Hard deceleration at 65mph when let off the gas. Transmission?

[bendecker]

I could use some expert advice here, please. I have searched extensively on symptom #1 and can't find a reason for it.

 

1. Intermittent problem: When driving at 65 mph, usually after the car has been driven 20+ miles. there is a really hard deceleration when I take my foot completely of the  pedal. The closest I can come to describing the feeling is like it shifted to a lower gear, although that is clearly not happening. It happens most often when it requires just a hint of gas to keep the  car at 65mph. Like when going down a hill that is not very steep. It can go days without doing it at all, or do it really bad on a single drive. Not too much in between, but it does happen.

 

2. The other symptom that may or may not be related: When starting from a dead stop, it feels like there is a slight delay in full engagement of the drive train. It starts moving, then there is kind a quick engagement of the rest of the drive train and I'm off and running. I can live with this just fine, but am curious as to what it might be.

 

[[ EDIT: I'm now think this is an engine stumble when accelerating from a stop, although it sure felt like a tranny problem. See later posts. ]]

 

Car otherwise shifts very smoothly and runs great. [[ EDIT: See later posts -- stumble and died on the freeway with no CEL ]]

 

STATS: 1996 Legacy wagon. 260K miles. Recently flushed the tranny and put in some trans-x. (Which helped #2 by about 50%)

 

Many thanks for your help.

Edited October 31, 2014 by bendecker

[Mr.Atlantis]

Perhaps something with the main driveshaft.

 

Check bolts holding shaft to rear diff.

Also check for play.

[Fairtax4me]

These cars will decelerate quite a bit when you take your foot off the throttle compared to other cars. The high final drive ratio combined with relatively high engine speed (~3,000 rpm) gives these a more pronounced engine braking effect than most other cars.

But the thing to pay attention to is engine speed. If the engine speed is around 3,000 that's normal for 4th gear. If engine speed is up around 4,500 rpm, that could mean the trans is staying in 3rd gear, and that will cause a much harder deceleration when you take your foot off.

 

The hesitation when accelerating could be due to the trans being stuck in 2nd or 3rd gear, then downshifting to 1st while trying to accelerate.

 

Does the AT TEMP light work? Does it blink 16 times when you start the engine?

[bendecker]

Perhaps something with the main driveshaft.

 

Check bolts holding shaft to rear diff.

Also check for play.

 

The rear main driveshaft (aka the propeller shaft, I believe) was replaced with a new one not too long ago, so I know that's not the problem. The car did make horrible clunking sounds on takeoff before that fix, so I would say that your recommendation is spot on as a first step if anyone else has those symptoms.

Edited October 27, 2014 by bendecker

[bendecker]

These cars will decelerate quite a bit when you take your foot off the throttle compared to other cars. The high final drive ratio combined with relatively high engine speed (~3,000 rpm) gives these a more pronounced engine braking effect than most other cars.

But the thing to pay attention to is engine speed. If the engine speed is around 3,000 that's normal for 4th gear. If engine speed is up around 4,500 rpm, that could mean the trans is staying in 3rd gear, and that will cause a much harder deceleration when you take your foot off.

 

The hesitation when accelerating could be due to the trans being stuck in 2nd or 3rd gear, then downshifting to 1st while trying to accelerate.

 

Does the AT TEMP light work? Does it blink 16 times when you start the engine?

 

I don't think it's deceleration based on the engine. I have a manual transmission '93 Impreza and when I take my foot off the gas in that car, the deceleration is very minimal compared to when this '96 is acting up. The change between normal and hard deceleration is very quick, if that helps.

 

I'll check on the RPMs and the AT light today. I don't ever remember seeing a light blink 16 times when it starts (we've had the car since 45k miles), but will check that today.

 

The deceleration is severe enough that my wife will no longer drive the car. Thanks for your reply. More info coming.

Edited October 27, 2014 by bendecker

[bendecker]

More info. There is an AT Temp light that comes on right when I start the car. It turns off a couple of seconds later. My wife said it has always done this since day 1 (until it started acting up, she drove it exclusively)

 

I tried to get it to do the deceleration thing, but no luck today. I purposely popped it into 3rd at speed and it kind of felt like the problem, but when I let up on the gas, it was much smoother and didn't decelerate as hard as the intermittent problem.

 

BUT!!! I was going up a fairly steep hill at about 55mph when all of a sudden, the  car had a dramatic loss of power. It kept getting worse and very quickly I had to pull off the road and it died. I tried to start it again. Nothing. Again, and it sputtered, then came painfully back to life. I got it turned around and headed back home asap. No CEL, but it threw a P0304 code (misfire cylinder 4) code. Given the severity of the symptoms, I think it threw that code because it didn't know what else to throw. 

