Sensor readings stuck at ZERO

[aschwerin]

1999 Subaru Legacy 2.2L manual

My wife was driving it for a few days with Check Engine light on. I pulled a P0170 (Bank 1 Fuel Trim Malfunction), P0325 (Knock Sensor), and an Evaporative Emissions code.

I cleared codes. Car now has very little power. Using Torque, I looked at readings, O2 1 and O2 2 are both reading flat zero (or near zero, but flat). All the Fuel Trims are also pegged at zero. I am getting other readings, revs, and such, and MAF looks good, in the lower single digits, going up to lower teens with some acceleration.

I checked engine (including o2 sensors) for bad wires and vacuum hoses, all looked ok.

[GeneralDisorder]

Fuel trims ought to be negative with a zero mV reading from the O2. Are you sure it was up to operating temp and in closed loop when you checked it out?

GD

[1 Lucky Texan]

also, would be no surprise a '99 needs a new knock sensor.

and ground connections refreshed.

[aschwerin]

It was a warm engine, my wife had been driving it a bit in the morning - I took a short drive, and I did see at one point it was up to 188 deg. Can't confirm closed loop.

Just to compare, I started a '96 Subaru, and while cold the o2's and the fuel trim had values, and the fuel trim would move significantly when I revved it. Unfortunately the '96 needs a full front end job so I'd prefer to get this '99 available.

It very well could use a new knock sensor, but that wouldn't affect the other readings? Or would excessive knock cause the engine to go into limp mode and that might zero out values?

I believe after I erased codes and drove it, when I checked again only the knock code was on again. If I remember right, that's an easily replaceable part that I could exchange tomorrow.

[GeneralDisorder]

Could be a limp mode. Hard to say if it would zero the fuel trims. I wouldn't think so but the programming on a lot of those older ECU's wasn't that great so it's entirely possible, Can't say I've noticed the fuel trims when inspecting a knock sensor code. 

GD

[aschwerin]

I pulled the knock sensor off my parts car but it was cracked so I didn't bother using it.

Then I thought, if the computer is in limp mode because of the knock sensor, it should be out of limp mode if I reset it.

So I left the battery disconnected for half an hour.

And I drove it, and I got readings. Front o2 sensor was going up and down as I understand it should. My STFT was mostly stuck up in the 20's. Sporadically, at the beginning of my drive, I would lose power with constant acceleration (uphill), just for a second - coinciding with a sharp drop in STFT.

Other than that, mostly normal for awhile... when I went up a long hill, o2 was pegged at/near zero. Maybe normal? Then down the same long hill, also zero, but more oddly, the STFT wouldn't go lower than -3% (its a manual, in gear).

Then, after a stop, in first gear it had very little power (around 1500 rpm) and o2 and fuel trim readings weren't moving. I pumped the gas and then it went back to being normal (or like 80% normal).

After all the driving, there were no fault codes in the ECU.

Cleaning ground connections still sounds like a good idea, other than that I'm not sure where to go.

/Andrew

[1 Lucky Texan]

some of that seems like a vacuum leak but, GD likely has more to go on now so, maybe he will reply.

the fact that it was better after the reset does somewhat point to a sensor though......?

[GeneralDisorder]

Fuel trims in the positive 20 range would indicate a possible vacuum leak. 

Also make absolutely SURE that the air box is properly snapped together - both the lower and the upper if you have them, and that the air filter(s) aren't creating weird flow through the MAF, etc by collapsing if they are the cheap FRAM style replacements. Factory filters have rigid frames for a reason.

Check that the vacuum line routing - especially to the MAP sensor and the MAP sensor switching solenoid, is correct as specified on the emissions sticker. 

GD

 

Edited November 8, 2020 by GeneralDisorder

[aschwerin]

I checked over the air box, and cleaned the filter, that looked to be a higher grade than cheap FRAM. I also was running the car and had left the airbox open, and was still getting the same running.

I quadruple checked the vacuum lines. All the routing was correct per the emissions sticker. One line said "to canister", but I didn't follow it past where it turns into metal line going under the intake manifold. The vacuum lines looked like they were from 1999, so I sprayed starting fluid along all them and never got a correlating rise in RPM.

So it was idling ok, STFT was still in the mid-20's, every 20 seconds or so the RPM would bump up a little then come back down. o2's were going up and down. And then at some point the idle got rough and the sensors had all flatlined again.

ECU going bad? I've never had that happen before, but I've heard it does. Also, I think this is unrelated, but the speedo only works intermittently, from my research there is a bad ground on the circuit board in the dash, but I wouldn't think that circuit board would have anything to do with the main ECU?

/A

[aschwerin]

I cleaned 7 grounds, no difference. 3 new codes, 302, 303, 304 - misfires on cylinders 2, 3, and 4. At this point I'm thinking the codes are indirect symptoms, not causes...?

[GeneralDisorder]

What's the MAF grams per second at idle?

GD

[aschwerin]

Cold start about 7.5 g/s, falling to around 3 g/s, then for a little bit sensors flatlined (around thermostat open) and it was up in the 4's. Functioning warm idle about 2.0 g/s. Revs around 700. Vacuum 22. STFT 30-32%.

[GeneralDisorder]

Warm idle should be closer to 3 grams. That's probably the source of the positive fuel trim. 

GD

[aschwerin]

I bought Walker brand MAF sensor off rockauto for $36. MAF now reads 3.6 to 4.9 g/s at warm idle. Fuel trim movement looks normal. Car runs great. Thanks for the diagnosis.

I've replaced 3 MAFs on Subarus but I don't think I've ever had to replace a different kind of sensor. I've always used the cleaner before replacing, and it has never helped. I guess these guys just have a short life span (10-20 years)

[GeneralDisorder]

Cleaning them requires using brake cleaner and scraping the elements with a pocket screwdriver. Spraying the weak sauce "maf" cleaner will typically do nothing at all. There is a layer of dirt on them and it needs to be mechanically removed. 

I have about an 80% success rate with cleaning the Denso MAF's. 

GD