porcupine73 Posted August 25, 2006 Share Posted August 25, 2006 Hi all, somebody please warn me if this is a stupid idea before I drive 35 mi to work in the morning Just removed the knock sensor from the block on my '96 Legacy 2.2 and wrapped it in bubble wrap and secured it loosely with tie wraps to a nearby hose (left connector connected to avoid CEL and resulting fully retarded timing). Then unhooked battery for a couple minutes to (hopefully) make the ECU forget it's previous life as a knock retarded processor and start it's new life tomorrow morning as a new zero knock retardation engine! Anywho, she's filled with 93/94 premium fuel, plus I poured in maybe 6-8 oz. xylene to try to get even more octane, goal being that if there's any knock it won't be because of too low an octane gasoline. I'll keep the radio off and windows down so hopefully I'll hear any deadly knock. Have tools in the car to bolt sensor back in if need be. Will report on results (provided I make it back home in one piece!!) Quote Link to comment Share on other sites More sharing options...
porcupine73 Posted August 25, 2006 Author Share Posted August 25, 2006 Actually I hope the knock sensor isn't just looking for a ground signal as now that I think about it it seems like the connector has only one wire.....maybe mesha better go have a look see.:-\ Quote Link to comment Share on other sites More sharing options...
OB99W Posted August 25, 2006 Share Posted August 25, 2006 Actually I hope the knock sensor isn't just looking for a ground signal as now that I think about it it seems like the connector has only one wire.....maybe mesha better go have a look see.:-\ Yes, one connection is to ground, and the ECU won't be happy with the knock sensor dismounted. It will probably give you a P0328 code (and naturally, a CEL to let you know ). If you know what resistance the knock sensor measures, you can try wiring a resistor equal to it between the connector and ground. Apparently some have reported using a 560,000 ohm (560 kohm) one successfully. Quote Link to comment Share on other sites More sharing options...
porcupine73 Posted August 25, 2006 Author Share Posted August 25, 2006 Thanks for the info OB99W. I think I'll just stuff a wire down in the bolt hole on the thing and jam some paper in there to secure it and then ground it with an alligator clip before heading off to work. If the ECU gets pissy I'll just pull over and remount the thing and unhook the battery for a minute. Quote Link to comment Share on other sites More sharing options...
skizix Posted August 25, 2006 Share Posted August 25, 2006 I got sporadic knock senor codes in my '90 legacy, and after a while it became constant. In my case, the k.s. housing had cracked (a common problem in those cars). I could tell my engine was not knocking -- instead the sensor was hosed and constantly sending a bad signal. At 230,000 miles, I knew I wasn't going to keep the car much longer. That combined with how hard it was to access the k.s., plus how expensive a new one was, prompted my to depricate the sensor. A cleaner way to do it is to replace the knock sensor with a resistor. If you look in your Chiltons or whatever, there should be a procedure to test your knock sensor, listing a resistance value that a good knock sensor should read when it is not activated. I went to radio shack, paid 50 cents for the closest value resistor they had. Left the sensor in place, but disconnected the lead to it, attached that to one end of the resistor, and the other end of the resistor to a ground screw in the firewall. Wrapped the resistor well in electrical tape, and I had a nice tidy, inconspicuous rig. Made the CEL go out for good, car drove fine. Much cleaner than bubble wrap. One thing about that approach is that when you take your car in for emissions testing, if they see your knock sensor all wrapped up in bubble wrap, they could conceivably fail you for disabling an emission control (it seems debatable whether the k.s. is actually an emisson control, but I would think they'd get uptight about disabling anything that makes the CEL go on). They certainly never spotted my rig. I always ran mid-grade fuel in that car instead of the recommended regular, as regular seemed to make it knock a bit. After the knock sensor disable, I kept running mid-grade, and the car seemed to run just great without a k.s. Just don't lug the engine, and I think the driver can become the knock sensor without too much worry, as long as everything else is running ok. Quote Link to comment Share on other sites More sharing options...