 

It did this a couple of times, months ago, but has not since a complete tuneup.

 

BTW, it has all new plugs (NGK), wires (NGK), fuel filter (aftermarket from Autozone -- don't say it... I know, I know...) and got a new fuel pump about 15k ago (OEM)

 

This sucks....

Edited October 27, 2014 by bendecker

[1 Lucky Texan]

could be a fuel supply issue I guess. That might explain the slow start - pump has to pressure-up the rails.

 

I'd want to investigate possible exhaust blockage - a vacuum gauge test could help with that.

 

Dropping the exhaust to look for a dropped valve guide might be on the list of things to check, but after other/easier/common stuff is eliminated.

[bendecker]

could be a fuel supply issue I guess. That might explain the slow start - pump has to pressure-up the rails.

 

I'd want to investigate possible exhaust blockage - a vacuum gauge test could help with that.

 

Dropping the exhaust to look for a dropped valve guide might be on the list of things to check, but after other/easier/common stuff is eliminated.

 

It felt like a sudden fuel issue. I wonder if there is some event that can cause a sudden fuel stoppage.

 

As for a possible exhaust blockage, ya' gotta love the Internet -- a quick search pulled up this video from Eric The Car Guy checking the exhaust on a 2.2l Subaru:

 

 

Fixing this mess is going to be a one step at a time thing. I'll be doing the exhaust test in the next day or two.

Edited October 28, 2014 by bendecker

[Fairtax4me]

I think I may know what's going on here.

 

The older cars are starting to have problems with the MAF sensors. A solder joint cracks in the sensor and the sensor signal drops out intermittently. It usually doesn't set a CEL unless it happens 3-4 times in a drive. And even then it may not do it.

When the ECU loses the MAF signal it cuts fuel entirely or almost entirely and the engine will flat out stall. If the car is moving this will feel like you turned the key off. Hitting the throttle makes no change, except the more you open it the more the car slows because the engine is sucking in and compressing more air. That compression is not being turned into power because there is no fuel.

 

Everything will still work just the same. No lights (battery charge, oil pressure, ABS, or CEL) will turn on because as far as the car is concerned the engine is still spinning, everything is fine. Tach still reads normally, speedometer still reads normally except for the needles swinging down as engine speed gets lower. All the right signals are still going to the same places, just the engine isn't getting any fuel.

After a few seconds the loose connection is remade, the ECU delivers fuel again and the car drives on as normal.

 

See if you can pick up a used MAF sensor from a nearby junkyard (usually under $40) and pop it on and see if the problem goes away.

[bendecker]

Dang. The MAF sounds highly likely. In case I can't find a used one (Subarus are not very common in junk yards around here), are aftermarket MAFs any good? Rock Auto has a Dorman for $112.00, a BECK/ARNLEY for $132.00 and a rebuilt A-1 CARDONE Part for 137.00.

 

If they are OK, I can likely source one for less.

 

Thank you to everyone for the input. I really appreciate it.

Edited October 28, 2014 by bendecker

[1 Lucky Texan]

nvermd.

Edited November 1, 2014 by 1 Lucky Texan

[uniberp]

Are your sure your brakes are not dragging badly?

 

I could use some expert advice here, please. I have searched extensively on symptom #1 and can't find a reason for it.

 

1. Intermittent problem: When driving at 65 mph, usually after the car has been driven 20+ miles. there is a really hard deceleration when I take my foot completely of the  pedal. The closest I can come to describing the feeling is like it shifted to a lower gear, although that is clearly not happening. It happens most often when it requires just a hint of gas to keep the  car at 65mph. Like when going down a hill that is not very steep. It can go days without doing it at all, or do it really bad on a single drive. Not too much in between, but it does happen.

 

2. The other symptom that may or may not be related: When starting from a dead stop, it feels like there is a slight delay in full engagement of the drive train. It starts moving, then there is kind a quick engagement of the rest of the drive train and I'm off and running. I can live with this just fine, but am curious as to what it might be.

 

Car otherwise shifts very smoothly and runs great.

 

STATS: 1996 Legacy wagon. 260K miles. Recently flushed the tranny and put in some trans-x. (Which helped #2 by about 50%)

 

Many thanks for your help.

[bendecker]

Are your sure your brakes are not dragging badly?

 

Absolutely sure.

Edited October 28, 2014 by bendecker

[bendecker]

Turns out there are a ton of used oem MAF sensors on Ebay. I think I'll  buy one and throw it in.

