More knock sensor madness

[screwbaru2]

I just got done putting a EJ22 into my 1997 OB that had a EJ25 originally in fact it had 2:mad: . So far the only issue I have is a hard P0325 CEL code.

I've been doing alot of reading about konck sensors and I still have a couple of ?'s and what seems to me a conflict of information. If there's anywhere I may get these answered / resolved its here.

Here's the sit rep: The OB had on CEL codes stored I checked. The donor car a 1997 lego also had no CEL issues. The OB has been sitting for over a year with a 1/4 tank of gas in it. With brings me to my 1st ?.

Can old gas loose enough octane or somehow cause enough knocking to make the knock sensor send so many signals to the ECU that the ECU thinks the knock sensor is bad and sets a code?

The Haynes manual states that there should be not continuity between the knock sensor connector and ground. Other things I've read talk about anywhere from 500K ohms > 1MK oms resitance between the knock sensor and ground. This would mean continunty no? Which is right? I have no continuity between the knock sensor and ground.

Here's some additional information. The knock senor I have in the car is from the EJ25 that was in the OB. It was good, the car was however overheated. Again not codes were set in the OB. The wiring harness was from the donor car and was good.

If anyone can help me I'd be really grateful. I have been:banghead: over and over with this thing. I'm going to need to go for emissions next week after I register it so I need to resolve this.

[idosubaru]

drive it, run it or drain it and run it with fresh gas. there's no point speculating about what bad gas can or can not do on a 10+ year old car. my guess is highly unlikey this would do it.

 

as for the sensor, just clear the code and see if it comes back. if it comes back a couple times, then replace the sensor.

 

did you install it at the proper angle and torque? the knock sensor is intended to be installed at a certain angle (the pigtail - pictures and info will come up with a search of USMB) and a particular torque.

 

did you check the bottom of the sensor for any cracks when you removed it?

 

knock sensors are essentially replacement items, rarely lasting the life of the engine. not that i want to see you throw parts at it or anything. maybe you have another motor to grab a knock sensor off of?

[porcupine73]

Can old gas loose enough octane or somehow cause enough knocking to make the knock sensor send so many signals to the ECU that the ECU thinks the knock sensor is bad and sets a code?

No, the code is purely an electrical issue. It never means the ECU thinks it is detecting knock. However yes gasoline oxidizes and gets stale over time esp in hot weather so generally if gasoline is not going to be used within a couple months using a stabilizer is a good idea.

 

The Haynes manual states that there should be not continuity between the knock sensor connector and ground. Other things I've read talk about anywhere from 500K ohms > 1MK oms resitance between the knock sensor and ground. This would mean continunty no? Which is right? I have no continuity between the knock sensor and ground.

Generally when it says 'continuity' it means no or very low resisatnce like a few ohms.

 

Dratted cracked knock sensor

 

pic Gary mentioned:

[OB99W]

[...]The Haynes manual states that there should be not continuity between the knock sensor connector and ground. Other things I've read talk about anywhere from 500K ohms > 1MK oms resitance between the knock sensor and ground. This would mean continunty no? Which is right? I have no continuity between the knock sensor and ground.[...]

The FSM indicates the sensor should have a resistance between 400 kohms and 700 kohms. If you're not seeing that between the sensor's terminal and the block, that's likely the problem.

 

People have used a 560 kohm resistor to "replace" the knock sensor when they thought it was overly sensitive, something I'd never suggest doing. However, for the purpose of troubleshooting, you could temporarily replace the sensor with such a resistor (connected from where the sensor usually plugs into the harness, to a convenient ground), and the code should not recur after being cleared if the sensor was at fault. If the code returns, there are wiring/connection problems -- or the ECU itself is faulty (possible, but unlikely).

[porcupine73]

hehe yep out with the old knock sensor, in with the good resistor. I meant to use 560k but accidentally used a 470k but it seems to work ok.

 

[screwbaru2]

No, the code is purely an electrical issue. It never means the ECU thinks it is detecting knock.

Porcupine reread my ?. I asked if the knock sensor can send so many signals to the ECU that the ECU thinks the knock sensor is bad and sets a code?

 

Generally when it says 'continuity' it means no or very low resisatnce like a few ohms.

 

I know so how can there be 500K > 1MK ohms resistance? As some information says.

 

Also no cracks, angle is good torque is right. Cleaned both the sensor and block suraface to shiny metal before installing.

 

I've cleared the code many times it comes right back which is why I called it a hard code. Takes about 10 seconds for the CEL to light after starting.

 

I asked about the gas because if it could cause this I'd stop trying to debug this thing.

 

I've already good almost $2000.00 into this project. I need to put 2 axles into it yet I'm not keen I ordering a sensor only to find out the problem was in the tank. After it gets dark (expired emissions sticker) I'll go into town and put $20.00 worth of premium in it and drive around, clear the code and drive it some more. Thanks all.

