About 6 weeks ago I had my dealer install a set of Stage I Cobb Tuning heads and Street cams. Due to some special circumstances, Cobb Tuning took the time to set up the valvetrain for me, so the dealer techs wouldn't have to worry about it. (On DOHC cars the Cobb cams usually require custom shims to get the valve lash set properly, and I wasn't going to have time to get those machined, so Cobb took care of it).

For the first week and a half or so the car felt fine. Then I started to get a very pronounced hesitation from about 2000rpm to 3800-4000rpm. The hesitation sometimes was pretty subtle, and other times the car actually would buck (much like when someone is learning to drive a manual, except it just bucks once instead of repeatedly). It did it regardless of how hard I was accelerating - gentle, normal driving or WOT. Once the engine hit about 4000-4200rpm it would take off. In 2nd gear it really felt like an on/off switch, the car felt kinda flat and then would literally kick me in the ass at around 4000rpm and pull strong to the redline. In 3rd gear I could still feel the change at around 4k rpm, but it wasn't as pronounced (I'm sure because of the higher gearing). In 4th and 5th I couldn't really feel the kick at all, but I could hear the engine note change around 4000rpm. I could also feel it on the highway, going uphill for instance the car would just feel kinda dead. I tried resetting my ECU, but it didn't help at all.

After a couple weeks (? - I wasn't tracking the time very well) the problem went away for a few days. I don't think I changed anything, so I'm not sure why the problem disappeared. After those few days, though, it came right back for a week or so, then seemed to disappear during my trip out West for the Open Track Challenge. When my co-driver returned the car to me 10 days ago, the problem was back yet again. The only difference now is that it doesn't hesitate as long as it used to - now the car takes off like a bat outta hell around 3400-3500rpm.

I can come up with a variety of ideas that would explain why the car would hesitate - bad fuel filter, misfire, bad fuel pump, bad coil pack, etc etc - but none of the reasons I come up with explain why the hesitation would go away so suddenly. From the dyno data, I know that at WOT with my current S-AFC settings the AFR is in the 12.75-13.0:1 range throughout the RPM band. So even at its leanest, the engine should be getting enough fuel so that there shouldn't be any detonation, especially with the 93 octane I'm running now.

The only thing I can figure is that the knock sensor is hearing something that it perceives as detonation up until ~3500rpm and as a result it's just dumping the timing. Since this started after the Cobb stuff was installed, I would tend to think it's something to do with that. Perhaps the shims aren't quite right, and as a result some sort of valvetrain noise is triggering the knock sensor?

Is there any way I can test the knock sensor, or test drive the car without the knock sensor? Hell, I don't even know where the thing is, so let's start with that. :) Would one of those Palm-based OBD II scanners allow me to see what timing is doing on a near real-time basis? Any other suggestions on what this could be?

Thanks,
Pat Olsen
'97 Legacy 2.5GT sedan

My Palm based PocketLOGGER reads and records timing and has been well worth the money. This also sounds a little like the chassis ground issue that many have reported. PS Have been enjoying the details of your trip. Hope this helps.

Originally posted by NoRice4U
This also sounds a little like the chassis ground issue that many have reported.
I added a couple extra ground wires a while back, but it didn't seem to do a damn thing.

Pat

My car has been doing the same thing, minus the going away for little stints. The dealership told me I have an oil leak at the crank case that has been leaking onto the timing belts and that is causing the issue. Probably not your situation, but maybe something to check.

Jared

Originally posted by Patrick Olsen

I added a couple extra ground wires a while back, but it didn't seem to do a damn thing.

Pat

When I read about this mod, I decided to test all my grounds...there is little to no resistance running from my battery to ground. :D

See if you can get your hands on a knocklink and a scantool to get the upstream data from the ECU.

mmm Your symptoms sound just like a blown MAF?? But you have a MY97 so maybee not. I doctored my Knock sensor to get rid of the notorious hesitation, I just used a gasket for an oil drain plug worked like a charm, and the sensor still works :)

You might also want to check your 02 sensor.

gasket from an oil drain plug? could you elaborate on this, please? :)

thanks!

peter

Alot of people were using Innertube's etc stuff that would degrade with heat. I started by using a gasket I made from Cork but the Knock sensor was all but disabled.

