TheSamba.com :: View topic

Wed Feb 01, 2017 1:22 pm

Hi all,

My 85 has an EJ22 in for about 10 years with about 20,000 miles. It starts fine, but suddenly the idle is rough and bogs down when rev'ed. It blows white smoke maybe blue/grey...I can't decide). I can "drive" it but it has no power. The smoke it bad enough I feel bad for my neighbor and don't idle it too long.

Here's what I've done so far, please feel free to second guess me. That's why I'm here:
1) verified timing belt is aligned
2) checked vacuum: steady at 15 in hg during idle. can't see a vacuum leak
3) check cylinder compression.
4) listened to injectors clicking with stethoscope
5) removed individual spark wires (doesn't seem to get worse with each disconnected)
6) looked at spark plugs - slightly sooty and black
7) used laptop with B10 Select monitor to read O2 sensor, IAC, and other sensors. Nothing that I see and out of the ordinary
8 ) threw a O2 sensor code, but I don't see that causing this. it would just go open loop.
9) disconnected each injector and observed a drop in rev and idle correction.

I have new spark wires and MAF. It happened suddenly after it sat for a bit and I jumped it. I think the smoke is oil, but that doesn't fit my "fuel or spark" theory on the fault. My fluids are clean with no signs of coolant in the oil.

my next checks might be to swap ignition module..

Thoughts?

Thanks!

Wed Feb 01, 2017 2:08 pm

is the white smoke slightly sticky sweet in aroma?
hold your hand in the exhaust and take a bigg huff.. I know Cali frowns upon this, but could be a blown headgasket.

Wed Feb 01, 2017 2:31 pm

Quote: is the white smoke slightly sticky sweet in aroma? hold your hand in the exhaust and take a bigg huff.. I know Cali frowns upon this, but could be a blown headgasket.

I would say no. I think the compression test would show a blown head gasket, no? I'd describe the smoke as noxious smell....like when you pull up behind a car from the 1960's, ya know, before those Californians made all those damn rules. :lol:

Wed Feb 01, 2017 2:41 pm

Based on the plugs and that smell it sounds like the mixture is very rich.....

Wed Feb 01, 2017 2:42 pm

What were the compression numbers from your compression test?

The oxygen sensor CAN cause some of the problems you describe. Replace it if it is not almost new.

Engines that sit unused for a while often suffer from varnish formation on the valve stems, usually intake valves. This can cause the valves to stay stuck slightly open. As some would say, "Believe me - - this problem with varnish is yooge, yooge!"

Subaru engines that sit unused for a while are prone to getting sticky rings, especially oil control rings, but also the top rings. The very best product on the market to address this is a product called "Yamaha Ring Free" available at any Yamaha dealer. Add it as directed to fuel and engine oil. Here are some links:

https://www.westmarine.com/buy/yamalube--ring-free-fuel-additive--2280683

Here is a quote from an outboard motor forum in which a Yamaha technician describes his experience with Yamaha Ring Free:

February 18th, 2013, 10:41 AM

Re: Ring Free, does it work?

I've used Ring Free as a cleaner all the time for fuel systems but not in the normal dosage. We learned how to use it in the Yamaha schools back in the early 90's when they first came out with it. I've taken engines with weak cylinders that was not due to damaged cylinders and ran pure Ring Free thru them. Yep, that right. Pure Ring Free. You can only do this with oil injected engines but drop the fuel hose in the bottle, and as it gets drawn into the engine you have to start closing the choke to keep it running. After the bottle is empty, shut the engine off and let it sit overnight. The next day you have to spray fuel into the carb to get it running, but it will blow all kinds of black gunk out of the exhaust. I've taken badly carboned cylinders with low compression and brought them back up to normal range. I've also run engines on a half gallon of gas with a full bottle of Ring Free and will actually do the same. Now for a little known secret. Techron is the same stuff. Yamaha claims that Ring Free is a heavier concentration but I've yet to prove it. Back in the 90's, the Yamaha techs use to tell us to use Techron if we couldn't get our hands on Ring Free.
Remember once years ago I had a 200 Yamaha with a mid range surge that I couldn't solve. Rebuilt the carbs several time with no solution. That was when a Yamaha Tech at the school told me to try the pure Ring Free and let it sit overnight trick. I did it and it cured the problem. So its not just for carbon in the cylinders, it will keep deposits out of the fuel system. Also, Ring Free is the only non oil based additive on the market. All the others use kerosene or mineral spirits."