[porcupine73]

Maybe in your case a resistor would be something to use for troubleshooting to see if the code stays off then. Maybe there is an issue in the upstream wiring for the knock sensor.

 

Generally when it says 'continuity' it means no or very low resisatnce like a few ohms.

 

I know so how can there be 500K > 1MK ohms resistance? As some information says.

 

I thought you were saying the haynes manual says there should not be continuity between the knock sensor and ground. 500 kiloohms resistance is not continuity. Continuity would usually be may a couple ohms or less, like when measuring the resistance of a piece of wire.

[hankosolder2]

No, the code is purely an electrical issue. It never means the ECU thinks it is detecting knock.

Porcupine reread my ?. I asked if the knock sensor can send so many signals to the ECU that the ECU thinks the knock sensor is bad and sets a code?

 

Generally when it says 'continuity' it means no or very low resisatnce like a few ohms.

 

I know so how can there be 500K > 1MK ohms resistance? As some information says.

 

Also no cracks, angle is good torque is right. Cleaned both the sensor and block suraface to shiny metal before installing.

 

I've cleared the code many times it comes right back which is why I called it a hard code. Takes about 10 seconds for the CEL to light after starting.

.

 

 

To restate porcupines point- "continuity" is generally considered to be a LOW resistance. A resistance in the 500Kohm to 1m ohm range (relatively high) would not be called continuity! The resistance value of the sensor should be in that range (assuming your multimeter is working correctly) if it is not, the sensor is OPEN and needs to be replaced.

 

Nathan

[hankosolder2]

Ah, I really restated your point. Looks like we posted almost simultaneously. Sorry!

[GeneralDisorder]

It was good, the car was however overheated. Again not codes were set in the OB. The wiring harness was from the donor car and was good.

 

I would replace the sensor. When I installed the engine in my 91 SS, the replacement had been sitting in a garage for several years. The knock sensor died all by itself - just from sitting. New sensor cured the problem.

 

GD

[four-fleet-feet]

Most dealers look up knock sensors by VIN. Since you've done an engine swap, maybe you were given the wrong sensor.

 

Subaru uses many knock sensors, they all look alike, but the wrong one will set codes until doomsday or until you pull it and replace it. If you know the VIN for the engine donor car, use that to get your new sensor. Or just mention the new engine and the donor car's year, and cross your fingers. My 97 Imp had at least 7 sensors used in its model year, or so I was told.

 

Also, if you don't leave that battery undone for at least ten minutes it'll just repeat the code. I know the manual says one minute to clear the code. Don't believe it. If your sensor is correctly oriented, torqued, and clipped into the harness correctly, ten seconds to CEL after startup seems like a repeat code to me.

 

Of course, if you pulled one off an older donor you didn't own from day one, perhaps the older engine didn't have an OEM sensor on it. That might be the problem, as well.

 

Hope this helps!

[GeneralDisorder]

Most dealers look up knock sensors by VIN. Since you've done an engine swap, maybe you were given the wrong sensor.

 

That doesn't appear to be the case:

 

The knock senor I have in the car is from the EJ25 that was in the OB. It was good, the car was however overheated. Again not codes were set in the OB.

 

Looking at the parts listing on www.subarupartsforyou.com it shows the same sensor for both EJ22 and EJ25 legacys (including Outback's) from 97 through 99.

 

I'm betting it's just a bad sensor. Get a new one as they aren't that expensive.

 

GD

[screwbaru2]

Most dealers look up knock sensors by VIN. Since you've done an engine swap, maybe you were given the wrong sensor.

 

Subaru uses many knock sensors, they all look alike, but the wrong one will set codes until doomsday or until you pull it and replace it. If you know the VIN for the engine donor car, use that to get your new sensor. Or just mention the new engine and the donor car's year, and cross your fingers. My 97 Imp had at least 7 sensors used in its model year, or so I was told.

 

Also, if you don't leave that battery undone for at least ten minutes it'll just repeat the code. I know the manual says one minute to clear the code. Don't believe it. If your sensor is correctly oriented, torqued, and clipped into the harness correctly, ten seconds to CEL after startup seems like a repeat code to me.

 

Of course, if you pulled one off an older donor you didn't own from day one, perhaps the older engine didn't have an OEM sensor on it. That might be the problem, as well.

 

Hope this helps!

 

Thanks for the info. I use a SnapOn MT2500 scanner so I don't think battery power is a problem as the battety can't be disconnected to use it to clear codes. I've cleared codes many times with this scanner after I fixed the fault and the code(s) didn't come back. You're right about the different sensors.

The 2 96's I have are both single pole sensors, 1 is an OB the other a Imp.