[bendecker]

Did a couple of things:

 

1. Put in a new MAF. (1 Lucky Texan -- I unplugged the old one it and it ran worse. )

 

2. Did a vacuum check for a clogged exhaust. No obvious problems found.

 

BUT, during the exhaust check, I discovered that #2 seems like an engine stumble problem from a stop (although it really felt like a tranny engagement problem). When doing the test, I manually hit the throttle and the stumble was really obvious and 100% repeatable.

 

3. Sprayed all the vacuum lines to check for a leak. No evidence of one.

 

I have not take it for any long drive where I could feel the hard deceleration problem or sudden death problem.

 

Maybe the coil...? Fuel Injectors..?

 

I think my next test will be fuel pressure, although I'd be surprised if it is low because of the new fuel pump and filter.

 

[ For anyone just looking at this thread: all new plugs (NGK), wires (NGK), fuel filter (all about 500 miles ago) and an OEM fuel pump about 15k ago. ]

Edited November 2, 2014 by bendecker

[Fairtax4me]

You replaced the MAF first, then did the vacuum check, andthe acceleration issue is still present with the "new" MAF?

 

 

Unplugging the MAF is a good test to see if it is causing a stall condition. Typically if the engine is running rough, unplugging the MAF will only make it worse. The ECU defaults to a baseline AF mixture and ignores all AF related sensor inputs with the MAF unplugged. This means you get a very rough idle and low power, but the engine will continue to run. If the MAF had been causing a stall issue, unplugging it would at least allow you to get the car to a safe place, or a short distance back home.

 

Most of the time a rough running or rough idle problem caused by an oxygen sensor can be made better by unplugging the front O2 sensor. But unplugging the MAF sensor will generally cause the engine to run rougher, so it's only a conclusive test if you have a stalling issue.

[bendecker]

You replaced the MAF first, then did the vacuum check, and the acceleration issue is still present with the "new" MAF?

Yes. The acceleration problem is still there.

Unplugging the MAF is a good test to see if it is causing a stall condition. Typically if the engine is running rough, unplugging the MAF will only make it worse. The ECU defaults to a baseline AF mixture and ignores all AF related sensor inputs with the MAF unplugged. This means you get a very rough idle and low power, but the engine will continue to run. If the MAF had been causing a stall issue, unplugging it would at least allow you to get the car to a safe place, or a short distance back home.

I unplugged it while running. Oops... it died right away. Then, I unplugged it before starting and, yes, it ran worse.

Most of the time a rough running or rough idle problem caused by an oxygen sensor can be made better by unplugging the front O2 sensor. But unplugging the MAF sensor will generally cause the engine to run rougher, so it's only a conclusive test if you have a stalling issue.

Hmmmm it purrs when at idle and, generally runs fine at anything above an idle (except those few times where it died). It's the transition from idle to rolling that is the main symptom. If I manually push on the throttle gently and slowly from the engine compartment, it chugs at the transition, then seems to run fine. If I try to gun it, it is far worse and sounds like it would stall, but then it smooths out to normal.

 

Does that still point to a possible bad O2 sensor? The "upstream oxygen sensor" was replaced by my mechanic a couple of years ago when it failed smog. The reports says it was "acting erratically". (It did, however, pass smog fine about 2 months ago.)

Edited November 1, 2014 by bendecker

[Bushwick]

A misfire can be a several things:

 

First off, just because something was replaced a year ago, doesn't mean it can't be bad now. So don't rule out a spark plug wire or an actual spark plug. A plug wire, if it gets damaged on the exterior, CAN cause intermittent shorts as it grounds out. Easiest way to check is see if the casing is rubbing on any metal or has chaff marks, OR came into contact with fluids like battery acid or oil. A water bottle spraying water (helpful at night when it's dark) on the wires can reveal a short to ground as the current arcs. Do NOT touch the wires while it's running as the current can kill if it's shorting out. Very important to use the wire separator clips to isolate them

 

 

Also, check and make sure EACH boot is firmly connected to each plug AND each point on the coil. It's also possible the wire itself has worked out of the crimping, and thus the connection is poor (you are supposed to use dielectric grease inside each boot AND pull wires off at the boot and NEVER pull the wire as it can pull right out of the crimp). You can very carefully reopen the crimp, and reattach the plug wire if you know how.