So what I did Is went to my local Autoparts Store, they should have new oil drain plug kit's, there will be gasket's/washer's that are about 1-2mm thick and have some sort of heat resistant padding on them that help's seal the oil drain plug.

I simpley removed the nkock sensor and put the gasket between the block and the sensor, the sensor will still be grounded through the bolt.

My MAF Recentley blew and My car was running real lean at certain RPM's, the knock sensor was still able to pull enough timing and I never heared a ping.

It has completely Cured my hesitation, I did do the SPG mod and didn't find it helped that particular problem.

Originally posted by RageHardIntoTheBendies
You might also want to check your 02 sensor.
I'm still at a loss for what might be causing this. I was just thinking about it again last night and realized it couldn't be the O2 sensor because it happens at all throttle positions, including WOT, at which point the ECU is in open loop and doesn't even use the O2 sensor.

I still need to order one of those Palm-compatible data loggers.

Pat

ragehardintothebendies,

where exactly is the knock sensor located? i bought some insulated oil drain plugs today....

thanks! :)

peter

Top of the engine, under the intake rails, driver's side IIRC ;). Pat, try just unscrewing the knock sensor from the block but leave it plugged into to it's electrical connector (tape it up and away from the block somewhere). See if you still have hesitation- if you do, it's not the knock sensor. If you don't, it must be the problem. I have also heard of people putting washers on the knock sensor before bolting back into the block, to "dampen" the noises it sees. Older Subaru's had definite problems with overactive knock sensors. A guy that went by the name of "HobieGary" used to sell kits on eBay (the thing I dismantled in the "paperclip knock sensor" page on the Ravensblade Impreza site, which amounted to a resistor which accomplished the same thing- dampening the signal the knock sensor saw.


Brian
http://www.ravensblade-impreza.com
http://www.imprezamods.com
http://www.scoobymods.com
http://www.deniedmywarranty.org

Well, I tried removing the knock sensor as Brian suggested (unbolted it, zip-tied it out of the way, still plugged in though) and the problem did not go away. If anything, the car felt even more sluggish, but that may have been my imagination.

I just don't know what it could be. :(

Pat

Pat,
I'm having similar issues w/ my car. Not that I have the heads or cams, but its' gotten me thinking. Maybe it's time to check my cat. Is it possible that since I've been running rich I've clogged it? maybe there is a loose wafer or something that when there is lower back pressure causes it to not be able to breath well?
AO

It's a possibility, Aaron, but I just don't know. Based on the way my car is behaving I can't help but think it's something sensor and/or ECU related that is causing the problem. I wouldn't think a mechanical problem would behave like this.

Oh, and the hesitation seems to be spreading. Now it's sometimes doing it around 4500rpm too. *sigh*

Pat

I hear ya. Something in mine is flakey as hell, and I've yet to figure it out. But I can tell you what it's not (mine that is) coil, plugs, or fuel filter. argh!... I'll tell ya what they say about mine at the end of the week, back to exeter for friday. Not sure what they are going to do, but... :rolleyes: I hope it works when it comes back.
AO

Pat - I guess its not throwing a CE light or you would have said so.

I know we have the stronger hot wire MAF sensor. But I'm thinking with your mods that MAF could be vibrating and getting confused with intake manifold resonance. If its progressively getting worse the MAF could be slowly dying. Heres were a good pocket logger would come in handy. I assume you have an A/F gauge installed to tune the AFC. Next time you get that hesitation take a quick look at the A/F meter and see if your running really rich or really lean. If your AF ratio is irradic it sounds like the MAF faily or beeing disrupted by manifold resonance or vibration. To fix this maybe try to stick the stock air box back on for a period. As you probably already know the stock box and panel filter did a good job of breaking up the turbulence caused by the intake manifold when the engine is hummy really good.

Let me know what happens.

Its definitely not the O2 sensor. I just went through replacing mine. It had a failed heater circuit. It was open all the time requiring the exhaust gasses to be super heated before the sensor would get a correct reading. The car would hesitate sputter and down right just not go. However when you enter Open Loop at WOT its like a different car that runs perfectly fine.

