Jump to content
Ultimate Subaru Message Board

Carbon vs Knock Sensor '96 OBW


Recommended Posts

Alright, let me begin by saying that I have spent many hours poring over whatever information I could find online, including this site re: Subaru hesitation, its causes and solutions. Over the past year or so good things have popped up here, and I'm hoping that some of you can shed some light. You might be bored with this issue, but then there are those that should read on.

 

My '96 OBW 2.5 Auto with 93k has had this famous hesitation thing as long as I can remember (bought with 36k in 2000). Between 2500 and 2700 or so revs during acceleration, especially going uphill, the engine feels likes it can't decide whether to go faster or not, like mollasses got in there or something. Then at the high end of that range it comes out of it and smoothly heads on through the power band. No bucking or pronounced stutter, just a bit of sluggishness, a nervous tach needle, right through that range. I have to floor it sometimes to really get the horses out. It's extremely irritating and gas-guzzling.

 

So, what have I had done to the car? Been a while since I've replaced anything (spent 6 months hiking the Appalachian Trail this past year), but at around 75k I replaced the timing belt, crank seals, fuel filter, plugs and wires, pcv valve, K&N air, trans. and diff. fluids, cleaned the TPS and had one of those fuel system vacs done at the dealer. All of this for love of the car and to try and fix the hesitation. No good. I run fuel system cleaner through every once in a while. Even when I use the recommened high octane for this engine, it is still there. I keep thinking that this is computer/electrical related.

 

I read in several places about knock sensors and their eagerness to share every little sound with the ECM, and perhaps non-knocking noises are retarding the timing. This kind of thing is apparently common with other makes as well. So then there are ideas about adding resistors to the wire to hush the signal a bit.

 

Well, I finally just decided to visit the dealer today and find a decent Soob mechanic who could answer some questions. Got to talking with this guy back in one of the bays, and I asked him about sensitive knock sensors, any upgrades from Subaru, and any relationship these have with the hesitation I described. No. They hadn't seen anything about it. Hmm. He pointed out that my '96 2.5 has solid mounts, and around the RPM range I spoke of the engine shudders and/or resonates, which I might think is a hesitation. Could the ECM pick this up as knock? Not likely. I could get liquid-filled mounts. Also, the engines are prone to carbon buildup. BG system cleaner and some Chevron gas should keep things clear. Hmm. And, bummer, I did not get the chance to take him for a ride.

 

I'm looking for more information, especially on these topics:

 

-- Would carbon buildup possibly cause the hesitation in that RPM range, but not outside of it? If so, why?

 

-- If there truly is knocking, and the ECM retards timing, would I be able to really feel that change while driving? Is there some way to monitor that to see if the ECM is making changes?

 

--If this engine shuddering thing is true, and perhaps that and other engine vibration are being picked up by the knock sensor as knocking, what do I need to find out? Can it be measured? How much resistance should be used inline if this does happen, and what is the Ohms spec on the sensor? Has anyone made their own resistor setup and what have you used? I have consulted these pages:

http://www.charm.net/~mchaney/knock.gif

http://www.geocities.com/MotorCity/6647/ta93/knock.htm

http://www.ravensblade-impreza.com/modifications/drivetrain/paperclip/paperclip.html

 

--Has Subaru made any changes to this sensor to make it less sensitive? Have they produced a replacement ECM with updated programming to account for the sensitivity?

 

--What measureable effect does a rubber washer or other material have at the sensor location?

 

--Do you have a '96 with the 2.5, and have you experienced any of the stuff I'm talking about?

 

Call me obsessive, but I just want my car to run like it should, you know? Somebody out there has answers. Please Share.

Link to comment
Share on other sites

You are on the right track. Sounds like you have done your homework.

 

I have a 97 OB with a slightly revised version of your engine.

 

I have never suffered the 2500ish hesitation problem, but I have suffered the carbon problem. The carbon problem was only noticable at idle on my car. That is, when starting off from a stop. It was between 600 and 1000 rpm. As it got worse, it crept upwards to 1500 rpm. Rather, I should say that the range it affected was 600 to 1500 rpm. A MotorVac cleaning took care of my problem for quite some time.

 

I'd be more suspicious of the knock sensor. My understanding is that the replacement units are not so sensitive. Perhaps someone could confirm.

 

One way you should be able to tell is by resetting your ECU. It takes a few times for the ECU to learn the knock problem. Therefore, it should feel fine the first few times you accelerate thru the problem range after a reset.

 

I'm sure others will have more feedback for you.

 

Commuter

Link to comment
Share on other sites

mine is not a newer model its a 87 gl-10 turbo

but i had the same hesitation problem it went away after i regrounded the engine so all the sensors and ignition had the proper grounding ..

maybe somethign to look into

maybe not

Link to comment
Share on other sites

Truk, here's a little info I've picked up about the sensors.

 

On Subes and GM's anyway, they're tuned to the frequency of the bore size, meaning that if you remove a cylinder head, rotate the crank to get a piston at BDC, then tap that bore with a brass hammer, it will have a fairly well-defined bell-like pitch which is proportional to the bore diameter (larger bore has a lower pitch).

 

If the pistons are clattering in the bore, the sensor picks it up. It's a crude system which is hampered even more by the boxer design. It would be better to have two sensors (one mounted directly on each cylinder bank. But as it's mounted centrally on the crankcase it picks up all the resonances of the rotating mechanisms within the crankcase as well.

 

My '02 was doing what you describe during the test drive when I signed for it at the dealer. It had a gummy, gooey, throttle non-response.

