92 Lumina 3.1L knock sensor

Reply

  #1  
Old 10-13-02, 08:09 PM
Gomer-techimo
Visiting Guest
Posts: n/a
92 Lumina 3.1L knock sensor

First post here... lets see what you can do =)
Found this site through a link @ www.techimo.com (great tech site).

92 Lumina sedan
3.1L
90K miles
well maintained (plugs, wires, fuel filter last within last 10K)

I started having an intermittent check engine (CE) light about 3 months ago. It would come on at random times. It didn't matter if the engine was cold or warm or how long I had been driving. It would typically stay on for about 3-4 minutes. Sometimes longer, sometimes shorter. One odd thing about it was it would never come on during interstate driving. I could drive for 3 hours and never see it, even if I had to slow and accelerate. CE light was always accompanied by a noted loss of power.

I pulled the codes from the computer and it showed a code 43 (ESC/Knock sensor). It took me a week, but I finally found the damn thing (rear of block below exhaust manifold, very hard to get at). It is very difficult to get at so I have yet to replace it or do any multimeter testing. One thing I noted is that the wire is VERY tight. There isn't any slack in it at all, it is tight like a guitar string. There are also a couple of nicks in the insulation near the base of the connector.

After wiggling the connector a bit the CE light went away for a few weeks. The engine was still a little doggy tho. Lost some zip even with the light off, but not as bad as when the light comes on (43 causes full timing retard and the lack of power). About 3 weeks after I fooled with it the CE light started coming on again, seemed a little more frequent too.

I broke down and took it to the dealership since it was intermittent and some mentioned 43 could indicate an ECM issue. I wrote a letter for the tech spelling things out like I had here but apparently he did not see it. He told me he found a code 43 as well... took it for a test drive and the knock sensor was showing an above average amount of noise (CE light did not come on). He tried to attribute this to internal noise, but seemed like he was just trying to get the car out of his bay. I know what engine noise sounds like and this engine is pretty tight(no noise). This tech was also not aware this was an intermittent deal.

When I told him of the intermittent nature of the problem he said that the CE light for the knock sensor should only come on at start, not while driving. He said that if the CE was coming on while driving it'd be something else and there should be a code stored (I have cleared the codes several times, all that shows is 43). I don't think that this is right. Shouldn't the CE/code 43 register if the ECM loses the connection to the knock sensor?

So after a trip to the dealership (luckily no charge), I am no better off than I was and maybe a bit more confused. I am leaning towards the connector being faulty.

So here are the questions I would like answered:

1. Would a bad connection cause a "high noise" signal to be sent to the ECM???

I know the bad connection would cause a loss of signal and the CE/code 43. My understanding of the knock sensor generates an AC voltage. The ECM sends 5 volts to the sensor and measures the difference when the AC voltage is generated.

2. If I replace the connector, is there anything special I need to do when replacing it?

Is it just a straight cut, solder, heatshrink deal? Or do special steps need to be taken?

3. Does the faulty connection diagnosis sound right? Would that generate a "high noise" signal? Would that set the CE light at random times while driving?

4. What is the best way to get at this beast?

Not very easy to get at from the bottom without a hoist. I tried geting at it by pulling the dogbone and rotating the engine forward (like to change the plugs). That was soso, but I doubt I could replace the connector from there. It sits low on the block under the exhaust manifold.

5. Any other thoughts???



One final thing... engine is leaking some oil from the top half of the engine. Not enough to make a spot in the drive, but enough to wet the block. I think it is the intake on the front bank and maybe the head on the rear bank (head is not leaking/mixing coolant). That is problem that will be remedied once I whip this driveabilty. This care used to sit for extended periods of time so I think the seals dried out from time to time.

This is long yes. I have tried to be very thorough in my explanation of things. I have a pretty fair base of automotive knowledge (above average IMO). So feel free to talk tech with me. Hopefull you can shed more light than the dealership did.
(BTW, I checked out the Autolibrary.org link. Very nice but no help on this subject.)

Gomer
 
Sponsored Links
  #2  
Old 10-13-02, 08:55 PM
Member
Join Date: Sep 2002
Location: Colorado
Posts: 431
I'm not sure if your engine has an ESC module so lets assume that it doesnt. The first thing I would do is to make sure that the electrical connection is tight and clean. Definitiely can get a voltage drop here with an increase of resisitence. Second if the wire to the sensor is too tight than it might be possible that when the engine shifts in the motor mounts during accel/decel you mike cause the circuit to make and break. Next things are a real knock in the engine(does it run smooth with no miss?), bad sensor, bad wiring or ECM. If you replace the sensor make sure you get the exact replacement as there are a couple of sensors with different values. Torque it properly per specs. Below is a service tip from Standard Motor Products on Code 43.



GM VEHICLES CODE 43-
ELECTRONIC SPARK
CONTROL SYSTEM
This tip concerns itself with those
systems that do not use an external spark
control module. Unlike the ESC systems
that use an external module, this system
can set a code 43 immediately after starting
the engine.
In order for the ECM to detect engine
detonation, it must be connected to a knock
sensor. In order to verify knock sensor
circuit integrity, the ECM supplies a 5 volt
reference signal. The knock sensor has an
internal resistance range of 3.8K ohms to
4.8K ohms. This resistance will lower the
5 volt reference signal to approximately 2.5
volts. If the DC voltage drops lower than
1.25 volts or rises higher than 3.75 volts, a
voltage comparator in the ECM causes a
code 43 to set.
Failure mode includes: no ignition
timing retard when the engine detonates,
and less spark advance timing during light
throttle loads. It is possible to install the
incorrect sensor since they look the same. If
you install a sensor with a higher resistance,
it will change the voltage reading to be out of
specification, and cause code 43. Also, make
sure the sensor is properly torqued to
14 ft-lbs. If the sensor is not properly
tightened, its sensitivity will be affected.