The quote above is from this link, a few posts down:

http://forums.iboats.com/forum/engine-repair-and-m...es-it-work

Wed Feb 01, 2017 2:54 pm

Quote: What were the compression numbers from your compression test?

The oxygen sensor CAN cause some of the problems you describe. Replace it if it is not almost new.

When I did the compression test, it suddenly wouldn't start. turns out I knocked off the start wire when hunting for vacuum leaks....when it rains, it pours with this thing.....anyway.
I had to manually crank the engine. I will redo it this weekend with the starter, but I got 75psi manually +/- 5% in all 4.

my O2 is newish. I can see the voltage in my laptop. it stays at ~.8 during idle and only drops when rev'ed.

I won't rule either of those out yet.

Wed Feb 01, 2017 4:09 pm

Lack of power, misfire (assuming) with grey white smoke is very suspicious for getting coolant in the combustion chamber (i.e. head gasket, cracked head / block). Compression test will not be accurate unless you have a fully charged battery with a good starter and the throttle plate wired wide open. Leakdown test would be even better if possible. Good luck and let us know what you find.

Wed Feb 01, 2017 7:02 pm

85Joker wrote: Quote: What were the compression numbers from your compression test?

The oxygen sensor CAN cause some of the problems you describe. Replace it if it is not almost new.

When I did the compression test, it suddenly wouldn't start. turns out I knocked off the start wire when hunting for vacuum leaks....when it rains, it pours with this thing.....anyway.
I had to manually crank the engine. I will redo it this weekend with the starter, but I got 75psi manually +/- 5% in all 4.

my O2 is newish. I can see the voltage in my laptop. it stays at ~.8 during idle and only drops when rev'ed.

I won't rule either of those out yet.

Your approach to the compression check yields unusable results. As noted in the post above, a compression check must follow a well-known procedure:

1. Engine should be warmed to operating temperature;
2. All spark plugs must be removed.
3. Throttle must be fully open.
4. Battery must be fully charged.
5. Install the tester to the spark plug hole for the cylinder being tested.
6. Crank the engine until the compression indicated on the gauge no longer rises. This will be from 2 to 5 seconds of cranking for each cylinder.
7. For safety, but not for accuracy of the test, I prefer to disable the fuel pump either at the relay or at the connector to the fuel pump or by disconnecting the connectors at the injectors. On a Subie, disconnecting the fuel pump is the easiest approach to avoid throwing a code.

After doing your compression test, you will want to see compression of no less than 165 psi if your engine is a Phase 2 engine and no less than 160-165 psi if the engine is a Phase 1 engine. My interest in the test result is to spot a sticking valve which would show up as an anomalous cylinder and lead you to follow up with a leak-down test. If the rings are just gummed up, a treatment with Yamaha Ring Free could bring your numbers right up to spec (171 to 206 psi for Phase 2)

The oxygen sensor reading at 0.8 volts indicates a very rich mixture. This could be due to the coolant temp sensor and/or its wiring. Check against the spec in the Factory Service Manual (FSM). Check the oxygen sensor wiring too, per the FSM.

To save money, don't install the new oxygen sensor until you are at a point where you are sure it is the only remaining problem. If the coolant temp sensor and/or its wiring are the problem, for example, the rich mixture will continue and may or will ruin the new oxygen sensor.

One other thought occurred to me. Old gasoline is very good at gumming up injectors, causing poor spray pattern, and/or uneven flow, and/or injectors remaining stuck open. Don't replace such injectors and instead, use a reputable injector cleaning service. Factory Subaru injectors, like most OEM injectors, are very high quality and cleaning them is preferable to buying new knock-off injectors.

Regarding "white smoke", an engine with a failed head gasket that allows coolant to enter the combustion chamber will produce a lot of white steam. There is typically a tell-tale anti-freeze smell as well. But a coolant leak like that will not usually, on its own, cause a significant loss in performance/power. This is why I was thinking that your engine has a serious oil control problem and possible loss of compression due to rings being stuck. The ring problem would dramatically reduce power and cause smoke from burnt oil.

Have patience with the diagnosis - - these are very tough little engines.