 

Above a certain RPM the sensor is ignored by the ECU as you describe, so in city traffic driving you get a lot of down-shifting by the auto-trans in order to make routine maneuvers. It makes for an unpleasant driving experience. :(

 

I measured my sensor at almost exactly 560k, cut the wire to the sensor, soldered in a 560k resistor connected at one end to ground as a dummy "sensor", and I haven't looked back. The sensor is out of the circuit. ;)

Link to comment
Share on other sites

I'm not one to advocate just replacing parts as a method to solve a problem, but... I don't think the knock sensor costs very much and you can change it yourself. They do age/fail. Why not just try replacing it?

 

Commuter

Link to comment
Share on other sites

Blitz, since it sounds like you cut off the signal for knock, did you get a Check Engine light? The ECM doesn't care? What about real knock that could damage your engine? And this gave you smooth acceleration, a real difference right away?

 

I am close to constructing something like you described, based on the stuff on the pages I noted, but I don't necessarily want to block out true knock sounds. So it would be like a 30% reduction or something.

 

And I have looked into just replacing the sensor, supposing it does not do the best job anymore, but if there is any info on sensor improvements by Subaru, I would like to know.

Link to comment
Share on other sites

Originally posted by truk

Blitz, since it sounds like you cut off the signal for knock, did you get a Check Engine light? The ECM doesn't care?

 

The ECU does a self-check for an open or short condition of the sensor circuit via a voltage drop across the sensor's resistance. As long as you replace the sensor with a matching resistor value, the ECU doesn't know the difference. Just make sure the resistor has a good ground.

 

What about real knock that could damage your engine?

 

I took care of that like real men used to (before the inclusion of granny features). I increased the fuel octane/de-carboned the cylinders as required.

 

And this gave you smooth acceleration, a real difference right away?

 

It fixed the "problem" for 19 cents.

 

I am close to constructing something like you described, based on the stuff on the pages I noted, but I don't necessarily want to block out true knock sounds. So it would be like a 30% reduction or something.

 

I studied those solutions, but decided that it would bother me to not know if the sensor was messing with my timing behind my back, so I just got rid of it. How do you determine when you've got it right? It's too crude of a system.

 

And I have looked into just replacing the sensor, supposing it does not do the best job anymore, but if there is any info on sensor improvements by Subaru, I would like to know.

 

I don't have any info on upgrade sensors or anything like that.

Link to comment
Share on other sites

THanks, Blitz, I might have to give that a try. The white connector is right there under the throttle wires and one could plug in a resistor there and ground it. Nice.

 

Anyone know if the 560k ohm reading for the sensor is right for a '96 as well? Who's got one of those shop manuals?

Link to comment
Share on other sites

Originally posted by blitz

Truk, here's a little info I've picked up about the sensors.

 

On Subes and GM's anyway, they're tuned to the frequency of the bore size, meaning that if you remove a cylinder head, rotate the crank to get a piston at BDC, then tap that bore with a brass hammer, it will have a fairly well-defined bell-like pitch which is proportional to the bore diameter (larger bore has a lower pitch).

 

If the pistons are clattering in the bore, the sensor picks it up. It's a crude system which is hampered even more by the boxer design. It would be better to have two sensors (one mounted directly on each cylinder bank. But as it's mounted centrally on the crankcase it picks up all the resonances of the rotating mechanisms within the crankcase as well.

 

My '02 was doing what you describe during the test drive when I signed for it at the dealer. It had a gummy, gooey, throttle non-response.

 

Above a certain RPM the sensor is ignored by the ECU as you describe, so in city traffic driving you get a lot of down-shifting by the auto-trans in order to make routine maneuvers. It makes for an unpleasant driving experience. :(

 

I measured my sensor at almost exactly 560k, cut the wire to the sensor, soldered in a 560k resistor connected at one end to ground as a dummy "sensor", and I haven't looked back. The sensor is out of the circuit. ;)

 

Do you run premium fuel now?

Link to comment
Share on other sites

Originally posted by myles

Do you run premium fuel now?

 

Yeah, 93 because I've got another 5 degrees or so of advance dialed-in via 56k resistance in series with the AIT sensor. That's 5 degrees on top of the 10 or so I was losing to the knock sensor (15 total).

 

Also I'm ducting cold air in from the fender while retaining the torquey stock intake tract, and have a cooler thermostat installed. It really woke this motor up and especially helped build solid tractor-like throttle response below 4k.

 

I have one last thing I want to do which is coat the inside of the aluminum manifold with an industrial thermo-coating. That's gotta be good for another 3 or 4 ft/lbs. in heat-soak driving conditions.

Link to comment
Share on other sites

Originally posted by truk

THanks, Blitz, I might have to give that a try. The white connector is right there under the throttle wires and one could plug in a resistor there and ground it. Nice.

 

Anyone know if the 560k ohm reading for the sensor is right for a '96 as well? Who's got one of those shop manuals?

 

Borrow an ohmeter and measure it (dissconnected from the ECU!). :eek: The exact number isn't super critical, you could probably be within 20% or so and easily fall within the ECU's allotted parameters (there is an allowable range).

 

Also if it makes you feel more comfortable, you could include a SPDT mini-toggle in the cockpit that connects the ECU lead to either the grounded resistor or the sensor (normal/bypass).

 

Always use solder, heat shrink, tie-wraps and ground-lugs. Don't wind, twist, tape, rig, or otherwise do a "hillbilly" wireup. You want the job to look professional be as secure as the factory would've done. :cool:

Link to comment
Share on other sites

a toggle switch is a good idea. Thanks.

 

And I measured with a mulimeter, though I must be sloppy. In x1k mode, which I assume is 1000 times whatever the number, the needle dipped just slightly down from infinity to hair beyond 1k when grounded. This would mean something less than 1,000,000 ohms. Not too good of a number. Need a digital perhaps.

 

Blitz says 560k for his. It could be around there.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...