The 97 OB and 99 Imp are both double pole. I only see 1 wire. What do they mean by single and double pole when only one side of the double connector has a wire?:-\

 

The sun is going down pretty soon it'll be time for my outlaw ride into town to put fresh gas in that problem car.

[screwbaru2]

Well I drove around with some fresh gas for a while last night. Motor sounds and runs great. Cleared the code started the car and the mother%^&* ,son-of-a- @#%&* code came right back. So it's on to subarupartsforyou.com to order 2 knock sensors one for the 97 OB and one for the 99 Imp, that's been throwing this same code for a week now. I must have some kinda enegy field over my house destorying my knock sensors. What's the name of the constllation subaru is in? Does it require periodic sacrfices? What or who should I sacrfice?

[porcupine73]

haha whatever you have that is destroying the knock sensors is destroying them everywhere it is not at all an uncommon problem. The only sacrifice required is money to purchase new parts muhahaha. I think the constellation is pilladaes or something like that isn't it?

[screwbaru2]

haha whatever you have that is destroying the knock sensors is destroying them everywhere it is not at all an uncommon problem. The only sacrifice required is money to purchase new parts muhahaha. I think the constellation is pilladaes or something like that isn't it?

 

Porc going on 11 years of running nothing but soobs. This is the 1st year I've gotten a knock sensor code. I guess I'm getting my share all at once!

[porcupine73]

haha..hm maybe you did do something to tick off pilladaes

[johnceggleston]

haha..hm maybe you did do something to tick off pilladaes

 

also known as subaru in japan.

 

http://en.wikipedia.org/wiki/Pleiades

[Steves72]

Rather than hijacking the existing thread I thought I would start a new one. I have an 2000 Legacy OBW with the P0325 code. Today I pulled the part out thinking that I would 1) learn how to do this job and 2) try RTV on the cracks to see if I could do a simple fix.

 

Well the part came out easy enough but it appears to be in perfect condition. No cracks in the plastic. So I tried to measure the resistance from the two different plug ends on the sensor to the mating surface. On one plug I got a reading of 1.4m ohms. On the other plug I got 3 ohms - essential open. The short pigtail wiring harness has two wires on the end that connects to the knock sensor but only one on the white plug at the other end.

 

So is the open reading on the one terminal normal? And, is the 1.4m ohm reading on the other terminal too high?

 

Update to this thread: I plugged the pigtail harness back into the knock sensor and measured the resistance from the single pole inside the white connector to the ground surface of the knock sensor. The reading was 1.4m ohms. So the open side of the knock sensor is not used and based on other threads to substitute a 560k ohm resistor for testing purposes I am going to assume that the 1.4m ohm reading is now considered out of spec.

 

Am i doing the correct thing to order a new part? I think I am.

 

Steve

[screwbaru2]

You could have used my thread, I've got 2 others you can check out. The answer to your question is covered in one of them. You should have 500K to 700K as I understand it. 1.4M would be too much

[screwbaru2]

also known as subaru in japan.

 

http://en.wikipedia.org/wiki/Pleiades

Thank you. So much knowledge here!

[Steves72]

I did and it answered a lot of my questions. The only thing that was tossing me was the two different resistance readings on each of the two electrical connections coming off the knock sensor. When I put the wiring harness back on to the unit and I got the exact same 1.4m ohm reading I knew that the part was probably out of spec. Your 500k to 700k confirms that assumption.

 

Steve

 

You could have used my thread, I've got 2 others you can check out. The answer to your question is covered in one of them. You should have 500K to 700K as I understand it. 1.4M would be too much

[Steves72]

Just a quick followup. I replaced the out of spec knock sensor with one I found in stock at Advanced Auto today. I had it apart and decided to call around and see if anyone had one. The new unit knock sensor fixed my P0325 problem.

 

Steve

[screwbaru2]

Just a quick followup. I replaced the out of spec knock sensor with one I found in stock at Advanced Auto today. I had it apart and decided to call around and see if anyone had one. The new unit knock sensor fixed my P0325 problem.

 

Steve

That's good to hear! I'll be ordering 2 today.

[OB99W]

You could have used my thread, I've got 2 others you can check out. The answer to your question is covered in one of them. You should have 500K to 700K as I understand it. 1.4M would be too much

For the models we've been discussing, 400 to 700 kohms is what the FSM says are the limits. That's why a 560k resistor (standard value) can be used in a test, since it's about midway in the acceptable range for the sensor.

 

By the way, the reason the knock sensor wiring may seem "strange" is due to the nature of the signal it sends to the ECU. The output produced is a small fraction of a volt, and can be interfered with unless precautions are taken. That's why the ground (shield) is not connected at the sensor end (minimizing stray electrical signal pickup by eliminating a "ground loop"), and why the positioning of the lead is important.