nipper Posted August 25, 2006 Share Posted August 25, 2006 The knock sensor doesnt need a ground. Its a peizo electric quartz sensor that creates an ac voltage that is sent through the ECU. When you put a meter to it and check to see if there is a circuit to gound, there should be none, if there is its a failed sensor. The AC signiture varies with the amount of knock. That said the ecu seeing no signal from the sensor, you will most likely get an CEL, and also possibly damage your engine. Not all sensors work off of resistance. DONT use a resistor, as it has no resistance. Even the haynes manual describes quite well how these things work. DO NOT GROUND A SENSOR WIRE OR PUT A RESISTOR IN ONE unless you are 100% sure as to how the thing works. You can damage an ECU this way. The knock sesnor hears knock before you do and corrects for it. Ping it cant do anything about because of the way ping happnes. Pinging may lead to knock, but ping by itself is realtivly harmless if annoying. If you want to argyue this point go right ahead as after umpteen posts on ping im not going to defend it, just do a search on it here. nipper Quote Link to comment Share on other sites More sharing options...
skizix Posted August 25, 2006 Share Posted August 25, 2006 Well, on the '90, the test procedure for the k.s. involved measuring resistance from the connector on the sensor to ground. I replaced it with a resistor that was close to the spec for a good, idle knock sensor, and it worked flawlessly for my purposes (which were to milk some more miles off a car that had done almost the equivalent of driving to the moon). Put another 10,000 on that car afterward (before selling it in still excellent running condition). Drove very well, and without a hitch. Perhaps the k.s. on a '96 is a completely different mechanism, or at least a different circuit design. No idea. Yes, I would recommend understanding the function of a sensor as well as possible before trying to remove, spoof, or otherwise alter it. I would not even consider doing that to my current car, as it is still in it's prime, and actually worth some $$$ (in the '90, at that point, a new knock sensor would've set my back around 15% of the market value of the car). As usual...YMMV. Quote Link to comment Share on other sites More sharing options...
OB99W Posted August 25, 2006 Share Posted August 25, 2006 The knock sensor doesnt need a ground.The wiring diagrams I've looked at all seem to indicate otherwise. The wire to the sensor is 2-conductor (shield plus center conductor), which might lead one to believe that a ground isn't needed (that it's seemingly provided via the shield). However, the shield is "floating" (not connected) at the sensor end, and a ground connection is made by the mounting of the sensor itself. Its a peizo electric quartz sensor that creates an ac voltage that is sent through the ECU. When you put a meter to it and check to see if there is a circuit to gound, there should be none, if there is its a failed sensor. The AC signiture varies with the amount of knock.Yes, the knock sensor is a piezoelectric device. Piezoelectric mineral-crystal or ceramic elements have extremely high resistance, and by themselves typically measure open-circuit by ohmmeter. That's one reason (there are others) manufacturers sometimes incorporate a high-value resistor in parallel with the piezo element; it allows a small current to flow (in the case of a 560,000 ohm resistance, and even assuming full automotive battery charging voltage is applied, only about 25 microamps), making possible remote sensing of wire breaks and/or other disconnects. The high resistance of the paralleling resistor won't swamp out the relatively low voltage generated by the piezo element. That said the ecu seeing no signal from the sensor, you will most likely get an CEL, and also possibly damage your engine.The ECU is probably looking for some "pull down" by a connected knock sensor, which the resistance (although high) can provide. If it doesn't see that, it "assumes" the sensor/connections are bad. Not all sensors work off of resistance. DONT use a resistor, as it has no resistance. Even the haynes manual describes quite well how these things work.Actually, what it has is high resistance. DO NOT GROUND A SENSOR WIRE OR PUT A RESISTOR IN ONE unless you are 100% sure as to how the thing works. You can damage an ECU this way.Agreed on . The knock sesnor hears knock before you do and corrects for it. Ping it cant do anything about because of the way ping happnes. Pinging may lead to knock, but ping by itself is realtivly harmless if annoying. If you want to argyue this point go right ahead as after umpteen posts on ping im not going to defend it, just do a search on it here.Agreed, knock and ping aren't the same. Quote Link to comment Share on other sites More sharing options...
nipper Posted August 25, 2006 Share Posted August 25, 2006 I should have explained myself better oh well... THis is how a knock sensor works. http://www.autoshop101.com/forms/h38.pdf#search=%22what%20is%20a%20knock%20sensor%20%22 Quote Link to comment Share on other sites More sharing options...
