started with the old (original) TPS, so the new TPS is not the cause. It has actually improved things. I don't think that the TPS is actually the problem, but perhaps its adjustment, and its adjustment definitely affects other things.
I have the throttle stop screw adjusted 1 turn from closed, which seems to set the idle about right for the minimum setting, although I think that it was originally more like 2 to 4 turns out. If I set it like that now, though, it idles too fast. I could be wrong about the original turn count, because I didn't keep good track.
I then reset the computer and let the engine relearn idle after taking it for a drive. It has great power and idles well. The point at which it reads .8 to .9 volts while maintaining speed down a hill has been greatly narrowed, so it is less likely that I hit this area, and less likely that I throw a CEL because of it. Something must still be wrong, because if I hover right in that area just before the computer cuts out the fuel, the O2 sensor reading stays high, like it would for idle, and the CEL illuminates after 40 seconds of this.
I reset the ignition timing by shorting the plug under the glove compartment. However, I wasn't actually able to check timing, as I can't find the timing marks on the pulley. The engine speed didn't change when I shorted the plug, though, so I'm guessing that it isn't off by much.
One other thing I did was change the O2 sensor back to a Denso unit. I changed the original Denso unit to a Bosch a long time ago, not knowing that there would be a difference. However, Bosch sensors run a tenth of a volt lower than Denso sensors, which would make the engine run rich. I notice smoother running with the Denso sensor.
I noticed back with the Bosch sensor that the engine ran kind of rich at idle, and ran leaner and better if I increased the vacuum to the MAP sensor with a vacuum pump. I have yet to test this with the new Denso sensor. The engine is running fantastically, though. The MAP sensor checks out well enough. Voltage readings compare to vacuum readings, as based off of the chart for this sensor. The chart gives an idle level of 1 volt, though, and the MAP sensor runs at about .92 volts at idle. If the engine is still running slightly rich, and smooths out with an increase of MAP sensor vacuum, I wonder if there are resistors that are faulty within the computer. It seems to be running leaner and better, though, with the new Denso O2 sensor.
The coolant temp sensor is new. I noticed that the old one measured below what it should. However, resistance drops with increase in temperature, so a lower resistance sensor would actually lean out the engine, as the engine is leaned for hotter and less dense air, so I don't think that is the cause, unless the over-all resistance slope became inaccurate with age. I did notice a slight improvement in engine operation with a new sensor.
I have yet to replace the air intake temp sensor. It also measures a bit low, but again, a lower resistance would indicate a higher temperature, and the computer would lean out the mixture for this, as hotter air is thinner.
The car does still lope when first turned on. It runs too rich. I'm not sure if this is because of the old air temp sensor, or if something else is wrong. Whatever it is, it's guessing wrong for the cold open-loop mode. Once it warms up, though, and goes into closed loop, it runs well.
In any case, I have yet to re-test the mixture since replacing the O2 sensor, and the replacement may have been the cure, here. Perhaps with the new and proper O2 sensor in place, the computer might also relearn its guesstimation for cold open-loop running, and perhaps it will run leaner. It is starting and running slightly better now that I've been driving with the new O2 sensor, but still is a bit rich.
Seems that the owner of this car claimed that it actually used to get about 500 miles to a tank with freeway driving, which is far more than it gets now (37-42 mpg, or about 370-420 miles per 10 gallons). It would be interesting to see if mileage increases with the proper Denso sensor.
...So things have improved, but the problem is not completely gone (high O2 sensor reading at near closed-throttle). Something must still be out of adjustment, or there might be something wrong with the computer. The resistors all seem to be carbon film, which are typically very stable. Many of the capacitors are ceramic disc, which almost never fail. I tested all of the electrolytics.
Thankfully now that the main relay is repaired, I don't have to worry about flicking the key, as before sometimes I'd wind up on the side of the road. If and when the CEL comes on when I'm driving down a hill, I flick the key because once the CEL comes on, the computer ignores the O2 sensor, and fuel economy drops. If I flick the key off and back on, it resets the computer and resumes monitoring the O2 sensor. Once I drove around all day with the CEL on, and mileage was as low as about 28! That's the lowest this car has ever achieved! I drive it hard in the city, testing out the new TPS , and it gets 33!
Engine still runs slightly rich at idle. When I increase vacuum to MAP sensor from what it sees at idle, the engine speeds up and runs smoother than it does at idle, though at idle it is pretty smooth. Not sure if timing is also involved. I should look at that. I could be mistaking one thing for another. In many cars, though, timing will increase to speed up an engine at idle that is under extra load. The load is sensed by a decrease in manifold vacuum--the opposite of what I am doing. ...Still I should check it out.
All along I have also experienced another problem. That is that if the car is taken right onto the freeway after very recently being started, it doesn't like to come out of open loop. I think that this has been the case for years, as I remember it always throwing a CEL if taking it onto the freeway too soon. Now with my DVM attached to the O2 sensor, I can watch what is happening. When I start the car and drive it on city streets, soon it goes into closed-loop rich-lean mode. Then when I take it onto the freeway soon there-after, the O2 sensor reading climbs to the .80s-.90s, like it would at idle, and stays there regardless of throttle position, except sometimes when I close the throttle, the fuel shuts off. If I maintain steady speeds, this will continue until the CEL comes on for 'Code 1--oxygen content'. However, if once I am on the freeway I take my foot off of the throttle every 10 to 20 seconds or so, I can coax the computer into closed loop mode, and it stays there from that point on.
I have yet to replace the intake air sensor. I have one arriving tomorrow. Both the original TA and TW are faulty. They measure the same and measure a rather low and incorrect resistance for any given temperature. At room temperature they measure about 250 ohms, which is far below the chart specification, which I found here: http://honda-tech.com/showthread.php?t=1817082
The TW has been replaced, so when I replace the TA tomorrow, hopefully that will make a difference. I don't think that it will correct the throttle problems that I am having, though.