Brad

I have the complete stock air intake installed, with the exception of the silencer being removed and a K&N panel filter installed in the stock airbox. That's the way things have been the whole time. (Well, I had a Weapon R installed for a day, but I didn't like it and it didn't seem to gain me anything on the dyno so I went back to stock). I have an Injen intake sitting in my garage but I want to get this problem squared away before I change anything else.

Yes, I have a CEL, but that's because the Cobb cams cause a misfire code on all 4 cylinders. That's something I need to clarify with Cobb - this hesitation problem could be a misfire, so how do I know if I have a real misfire or just a perceived misfire? And why exactly do the cams cause a misfire code?

I pulled the plugs last night. The electrodes on all 4 were pretty white, maybe the slightest hint of tan on one or two. If my memory serves me right, that indicates the car is running lean, yet my exhaust tip turns black over time. :confused: The plug gap was pretty good, but on the high side of the recommended range, so I re-gapped them all to 0.039" (minimum spec) with the thought that perhaps this is a misfire problem.

I think that the general consensus here is that when you reset the ECU the car initially runs on the safe side (rich). I tuned the S-AFC on the dyno immediately after doing an ECU reset. Perhaps as the ECU learns, it leans out, making my S-AFC settings too lean? One would hope the ECU was smart enough not to learn fuel settings that were so lean it caused detonation and timing retard by the knock sensor, but who knows. Whatever the case may be, this morning I richened the mixture across the board, and I'll do an ECU reset today to see how things develop over the next few days. (After doing an ECU reset the car does feel better - not cured, but better - for a day or two, but the probably gradually returns. Perhaps with richer settings in the S-AFC it won't get worse, I dunno).

Pat

Very Strange. Not sure the reason for the ECU reading mis fire on all four. I would definitely check with Cobb.

You said this hesitation happens both while cruising and at WOT? If so your problem has nothing to do with the ECU learning as it doesn't learn anything in Open Loop.

Brad

pat...
just an additional note. I've got misfire codes on all three aswell.
I'll let you know what they say.
AO

My problem persists. I had my valves adjusted last week in the hope it was some sort of valvetrain noise that was triggering the knock sensor. No such luck - the valves did need adjustment, but the hesitation came back within a couple days (as it normally does after reseting the ECU).

I recently acquired a www.pocketlogger.com for my Palm Pilot and have confirmed that it is timing retard that I'm feeling. I still don't know what's causing the timing to be retarded, though.

I filled up the tank this afternoon (93 octane) and added a can of octane boost. It still hesitates.

I'm still confused. :confused:

Pat

What about upping the octane? Can you get 100? Did you try getting an oil plug gasket, and putting that between the engine and knock sensor? Can you get your hands on another ECU, and trying to see if that's the problem? Damn...:(

Pat,

Try the oil pan gasket mod on the knock sensor. Reset the ECU using the battery dance method. Then see what happens. If you still are having issues......the only suggestions I can think of is carbon or some sort of build up........or.....bad ECU.

Another thing to try that has worked for me in the past is resetting the ECU and as soon as you start the car drive it at 50% throttle going to redline in the lower gears with the AC on full blast.

Basically load it as much as possible but dont go to WOT.

This sounds similar to something I encoutered (well, caused) when I had my RS. I has a S-AFC and a JC Sports intake installed. I was messing around with leaning it out and making it richer over a few weeks time to see if I could feel any difference. I went to a point where I went a bit too far on the lean side and the car detonated once. After that, it ran like crap and just got worse. I added fuel back, still ran poorly. Similar to what you explain. i reset the ECU and put the AFC back to zero and started over. I know you reset the ECU alradys, so thats puzzling. I know the DSM guys have an issue of the ECU overriding the AFC at closed loop operation. They have to add a ton of fuel in those areas to max out the ECU's compensation capabilities, then the AFC can start to add. So basically, you add say 15% with the AFC, the ECU can take away 15%, so to get a true +5% you have to set the AFC to +20%. It would be interesting to see (maybe you already have) what your AFR looks like on the dyno when your car is in its "worst shape" (most hesitation). Maybe it is lean....