porcupine73 Posted August 25, 2006 Author Share Posted August 25, 2006 Interesting notes everyone! Those knock sensors are interesting. Functioning seems not entirely unlike an accelerometer. Anyway, drove 100 miles today, didn't get a CEL, at least not yet (I stuffed some Al foil into the inside of the knock sensor where the bolt goes and then hooked up a jumper lead between that and ground in case it was needed). My guess is if the CEL comes on it goes to maximum retardation as a fail safe. The only real thing I noticed was maybe the low end, below 3000rpm seemed to be more smooth and maybe more power. Before, the thing liked to bog down almost if getting on the gas below 3000rpm; now it doesn't seem to do that. Of course it could just be that I disconnected the battery last night so I'm not sure. Anyway I guess there really should only be a difference if the knock sensor's output was making ECU retard timing and now it isn't. Quote Link to comment Share on other sites More sharing options...
nipper Posted August 25, 2006 Share Posted August 25, 2006 of course you can always get a new knock sesnor, but hey what fun is that:rolleyes:. Keep in mind now if you ever do develop heavy ping or knock, there is nothing your ecu can do to stop engine damage. nipper Quote Link to comment Share on other sites More sharing options...
OB99W Posted August 25, 2006 Share Posted August 25, 2006 I should have explained myself better oh well... THis is how a knock sensor works. http://www.autoshop101.com/forms/h38.pdf#search=%22what%20is%20a%20knock%20sensor%20%22 Nipper, hopefully nobody here was questioning your description of how a piezoelectric knock sensor produces its output. The original discussion was about the consequences of dismounting one. As I'm sure you know, various sensors in the car are monitored for function. Even those whose output is a varying AC signal can be checked on a basic level (connection, etc.) via DC means, if they have some level of conductivity. One type that comes to mind is an ABS wheel sensor. When the car is first started, before a wheel is moving (and therefore before any AC "signal" is developed), the ABS ECU checks the wheel sensor(s). It's done via monitoring an applied DC voltage; if it's out of a range that the ABS ECU expects, it indicates the sensor's resistance is incorrect, due to either a sensor or wiring fault. The knock sensor can be similarly (basically) verified by the main ECU. A particular knock sensor that Subaru uses has a spec resistance of 500-620 kohms, the middle of that range being 560 k. It seems the ECU won't flag a problem if a resistor of that value is substituted for the sensor. By the way, Nissan has used a k.s. with the exact same resistance spec, although I'm not sure it's identical in other respects. Quote Link to comment Share on other sites More sharing options...
porcupine73 Posted August 25, 2006 Author Share Posted August 25, 2006 Hm nipper, the idea of a new knock sensor sounds good....I mean it seems there is not a small number of posts about those things going batty. And this one is 10 years old. Then I'd have the protection of the knock sensor without the possible damage of nasty knocking. Actually I'm not even sure; would I notice knock in the passenger compartment? I just pulled it out last night since it was just one bolt and it seemed like fun and using premium gas and well maintained vehicle I figured hopefully I had no reason for knock. Actually, as you mentioned about checking mpg using the scangauge in other threads, if I ran at some given load%, could/should I expect that to be a good indicator for checking the effective timing retard on the scangauge? Thanks buddy... Quote Link to comment Share on other sites More sharing options...
nipper Posted August 25, 2006 Share Posted August 25, 2006 Hm nipper, the idea of a new knock sensor sounds good....I mean it seems there is not a small number of posts about those things going batty. And this one is 10 years old. Then I'd have the protection of the knock sensor without the possible damage of nasty knocking. Actually I'm not even sure; would I notice knock in the passenger compartment? I just pulled it out last night since it was just one bolt and it seemed like fun and using premium gas and well maintained vehicle I figured hopefully I had no reason for knock. Actually, as you mentioned about checking mpg using the scangauge in other threads, if I ran at some given load%, could/should I expect that to be a good indicator for checking the effective timing retard on the scangauge? Thanks buddy... oh would you EVER notice knock, its affects performance, then there is that entire melting of the piston thing nipper Quote Link to comment Share on other sites More sharing options...
porcupine73 Posted August 25, 2006 Author Share Posted August 25, 2006 Thanks nipper... hm...melting of the pistons...I'm not a real experienced car guy, but I'm assuming that's bad... Just curious...what did cars do before knock sensors? Did you have to kind of be aware of what was going on or what? Quote Link to comment Share on other sites More sharing options...
nipper Posted August 25, 2006 Share Posted August 25, 2006 Thanks nipper... hm...melting of the pistons...I'm not a real experienced car guy, but I'm assuming that's bad... Just curious...what did cars do before knock sensors? Did you have to kind of be aware of what was going on or what? its not that simple. Cars wyould run far richer then they do today, so its not an issue. As pollution controls were added, the engines strated running leaner. Right now they run so lean that some pinging will happen and is considered acceptable ubder some circumstances. If you had an older car that ppinged you turned the distrubutor to get rid of it. Now the puter does all that stuff.... and melted pistons bad.... nipper Quote Link to comment Share on other sites More sharing options...