 

 

Another possibility is the ignition coil. I know on older Fords, a car could run perfect, shut it off, and it'd never restart again. Saw that happen more than once. On other cars, you end up with a random misfire OR a specific cyl. misfire. Having a coil directly attached to the engine was never a good idea as it can heat soak and act differently when cold vs. hot. Thankfully the coils are all over junk yards. If no local yards, try car-parts.com and locate one closest to you. Most of the yards there should be willing to ship that to you. For $15-20, plus $5 for shipping, it's a worthwhile gamble if the coil is acting up after too much heat soak. Also check the wiring harness and plug that attach to the coil for corrosion, looseness, damaged or pinched wires, etc.

 

 

A cylinder misfire can also be an up or downstream O2 sensor. I had this happen on my Saab with it's downstream O2 sensor and it was PITA to diagnose as it never threw the correct code pointing to it. It did however throw a code for a failed pre-heater element, but nothing saying it was causing the misfire (it ran perfect for a year with the failed heater element code BEFORE ever getting the misfire) . Just that a random misfire occurred and sometimes it'd point to a specific cylinder. In my case, the actual metal casing surrounding the upper half of the O2 sensor, where the wiring pigtail entered the sensor, was cracking and the chrome was peeling off. It originally started out as a very random misfire, that would clear itself out after blipping the throttle, then wouldn't occur again for days. Near the end, it'd do it after the 1st 5 minutes of driving, then started doing it again while driving. Fortunately, the car was catless briefly when it started misfiring badly, I goosed it in neutral and it backfired so badly it literally blew the muffler apart at the seams. Surprisingly it didn't hurt the turbo, and had the cat been there it'd have been damaged. Finally went ahead and replaced both O2 sensors (replaced the rear last as upstream didn't stop the misfire) and the issue was fixed. Also, when the shorting occurred, the car would slow down a bit and loose power, but wouldn't completely stall. Putting it in neutral while coasting and it'd still idle OK. So basically similar to your issue. Idle was OK (towards the end) but it'd stumble under throttle or actually not react to the gas pedal. I honestly never thought a downstream O2 could cause that much havoc. Thankfully rockauto has Subaru O2 MUCH cheaper than Saab O2

 

You can try disconnecting the rear O2 sensor, drive it and see if the issue stops. Car should enter a limp mode with preset values.

Edited November 1, 2014 by Bushwick

[1 Lucky Texan]

TPS ?

[Bushwick]

Another thing that can cause a misfire is oil entering the spark plug boot. If the valve cover grommets are leaky, oil might be getting past them and pooling up in the plug boot, OR seeping down the threads (I believe it's possible for the plugs to back out if they were torqued down incorrectly or the head was the wrong temp at the time of the plug change). Both instances can cause misfiring, especially if the plug electrode gap is getting bridged with oil on occasion as it'll coat the electrode and the engine won't fire correctly. It'd be worth the effort to pull the plugs anyways and inspect for any type of fouling or incorrect colors. You did gap the NGK yourself, right? Plugs never come pre-gapped as they fit MANY other engines from different manufacturers.

 

Though to be honest, given how similar your symptoms were to mine, I wouldn't be surprised if it was the rear O2. Did you try unhooking the rear O2 to see if the issue clears up? If it were shorting internally like my O2 did, unplugging it will prevent the short. It'll throw a code and go into whatever limp mode Subaru has and should at least rev and be drive-able. You can also try and read the resistance of the O2 sensor across it's pins, but it may still read within spec when cold. If it has a hard short to ground, it'll be obvious with a voltmeter. You'll need to find the values though for it as I'm clueless where they should be. If by some chance Subaru uses the exact same up and downstream O2 sensors, you can take the numbers from the newer upstream and compare against downstream. go to rockauto.com, look up your car/year, etc. Find O2 sensors under emissions, and see if up/down have the same part number. 

Edited November 2, 2014 by Bushwick

[bendecker]

None of the wire/coil/plug issues you mentioned apply. They were all swapped out less than 500 miles ago and everything was done as pointed out. (Great tips, BTW)

 

I decided to check the coil today. Primary wiring appeared to be out of specs (should be .73 ohms, max, but was 1.8 ohms). But, down at those levels, I did not completely trust my multi-meter, so I tried it on another one -- 1.0 ohms. Then I checked the secondaries. Easy check and both were in specs, but after I put the boots back on and did a few things, I forgot exactly what the numbers were (doh! Getting old...). So, I pulled off the boots again for a quick check. Open circuit on both. What?!?!? Did everything I could to rule out poorly placed probes. Got a flash here and there on the meter, but nothing steady. So, I tried the other meter with longer probes. Same thing. Just for grins, I put the spark plug wires back on the coil and pulled them off at the plug ends and did a check through the wires (thinking the ends in the coil might push out the cups and force a connection). Same thing. Tried the test on my '93. Perfect.