Wed Feb 01, 2017 7:28 pm

If everything is working correctly, the reading from the oxygen sensor should swing back and forth every second between .1 and .9 volts. Since yours is reading high all the time, it only confirms that the engine is running rich and for some reason the ECU is not able (or not trying) to lean it out. You might check to see if one, or more, of your injectors is leaking.

Fri Feb 03, 2017 8:28 pm

I had a chance to rerun to compression test as you all suggested:

1: 215 psi
2: 200 psi
3: 209 psi
4: 205 psi

I'm not sure if it's phase 1 or 2, but the Haynes book says spec is 156-185psi. I'm assuming the slightly higher readings are from the soot that is probably built up in the cylinders running rich. I didn't find Yamaha ring free yet, but I sprayed Sea-foam in each cylinder to let that sit. otherwise, the results look good to me.

I tested the injector resistance and all are at 11.2 Ohms within spec.

the smoke is not coolant, i think it's just running very rich, but white/blue smoke....so maybe oil? Something is off with the fuel air ratio and it effects all 4 cylinders.

My next step is to swap the temp sensor. I swapped the ignition module with no effect. I think it is a sensor given the suddenness of this.

Here is a 30 second video of the smoke and running.

https://www.youtube.com/watch?v=-ipfW4kc2wk

Thanks!!

Sat Feb 04, 2017 11:16 am

This is exactly what happens to mine if I overfill it with oil. Mines a 97 ej22. After fixing it takes a while to burn ty he oil out of the exhaust

Sat Feb 04, 2017 11:53 am

Does the smoking diminish when the engine warms up? I wonder if the ECU is not going into closed-loop mode--essentially, staying in rich-running start mode. Replacing the temp sensor is a worthy direction. They are cheap and easy to replace and it is vital to the system working correctly.

Sat Feb 04, 2017 12:34 pm

Quote: Does the smoking diminish when the engine warms up?

Everything stays constant from start to warming up and never changes. the smoke is always there. I'm going to check all my emission and timing sensors today. I would think the O2 and temp sensor would cause a performance problem once warmed up since I thought the ECU picks a open loop fuel map untill they warm up. I guess that is true for the o2 more than the temp. The laptop shows a reasonable temp from the ecu. I'm also Going to check my crank, knock and cam sensors.

I think my mechanicals are ok (rings, head gasket, valves). It's running like crap, but I still hope it a cheap sensor. More to follow.

Thanks!
Andy

Sat Feb 04, 2017 5:17 pm

85Joker wrote: I had a chance to rerun to compression test as you all suggested:

1: 215 psi
2: 200 psi
3: 209 psi
4: 205 psi

I'm not sure if it's phase 1 or 2, but the Haynes book says spec is 156-185psi. I'm assuming the slightly higher readings are from the soot that is probably built up in the cylinders running rich. I didn't find Yamaha ring free yet, but I sprayed Sea-foam in each cylinder to let that sit. otherwise, the results look good to me.

I tested the injector resistance and all are at 11.2 Ohms within spec.

the smoke is not coolant, i think it's just running very rich, but white/blue smoke....so maybe oil? Something is off with the fuel air ratio and it effects all 4 cylinders.

My next step is to swap the temp sensor. I swapped the ignition module with no effect. I think it is a sensor given the suddenness of this.

Here is a 30 second video of the smoke and running.

https://www.youtube.com/watch?v=-ipfW4kc2wk

Thanks!!

Having looked at your video, I strongly suspect that the problem is a failed or failing Throttle Position Sensor (TPS) and/or the wiring to the TPS. The Subaru engines are very sensitive to the TPS signal. Remember that the TPS signal is used by the ECU to provide temporary enrichment for transitional throttle response. This is why your engine is stumbling when you open the throttle. The TPS signal is also used by the ECU to allow full-throttle (or near full throttle) enrichment. In the Subaru application, the TPS signal also tells the ECU how quickly the throttle was opened so that the ECU can determine just how much transitional enrichment should be added. When the Subie turbo engines have the ECU re-flashed with performance software, for example, this aspect is used to provide a lot of transitional enrichment - - a lot more than the EPA would like.

Anyway, the Subie TPS units do not last forever the way the Toyota ones do, by comparison. They develop "drop outs" and the odd signal can cause a DTC, but a failed TPS does not always cause a DTC to register. The failure mode, AKA "limp home mode" is for the ECU to generate a proxy TPS signal that is constant. This means that once the ECU detects a problem and goes into limp mode for the TPS, you will have no transitional throttle enrichment.

