legacycentral bbs

when I got the car it was doing arround 15l/100km....I think that's arround 15mpg. You all know that sucks. I checked the O2 sensor...burned. I changed it but the car is still doing 15mpg,highway mileg mostly. I reseted the ECU. Twice.
I have two problems that I know of
1. I meassured the voltage on the signal wire of O2. It goes from arround 200 mV...to 500 steadilly (stabile). But at 550 it starts growing and droping kind of randomly, goes even -72mV....and it keeps growing to 7-800 mV, but all the time varring, 120-270-380-420...all the way up...drops again...so it fluctuates. And it does not become stabile again. Should I wait longer? I had the engine on for 10 minutes probabilly.
2. On warm engine, the CEL comes on. It's code 22...knock sensor. They say to check for continuity on knock sensor...and replace it if there is connectivity. But there's not. And the CEL comes on only if I stop the car but with the engine ON. So basically if the engine goes hotter then ...but the temp gauge shows normal...exactly half way.
So...what's happening?
Anyone from Canada? Where can I buy knock sensor cheaper than dealer? 107 cad for it
Help please:(

The voltage at the oxygen sensor will fluctuate between like 0.2 and like 0.8 volts when the engine's idling warmed up. That's normal. -72 millivolts is a little low but it's probably fine.

If the ECU doesn't trust the knock sensor it'll retard back to base timing, which results in reduced efficiency.

Take an ohmmeter and measure the resistance between the knock sensor wire and ground. It should be about 560K. But it's also possible you have an intermittent failure.

Do you have the original knock sensor? The original design has a gray electrical connector. They redesigned it to be sturdier and the redesigned one has a white connector. It's part number 22060AA031 and lists in the US for around 50 dollars, so your dealer's probably overcharging for it. You might try other dealers or maybe contact an online dealer in the USA.

So is it normal for the tester to jump from 200 to 700 in steps of 100?

Steps of 100? What do you mean exactly?

This page shows what some automotive waveforms look like:

http://www.katho.be/vhti/automechanica/fluke97.htm

Look at the one where it says "good oxygen sensor."

So, let's say in a 3 seconds interval...the tester will show: 120 for a split, 260...for a bit, and so on up to 700 and some...then drop...sudden or in steps as I described above....And the 3 second interval it's done:)

BUT...when I start the engine it shows..200 let's say. Wright? The first 2-3 minutes...the tester kept showing growing numbers 224-225-227...etc but in a constantly growing manner, up to arround 550...and then it started to do what I said in the beggining.

Hmm, that seems a little slow... a new sensor ought to cycle a couple times a second. But it may take some time before that happens if the ECU hasn't learned all about the new sensor yet. How long have you been driving with this sensor since you installed it?

When the engine first starts it takes time for the sensor to heat up, so you should expect low readings at first.

Oooooh...so it is normal that the tester displays...jumping results and not stady numbers. Wright?

Right. Take a look at this page:

http://www.smp-training.com/Counterman/ ... sor03.html

Basically, the ECU tries to keep roughly stoichiometric mixtures, and does this by bouncing on either side of stoich.

I donno...I am gonna sleep on it for a while...meanwhile I'll try to find a cheeper knock sensor.
Thx for help man.

Cool. You're welcome.

I bought the knock sensor from http://www.subaruwrxparts.com/ for somewhere around $35 US.

I bet the volt meter you have is too slow of sample rate and display rate. The voltages are probably jumping quite a bit and the interpolation of the meter is giving you incorrect results.

If the meter were integrating you'd expect him to see smaller fluctuations, no? Although I guess if it's a really cheap meter it could be doing who-knows-what with the signal... good call, ciper.