one thing worth checking.. plug wires

A couple of good point, Steve. The hesitation actually started before I installed and tuned the S-AFC. As a matter of fact I have a couple of dyno runs that I think show the hesitation and the AFR is pretty solid bouncing around 12.75:1 (just like it is for the rest of the run). I say "I think" the dyno graph shows the hesitation because I wasn't actually driving the car on the dyno, so I didn't feel it, and the dyno operator didn't say anything about it. Comparing the graph to the previous run, though, there's a pretty pronounced dip in the torque band from ~2400 to ~3500rpm. Anyway, with all that said, I did reset the S-AFC back to all zeros today, and I'm going to go unhook the battery now to reset the ECU and we'll see what happens in the next couple days.

As for the plug wires, I hope they're OK, as they're relatively new Magnecor wires. They were installed a few thousand miles before this problem developed. I have been thinking of re-installing the OEM wires just to see if that makes any difference. If the zero'd S-AFC test doesn't work, then I think the plug wires will be the next thing I change.

Pat

Dang, you're getting misfire on your street cams too? Right now I have a intermittent misfire on #1, but I had all 4 going for awhile. Seems valve lash has something to do with it, and its VERY sensitive to changes. That said, I think Im losing alot of top end on my RS. I got 189 lb/FT of torque but only 176 hp. Thinking it pulling timing up top. Maybe a similar problem. Check out road dyno, neat little tool to see what your engine is doing.

Here's a graph of timing and RPM vs. time for a 3rd and 4th gear pull at around 40-50% throttle:
http://www.submariner.org/thepno95/Pictures/Subaru/timing%20retard.jpg

The data comes from my PocketLogger ( www.pocketlogger.com - pretty cool OBDII scanner and datalogger). As you can see the timing is seriously retarded in the mid-range. I've done a couple more runs that I need to convert to graphical format (I don't have Excel here at home, so I have to take the files to work). For whatever reason, today the car feels absolutely perfect, and it felt the same two days ago. However, yesterday it was its "normal" hesitating self, so I don't know what the deal is. It's amazing how much more powerful the car feels when it doesn't dump so much timing in the mid-range (which is where RPMs are during normal driving).

Pat

I started a new thread with this info, but I figured you might have a better chance of seeing it here....

For all you long time "hesitation" sufferers I have a solution for you. My g\f's '98 Legacy 2.5 GT with only 70k on it, started showing really bad hesitation at low rpm's 1000- 3500. It was terrible, you'd practically get whiplash every time you started from a stop, and if you tried to punch it the engine would cough and choke. Then it finally through a misfire code after a few hundred miles.

Long story short, I looked into the common ground, MAF, and knock sensor fixes I saw posted on this site, but before I tried them out I talked to a Subaru mechanic and he said he had never heard of any of these. So he got to thinking and told me that at least on the 2.2L engines sometimes they would get build up on the exhaust valves and that it would screw up the seats and make the valves stick, but hadn't ever heard of this happening on the 2.5L.

So he recomended that I take some industrial strength injector cleaner and run it straight into the intake manifold. Now I am NOT talking about the $5 crap that you add to your gas, I mean the stuff that runs about $20-$30 and comes it a compressed can and you need special equipement to hook it up to your injector rail.

I took the car to my buddy's shop told him the story and luckily he had a can of the right cleaner. He basically hooked a hose up to the can and ran it into the main vacuum on the top of the intake manifold. He emptied about half the can while the engine was barely idling, the cleaner bogs the engine down so you might need to give it a little gas, but keep rpm's as low as you can it lets the cleaner kind of soak in. After half a can, he let the engine idle on its own for a few minutes to let the cleaner work in a little, then he revved the piss out of it to blow it out. Right away we could tell that the hesitation was gone!!!! We couldn't believe that it was that simple. We went ahead and used the other half of the can up to make sure we got all the buildup out. Keep in mind that your car will smoke bad for a little while while all the crap burns out of the exhaust pipe, so don't freak out.

I hope this helps, I'm sure any shop can do this. Let me know if you have any questions.