Your compression numbers are excellent and indicate that the rings and valves are sealing properly. The exhaust in your video does not look like the tell-tale steam-generator exhaust you would see with a serious head gasket failure. The compression is not "too high" and if it appears too high, it is not caused by carbon build up. It's probably a compression gauge that is a bit off. Remember that just 5% inaccuracy is 10 psi.

The Coolant Temp Sensor specs, for your checks, are 20,000 to 30,000 ohms at 68F temp. and 400 ohms at 176F temp. If those specs aren't met, replace it too.

What year is your engine?

Mon Feb 06, 2017 12:36 pm

OP update #3:

Ok.....

So I removed the air filter box and about 1 quart of water poured out of it. I have a K&N filter that was soaked with water. Fortunately, given the angle of the intake, water didn't drain into the engine but appeared to restrict the air flow. I've had setup a while, but have only recently been parking in my new driveway which sits at a 10 deg angle....all that recent California rains found new ways in..? Add in the fact that I cut the "J hook" off the air box to attach my version of a fresh air intake to the VW "snorkel" on the driver side. The water got into the joint from the air box to the flex hose. oops.

So I dry it all out and re-oil the K&N filter. And thought I had the "smoking gun" (see what I did there) to explain my rich mix. I was praying the van would purr like a kitten again....nope. Same issue. I reset the ECU to clear the fuel map...no joy.

To add to my drama, the passenger seat battery box was filled with water. I don't keep my battery there, but now I suspect water in the fuel

Time to drop and drain the gas tank. I thought about just using a product like Heet that claims to absorb the gas, but I think I should do this the right way.

Other than that, no ECU codes, all sensors nominal and reading normal on the laptop scan tool (TPS, temp sense, IAC, O2 reading rich, no ping)

I hope against hope that fresh fuel makes this all go way.

-Andy

Mon Feb 06, 2017 12:40 pm

Quote: What year is your engine?

It was a JDM crate motor that was marked a 1991 auto.

Mon Feb 06, 2017 1:10 pm

When you do the work on the gas tank also remove the expansion tanks that reside in the drivers and passengers front wheel wells. There is a valve on the top on those that has a grommet that goes bad and lets water splashed up there into the expansion tank where it nicely drains to the gas tank.

Mon Feb 06, 2017 7:43 pm

Water is heavier than gas and will pool at the bottom of the tank. The engine will stall quickly if one of those globs of water gets pulled in.

I wonder if the water in your air box has mucked-up your air flow sensor. It is very sensitive, and I can see how extra moisture combined with oil from your filter, might have coated it with gunk. You can clean the thin sensing wires with a spray of electrical contact cleaner. Be sure to use the kind that says: leaves no residue.

Mon Feb 06, 2017 8:44 pm

I reckon that the water in the air filter media and the wicking action of that media absorbing the water in the filter enclosure explains the "white smoke" that I would describe as "steam". Gasoline engines can actually ingest a fairly large amount of water and still run - - that's how anti-detonation water injection works.

The most likely culprit for the odd running remains the TPS, in my opinion. They cost between $250 to $300 new, so if you can borrow one known good to test in your engine, that would be best.

You can test your existing TPS using an analog VOM to check for correct operation of the existing TPS. Check for drop-outs in resistance while slowly opening the throttle. This test is done with the engine NOT running.

Do you have a three-wire or a four-wire TPS on your engine? If you have a 3-wire TPS, then the measured resistance at closed throttle upon the first couple of degrees of movement is very important since the ECU uses the resting resistance of the TPS as a proxy for closed throttle. The older 4-wire units have an actual closed throttle switch in them. Is there anything in your throttle linkage that might be holding the throttle very slightly open?

Wed Feb 15, 2017 8:25 pm

FIXED!!!! (almost)

It's the MAF.

The water must have gotten to it. seems pretty obvious now in hindsight. The curious thing is that it didn't throw a CEL code for the MAF. I wasn't sure how to diagnose a bad MAF and it's pricey to replace without knowing. But I ended up simply unplugging it and started the car and the computer guessed the air flow and ran normally. when I ran with the MAF....it runs terribly (rich).

I think you can use the same method to diagnose a coolant temp sensor (unplug it and see if it run better)

I should get my new MAF Friday so I'll confirm again.

Thanks to everyone that helped.

Andy