Considering what happens when I increase vacuum at the MAP sensor, I decided to check some of the resistors around where the signal is fed into the computer. They all checked out fine. I also checked MAP and TPS voltages at the computer. They were correct. I'm about ready to disconnect and go through every resistor in the ECU at this point. I'm not sure if this is how this model of car is supposed to behave or not.
One thing I did notice all along is that when the engine is cold, the voltage at the oxygen sensor is 0 and not .45, like more modern cars. I'm not sure if it should be this way or not. More modern cars have a very high resistance that feeds this 'bias' voltage to the circuit. It is overridden by the O2 sensor once it warms up. I tested the highest resistances I could find within the computer--some 1 Meg and 2.2 Meg units, and they all checked within tolerances.
I put the old coolant temp sensor into my freezer, which is set for about 0*F, and let it stay there overnight. I measured it this morning, and it measured only 300 ohms! Last night I also applied my soldering iron to it, and lowest it got was 50, so it is definitely out of range. The new one I put in measured something like over 2K at room temperature, so the old one is definitely faulty. I am guessing that the IAT or AT, or whatever it is called, also reads the same, since both the AT and the WT measured the same when the engine was cold. I have yet to remove the AT because the Philips head screws are rusted in place. I will probably wind up chewing it out with a wire cutter, and then soak the screws in PB Blaster.
Though no codes were thrown for these sensors, I am guessing that either the computer was completely ignoring them, or fuel trims, if this computer performs any, have been seriously modified. I only hope that with a new air sensor, the computer will properly recognize it again. I did note improvement with the new water temp sensor. Hopefully the air sensor has something to do with the computer not wanting to come out of open loop.
I wonder if with the air sensor being faulty, the computer is just 'guesstimating' on mixture at idle, which would explain why it runs slightly rich. Actually, when the car is thoroughly warm after driving for a bit, it runs perfectly, but after sitting for a few seconds, I can hear it switch to a richer mode, as the engine gets slightly rough. My guess is that the computer times out the O2 sensor at idle, and relies on the other sensors, since the O2 sensor tends to cool at idle and become inaccurate (even a new one). With the AT sensor being faulty, the computer probably has to guess pretty hard on idle mixture. What is amazing is that this car, with all of its faulty sensors (AT, WT, and TPS) has achieved fantastic mileage, with 33 beating it hard in the city, and typical 38, with a high of 42, with a mix of mostly freeway and some city driving.
Last night I put a 1.2K resistor in place of the air sensor just to see what would happen, but it made no difference as far as my original problems are concerned (O2 sensor reading at almost closed throttle, and not wanting to come out of open loop when first started). Hopefully a real properly functioning AT sensor will make the difference.
Well yesterday I put in a new air temp sensor. Made no difference.
Later on I attached a potentiometer between the signal and ground terminals of the MAP sensor, and attached the signal to computer wire to the wiper of the potentiometer. I adjusted the potentiometer so that the computer would see a voltage of .77 at idle, with the engine full warm, no accessories on except the cooling fans. This is what I adjusted the MAP sensor for with a vacuum pump earlier, when I noted that the engine ran a bit better with this voltage (which likely runs the mixture leaner). I drove the car with this potentiometer in place. It drove well, and the problem of high oxygen sensor reading at near closed throttle was significantly reduced to the point that it was very difficult to actually make it happen. However, the engine didn't seem to have as much full power. I didn't check to see what the potentiometer did to the no-vacuum reading of the MAP sensor. I'm sure that the adjustment has probably leaned out the engine at that point, too, which is likely undesirable.
I put the MAP sensor back to its factory connection arrangement.
Last night I decided as one last resort that I'd replace the O-rings on the injectors. The fuel pressure holds well, and will actually hold significantly overnight. There also is no indication of injector or O-ring leakage. With the throttle closed and the car coasting, the O2 sensor reading drops to 0. If I unplug the injectors, the engine dies, with no indication of receiving any fuel. However, I thought I'd replace the O-rings anyway. Made no difference
...So I'm at my wit's end with this. I really don't know where to look anymore. I also believe that the fuel pressure regulator is working properly, based on the readings I saw with a gauge connected. I forgot the specifics, but it's like from 30psi with full vacuum to 40 psi with no vacuum. Probably not exactly that, but it's in that range, so I doubt that the injectors are getting too much pressure.
Sounds like a computer issue, but I really don't know what. I read in my useless Chilton's that the 87-89 Civics have some sort of 'idle fuel mixture adjuster.' I didn't see any indication of one for the '90 model year, and I don't see one on my car where they say it's located. That manual is so useless, though. They give bits and pieces of information for different model years, and you have no idea if it applies to other model years or not. For instance, temp sensor readings are given for 87-89 model years, but not for any others. So what does that mean?????????? Who knows.
I guess I should just live with this problem and flick the key whenever the CEL comes on. It comes on quite a bit at speeds of 45 MPH. The throttle is held nearly closed on flat land at this speed, and unless I move the throttle around a bit, or encounter lots of hills, the CEL comes on.
Very good read. I too am having problems with my 02 sensor. Took it to honda an they diagnosed it as 02 so replaced with used one and still blinking code 1 blink with 2 seconds later another blink. Even so I noticed alot more power. So replaced with new Bosch sensor and noticed a power drop and same cel reading. I am going back the peppy used one. Not sure what to do about my cel problem especially when it was running fine before I replaced the PS unit and Catalytic converter. I know the old cat was clogged though. Thanks for the read....Hatch.