check engine light adventures

I've decided to start posting again.  I've recently had car troubles exacerbated by poor decision making.  I was driving home from a visit to my grandparents house recently and my check engine light came on.  Prior to embarking on the 4.5 hour trip from Akron to South Bend, I thought I noticed that my car was losing power and then surging back.  In addition, on the way up I thought I had poor fuel mileage, but that could have been a result of many things since the prior tank had very good mileage and it could have been a related error, or poor mileage could have resulted from the weather here being very cold and the fact that I only make short infrequent trips when not driving to and from Grandma's house.  When the check engine light came on I was on the turnpike and 14 miles from the next exit, also about 30 miles from my grandparents house.  I decided to exit and try and go to an autozone and have the DTC OBD II codes  read in order to determine if the problem was a serious one.  I asked for and received directions to an autoparts store from the toll booth operator.  However it turned out those directions were not so good and I spent a good deal of time driving around the country side.  I then thought my uncle has a code reader as well as tools and such so I ought to head back to Akron and diagnose the problems there as I could expect more help from my uncle than from the autozone clerks, assuming I ever found one. My uncle read the computer with his code reader and learned that I had a "fuel trim malfunction" and "system too lean" error.  However I did not record the actual codes.  He advised me that I should just put some fuel injector cleaner in and be on my way as he had a car that gave these errors when he didn't use the stuff.  I figured this explanation seemed reasonable because I drive so infrequently that my gas sits and can oxidize and vaporize and collect moisture and that my maintenance was so behind that I may have forgotten to swap fuel filters and plus my last tank was from sheetz, not necessarily the most reputable gasoline vendor perhaps.  Given these things, I took his advice.  However, the car ran rougher and rougher.  It wasn't so noticeable while going down the highway at speed, but when idling the car would nearly stall out as the engine rpm's dropped so low.  In addition I got atrocious mileage-27 rather than my customary 40+ mpg.  As such after the trip it was clear that I needed a little more than a gas treatment.  Thus the fun began.  I found a service manual online for a 1998 honda civic as I have.  It was 2000 pages and with poor indexing.  As such it wasn't that helpful as my computer's a bit slow anyway.  I called the honda dealer and they told me that they needed to know what the exact codes were in number form to help me out and that it couldn't be the fuel filter because my car has a lifetime filter built in to the gas tank.  I asked what that fuel filter like looking thing was on my firewall and they told me it was not a fuel filter.  Anyway, they suggested a whole list of problems which could cause such codes.  I then began googling and read all sorts of conflicting accounts.  After this great help I went back to autozone to have the codes read again.  They were P0170 and P0171.  The clerk at the store printed out tons of useful papers and discouraged me from buying any of the tools I'd eventually need to finish this job by telling me falsehoods.  I went home and did more googling.  I resolved to print the 2000 page document on the school's dime and go from there.  After a paper jamb and lots of reading I decided I wanted to test the Manifold Absolute Pressure gauge (MAP) and the oxygen sensor (HO2S) as both of these things could well be the culprit and both can be tested without loosening any bolts and such.  The easiest way to do this is with an OBD II scanner, so I called various autoparts stores to see if I could borrow or buy and return one.  All said no, so I called my machinist friend to see if I could borrow his.  He referred me to his out of work mechanic friend who talked to me for a long time, suggesting all sorts of things that could be wrong.  However he didn't have a scanner either and he didn't think one would be useful.  He did however offer to look at the car for me.  I disagreed and resolved to take advantage of Harbor Freight's easy return policy and purchased a "scanner" (sku 99722) there.  However after purchasing this scanner didn't have many of the features typically associated with a scanner, so it was actually a code reader with a higher price tag than their other code readers.  Nonetheless I used it to reset the codes in my car as the driving was becoming less and less smooth.  On the way home a new code became pending, P0133 (this code must appear in 2 consecutive trips to activate the check engine light and until then it is pending).  This code pertains directly to the HO2S and basically excluded other possible problems besides perhaps an exhaust manifold leak.  As such I undertook to make my own back probes, and connect them to an oscilloscope in order that I could monitor the signal from the HO2S directly.  This would have been cake with a proper scanner, but hey.  The sensor only read lean, so I tried to make it rich by blocking off the air intake.  This didn't seem to work, and that bothers me because it seems that the car should have stalled or something.  Perhaps the pressure in the intake just dropped a bit.  As such I had a good clue that the problem was was the HO2S, but I wanted to do a bench test in which one removes the sensor from the car and places a volt meter on the signal wires.  Then a propane torch is lit and directed at the business end of the sensor.  In a properly functioning sensor this produces a result on the voltmeter.  I did this test and my senor passed.  As such I was a bit miffed.  I did notice if one got the sensor too hot the response quit.  I believe this was the heart of the problem.  I put the sensor back in the car and tried to test again with the voltmeter.  Initially I read that a rich signal, but after a while I read a lean signal.  Perhaps this had to do with the car still warming up or some such, or perhaps the sensor finally got to the same point where it began failing the bench test.  Hard to say exactly what's wrong, but I just replaced it and now the whole car seems to work just fine.  I think I'm going to get a cheap scanner that one hooks up to a laptop and make some recordings of the various parameters available for diagnosing purposes so that I can perhaps more easily diagnose any problems next time.  I rather hope I learn from these mistakes, and you might too.  If you want to do auto repairs, get a scanner.  Lots of other lessons here too, hopefully I'll remember them.