OB99W Posted August 25, 2006 Share Posted August 25, 2006 [...]Anyway, drove 100 miles today, didn't get a CEL, at least not yet (I stuffed some Al foil into the inside of the knock sensor where the bolt goes and then hooked up a jumper lead between that and ground in case it was needed). My guess is if the CEL comes on it goes to maximum retardation as a fail safe.[...] So now the ECU sees the k.s., but it can't respond to engine vibrations. Whether that's a good thing is a discussion I won't get into . Even assuming a k.s is "tuned" to respond to particular frequencies, anything banging around causing vibrations that reach the engine could make the k.s. output something that the ECU sees as engine knock (detonation). Sometimes simple things such as exhaust system rattles can be a problem. A timing light can tell you a lot about how the k.s. responds; hook one up, try rapping near a mounted k.s. (not directly on it), and see how the ignition timing changes. When mounting a knock sensor, be sure to tighten the bolt within torque specs. They can give grief if loose, but can be permanently damaged if sufficiently overtorqued. Quote Link to comment Share on other sites More sharing options...
porcupine73 Posted August 25, 2006 Author Share Posted August 25, 2006 Thanks as always for the info nipper. Now I wish I woudn't have sold my brand new timing light thinking 'i'm done with distributors! when I sold '88 town car. Well I think I should maybe put that knock sensor back in and maybe do your hammer test. I could put it back in and see if I start getting that <3000rpm hesitation again. And/or scangauge timiing readings. Or some combination of options..... Quote Link to comment Share on other sites More sharing options...
OB99W Posted August 25, 2006 Share Posted August 25, 2006 Thanks as always for the info nipper. Now I wish I woudn't have sold my brand new timing light thinking 'i'm done with distributors! when I sold '88 town car. Well I think I should maybe put that knock sensor back in and maybe do your hammer test. I could put it back in and see if I start getting that <3000rpm hesitation again. And/or scangauge timiing readings. Or some combination of options..... Anyone remember the group "Devo", and "We're all Devo"? Apparently, we've all become "nipper" here . --The guy formerly referred to as "OB99W" Quote Link to comment Share on other sites More sharing options...
nipper Posted August 25, 2006 Share Posted August 25, 2006 Anyone remember the group "Devo", and "We're all Devo"?Apparently, we've all become "nipper" here . --The guy formerly referred to as "OB99W" Your all scaring me again ....... nipper Quote Link to comment Share on other sites More sharing options...
porcupine73 Posted August 26, 2006 Author Share Posted August 26, 2006 Whoops my bad! Sorry about that OB99W! We're just all one big happy family on this board and it gives me a warm fuzzy just thinking about it. Thanks for the info about the timing check idea. I could probably get one 'rental' at autozone maybe. Say what's the 0-60 supposed to be on the '96 Legacy 2.2AT? This thing feels pretty zippy. Quote Link to comment Share on other sites More sharing options...
OB99W Posted August 26, 2006 Share Posted August 26, 2006 Your all scaring me again ....... Me too; think I'll take a break for the rest of the day! Quote Link to comment Share on other sites More sharing options...
OB99W Posted August 26, 2006 Share Posted August 26, 2006 Whoops my bad! Sorry about that OB99W!No problem, now that I have my identity back . We're just all one big happy family on this board and it gives me a warm fuzzy just thinking about it. Yep, the cooperation is amazing, although we do have our little disagreements (umm, "discussions") now and then (I suppose just like real families). In addition, from what I've seen in reading other Subaru-related forums, there certainly seems to be a higher concentration of technical expertise here than on the other boards. Thanks for the info about the timing check idea. I could probably get one 'rental' at autozone maybe. Say what's the 0-60 supposed to be on the '96 Legacy 2.2AT? This thing feels pretty zippy.Of course a real-time OBD-II scan can be very revealing, but even with engine management systems the basic tools can be useful. I still find things like a vacuum gauge and timing light helpful. 0-60 for a '96 2.2? Maybe 10-ish? My '99 2.5 is around mid 8's, but I don't make a habit of checking it, given current gas prices . Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.