 

So, a new (used) coil is on order through ebay -- $20.00 delivered.

 

I also jacked up the car and disconnected the rear O2 sensor. Seemed to be no noticeable change when gunning it in the driveway (although the car is currently running a bit better for no apparent reason). Didn't take it for a drive. Will do that test again with a full, long test drive if the new coil doesn't fix it.

Edited November 2, 2014 by bendecker

[Bushwick]

IF there was a short in the O2 sensor, it could be where the wiring enters through the epoxy. At some point, the rubber coating gets stripped back to enter a crimp that's encased in the epoxy (or whatever the O2 has). If the epoxy gets brittle and the exposed wire isn't sealed from the elements, it'll corrode over time, or worse get wet. O2 wires are extremely sensitive and moisture on them can mess with the signal, let alone whatever corrosion does to the signal.

 

I suspect if the coil doesn't fix it, a rear O2 might. Parts like these will wear out at some point, and often don't outright fail but gradually fail over time, increasing fuel consumption. Point being, it doesn't hurt having new ones. Mine runs OK, but at 180k?+ miles I'll be doing my O2 soon to avoid the issue altogether.

 

Also, did you check the O2 sensor's values across the pins? I suggest wiggling the wiring harness where it enters the O2 to see if the values change at all. Would also be a good idea to check the values cold, and again after the exhaust has been brought to normal temps. Whatever you do, do NOT cut the O2 wires or strip the coating back for test probes as they are too sensitive to resistance changes.

 

And one more thing, any time you disconnect a sensor or reconnect one, disconnect the battery neg cable to rest everything and clear any stored codes. I'm not sure how sensitive the Subaru stuff is, but if it's detecting a short (your misfire), and is already in a limp mode, disconnecting the MAF and then reconnecting a new one, then disconnecting the O2, might be making the ECM go bonkers. Resetting the neg cable will at least put every thing back to square one before you experiment. 

Edited November 2, 2014 by Bushwick

[bendecker]

Thanks for the tip on the battery neg side. I have just been using my scan tool to reset everything.

 

And, just checked on RockAuto and a new downstream O2 sensor by bosch is $55.79. For that price, throwing in a new one sounds like the way to go. I have no idea when it was last changed, if ever, and 260K on a part is more than a good run. Didn't check it across the pins. Will do that as time allows.

 

Thanks for all the help from everyone. Very appreciated.

Edited November 2, 2014 by bendecker

[Bushwick]

I'm not that sure how "smart" the older Legacy ecms were. I know some cars have to "relearn" certain things after a hard reset, which takes x amount of miles or several restarts, etc. to "learn" everything again. Resetting the codes with the scanner might be enough, but at least with a hard rest you know the ecm is at square one again, so if it did have certain things stored, they won't be around anymore, unless the fault is still present, which might take up to 50 miles to trip again. 

 

260k miles on what's probably oem O2 or possibly a replacement is well pass it's expected life span. Keeping basic sensors fresh along with filters, fluids, plugs, & wires will keep the car at it's best MPG and hopefully most reliable. It also helps avoid the needle in the haystack scenarios (a giant magnet is useful there ) so when a funky situation does arise, it's often a brittle wire or rotted vacuum line, etc. GL and update if/when the issue straightens out as it'll help others should they run into the same problem.

[bendecker]

Put in the "new" coil (used ebay special) and pulled off the battery terminal for a hard reset. There was a dramatic improvement in the hesitation at takeoff after that. Now there is just a hint of hesitation -- nothing like before.

 

At this level, I'm reminded of a video series by Erik The Car Guy on YouTube (great channel), where he kept trying to track down a hint of hesitation at startup. It ended up being a skipped tooth on the timing belt. I'll be doing a timing belt in the next few weeks (belt, tensioners, oem thermostat -- first time attempt on that project) and am curious to see if everything is lined up properly.

 

I have not taken it on a long drive to check for the hard deceleration or the freeway shutdown from hell. Frankly, I'm hesitant to drive with anyone else in the car until I have more confidence in it. So, will likely take a long drive in the next few days with just me in it.

 

If I were to make up a story about what's happened to date, it would be:

 

1) Hard clunk at startup and continued clunking/shudder at slow speeds: new rear propeller/driveshaft (fixed before this thread).

 

2) Very strong buck on startup that was so bad it felt like a transmission problem: coil and probably the MAF made it worse.

 

3) Complete shutdown on the freeway: I'm hoping this was the MAF, but need more testing.

 

I'll post more results so the next poor sap that has these symptoms can see what was done to get my '96 rolling again.

 

As always, thanks for the input.

Edited November 6, 2014 by bendecker