'99 Forester S: fuel economy nosedive, unsure of cause

[casm]

Posted this originally on subaruforester.org, but didn't get any bites so am asking over here.  Any ideas gratefully received.

 

I've been measuring a big drop in fuel economy over the past two months - 19mpg on average down to (as I filled up this evening) 15.1mpg. I'm at a loss to explain what might be causing it. Before I get into the suspicion that I have, however, here's what has remained constant:

- 1999 Forester S, 4EAT, N/A. 151,000 miles on the body; 15,000 miles on fully-rebuilt SOHC EJ25 (heads and block). Compression and timing are good.

- Drop in economy is noticeable in daily commuting or on longer trips at constant speed. Total loss in fuel economy is approximately 20% across the board.

- Oil is changed every 3,000 miles; level is good and oil is clean. No leaks or burning.

- Tyres are brand-new (less than 800 miles on them). Economy drop was measured before changing tyres.

- Brand of fuel and station it is purchased at have varied, so it's not a case of one station having poor fuel storage leading to contamination and poor economy.

- Front oxygen sensor was replaced after the rebuilt engine was swapped in. Rear sensor has not been touched.

- Air filter is clean, spark plugs have normal brownish-grey residue, and the fuel filter was also replaced after the engine swap. I believe that the original injectors were reused with the current engine, though I'd have to check some fairly buried paperwork to confirm.

- Engine gets up to temperature normally, and there are no signs (whistles, surging idle, etc.) of possible vacuum leaks.

- There are no smells of fuel leaks, flooding, etc.

- A/C is not a factor as it hasn't been turned on since somewhere in the October timeframe.

There are two of us using the car on a regular basis, we're both seeing the same drop in terms of percentage, and driving styles haven't changed. However, for reasons that are completely inexplicable at this point, we've both seen the fuel economy go through the floor over the past eight to ten weeks. Frankly, the Jeep Cherokee that this Forester replaced managed 17-18mpg in mixed driving, and that was with a 4.0; in the nearly two years I've owned this vehicle, it's never been anywhere near this bad (with the exception of when the previous engine exploded).

One thing I will mention: the catalytic converter was not replaced after the engine replacement, and the old engine (which had failed rings and burned valves) was definitely dumping raw fuel and oil down the exhaust. I would not be surprised if the cat was burned out internally, but there are no rattles from it or signs of external damage. Ditto no exhaust leaks, either from the pipes or gaskets. I'm not entirely certain that I can blame the cat at this stage, though given that it's probably the original one it's likely time to replace it anyway.

 

That said, does anyone have any ideas on what to check? The vehicle is otherwise behaving perfectly normally apart from the huge drop in fuel economy, and I'm at a loss as to where to start probing since all the usual culprits seem to be in working order. Seeing a drop this sharp and this sudden with no obvious indicators of what may be causing it has me completely stumped.

[Gloyale]

Wheel bearings?  Brakes?

 

Rule those out at least.

 

 

Could be a knock sensor?

[1 Lucky Texan]

yeah, knock sensor or Engine Temp Sensor , less likely, stuck open thermostat - but you didn't mention low temp gauge reading.

 

if you have a scanner that can read freezefram data, you could try to force a CEL and read the FF dtat and look at the long term fuel trims.

[casm]

Wheel bearings?  Brakes?

 

Rule those out at least.

 

Good points, particularly since I didn't mention them.  Front rotors are new (within 6,000 miles); pads were replaced all around within the past year.  There's nothing to suggest a sticking brake (no loss of braking or pedal feel; no burning smells), and none of the usual audible wheel bearing symptoms are present.  That said, I haven't jacked it up and done the wobble test in quite a while to see if there's a possibility of them going out.

 

Could be a knock sensor?

 

Hadn't thought of that - is there a good test for them? No codes are being thrown and there's no CEL, though that doesn't mean that there isn't something out of spec that the ECU isn't sensing properly.

Edited March 24, 2013 by casm

[casm]

yeah, knock sensor or Engine Temp Sensor , less likely, stuck open thermostat - but you didn't mention low temp gauge reading.

 

Correct - all gauges read normally, and the car runs to temperature as expected.

 

if you have a scanner that can read freezefram data, you could try to force a CEL and read the FF dtat and look at the long term fuel trims.

 

Good point.  I'll have to see if mine can do that; it's something I've never tried to do with it before.  FWIW, it's one of the cheap Chinese ELM327 Bluetooth readers that I've been running with the Torque app under Android for a couple of years.  Worked great so far but I've never tried to do that with it so have no idea if it's doable or not.

[casm]

Just had a quick eyeball of the knock sensor.  Didn't remove it, but I learned two things: first, I have the later-type sensor without the pigtail attached to it, and, second, from what I could see it appears to be cracked along the top from the centre going out to the electrical connector.  I'll pull it out later and verify, but, figuring that it was likely swapped over from the old engine and it looks to be factory-original, there's 151,000 miles on it so replacement makes sense.  I've ordered a new one and it should be here in the next couple of days; I'll update once I've had a chance to swap it in.  It looks to be a no-brainer in terms of what's required to actually do the job, so as soon as it hits my hands it'll be in.

 

Also took a shot at figuring out how to access the coolant temperature sensor.  It appears as though the only two options are either pulling the entire intake manifold (best access, most effort) or removing the alternator (harder access, less effort).  Any recommendations on how to approach this?  I'm not ordering it until I get a chance to pull some OBD-II data and figure out if what the gauge is showing more or less matches what the sensor is reading since it looks like one of those jobs I'd rather avoid until I just can't ignore it any longer.  I've also got a fairly long trip coming up this weekend, and with work requirements having the time to take care of all of this is pretty limited.

Edited March 24, 2013 by casm

[hohieu]

You haven't mentioned longer term fuel economy trends, but it may just be a difference between winter (gas) and summer fuel mileage.  A 20% reduction during winter would be typical.  Oil is more viscous at lower temperatures and takes longer to reach operating temperature.  Also, ethanol has a lower energy content than gasoline.

[1 Lucky Texan]

I've read that the knock sensor can be delicate so, try to find the torque value for installation.

[casm]

You haven't mentioned longer term fuel economy trends, but it may just be a difference between winter (gas) and summer fuel mileage.  A 20% reduction during winter would be typical.  Oil is more viscous at lower temperatures and takes longer to reach operating temperature.  Also, ethanol has a lower energy content than gasoline.

 

I thought of this, but we're only just going back to Summer gas here so pretty much the entire drop has been measured on Winter gas. I'm used to the annual 20% drop once the blend changes; I swear at it, but can't blame the car for going from 23mpg on average back down to around 19mpg.

 

Also, re: Ethanol: there's nowhere I can conveniently get ethanol-free gas around here, so all numbers are on E10.  No E15 in this area yet (thankfully).

Edited March 24, 2013 by casm