When I splice in new electrical connectors I like to use heat shrink, an uninsulated butt connector, crimp it, small drop of solder and then melt the heat shrink...but thats me.
 
  #3  
Old 10-14-02, 05:51 AM
redneck
Visiting Guest
Posts: n/a
Sounds like steveo has some input on your elec. problem. As for the oil--all 3.1's do this--look directly under the throttle body. There is a plug where the distributer used to be on these motors. That is where your oil leak is at. You have to remove the throttle body and fuel lines to get at it. Remove hold down bolt and pop up the plug--there is an O ring on it--don't bother replacing O ring just silicone it real good and put back together!!
 
  #4  
Old 10-14-02, 02:25 PM
Gomer-techimo
Visiting Guest
Posts: n/a
Useful info Stevo2

In order to verify knock sensor
circuit integrity, the ECM supplies a 5 volt
reference signal. The knock sensor has an
internal resistance range of 3.8K ohms to
4.8K ohms. This resistance will lower the
5 volt reference signal to approximately 2.5
volts. If the DC voltage drops lower than
1.25 volts or rises higher than 3.75 volts, a
voltage comparator in the ECM causes a
code 43 to set.
Does the ECM only supply the 5V at startup to check the integrity of the sensor? Does it apply the 5V to check the sensor, then switch to measuring the output voltage of the sensor, or does it supply 5V continuously?

What causes a code 43 to be set while driving? My guess is that the knock sensor showing a high voltage to the ECM (abnormal noise) sets the code.

The light hardly ever comes on at startup. It usually comes on no sooner than 5-10 minutes into a trip. I highly doubt it is picking up abnormal noise from the engine. The engine is quiet and runs smooth except when the CE light comes on. If it was detecting abnormal noise from the engine you'd think it would do so under hard acceleration and it usually doesn't. It will come on while I am sitting in the drive through at idle, stay on 30-40 secs and shut back off again.

If it is the case that this wire is supplying a voltage constantly, then it indeed sounds like there is an issue in that wire/connector. If it only sends the voltage at startup and then just measures voltage from the sensor, then I have other issues.

redneck:
Thanks for the info on the oil leak... I'll look into it.
 
  #5  
Old 10-14-02, 06:51 PM
Member
Join Date: Dec 2001
Posts: 2,538
the ecm supplys the 5 volts to knock sensor anytime the key is on, and could very well be a bad connector about 10 bucks at any parts store, i would suggest you replace it and also reroute or lengthen the wire so it isnt stretched to tight.
 
  #6  
Old 10-14-02, 07:15 PM
Member
Join Date: Sep 2002
Location: Colorado
Posts: 431
reference voltage: In computerized engine-management systems, a five-volt signal sent out from the electronic control unit to a variable-resistance sensor such as a throttle-position sensor so the computer can read the voltage value of the return signal.

5 volts is a constant value sent by the ECM. The knock sensor is a variable resistor that knocks the voltage down, its a resistive ground point for the voltage. The ECU is expecting to see a return or actual voltage between 1.25 volts and 3.75 volts. Any voltage out of that paramater will set your Code 43. The knock sensor performs its task just like the sensors for your oil GAUGE and temperature GAUGE...i'm not talking about idiot lights but GAUGES. These GAUGE sensors are variable resistors also.

The article that I sent you says the code "CAN" be set at start up...it doesnt say it must be set. The purpose of the knock sensor is to monitor your engine at any time its running and the Code 43 can be set at any time.

Knock sensing may be initiated by the ECM at startup(open loop) or after warmup(closed loop) it just depends on the negine and the way the engineer programmed the ECM.

Looking at the Chilton's online wiring diagram(if correct) you should have only one Dark Blue wire running to the knock sensor which connects to pin A11 of the ECM.
 
  #7  
Old 10-14-02, 07:59 PM
Gomer-techimo
Visiting Guest
Posts: n/a
Thanks again Stevo2... What you have said makes sense. My confusion stems from what the tech at the dealership was telling me. It does indeed sound exactly like a connection issue now. I will try to take care of that this weekend. Hopefully it is mild here (MI). It is a single wire connector with one dark blue wire as well.

One thing about what you said though... is it really a variable resistor? What I have seen regarding testing it says that it generates an AC voltage in response to knock. The testing calls for the connector to be pulled and to tap the block near the sensor. A good sensor would yield a varying AC voltage.

Anyways, thanks a lot for the help. Like I said I am going to replace that connector this weekend if all goes as planned. I will post back results.
 
  #8  
Old 10-14-02, 09:08 PM
Member
Join Date: Sep 2002
Location: Colorado
Posts: 431
Gomer, I was trying to keep this simple and that is why I referred to it as a variable resistor. It is in fact a piezoresistive device that allows a current change. Tapping on the engine block and monitoring any voltage at the sensor is not a valid test. You could be measuring an ESD charge for all you know. You must measure resistence in its static state and you can monitor the voltage changes on line to the ECM in DC volts. Is it your intention to hook a scope up to the sensor? You'll have to get the waveforms from a Chevy service manual. You can also hook up a scanner while you drive to monitor voltage or you might be able to find a Flight Recorder type scanner that will log errors as they happen during a test drive. There are also different types of knock sensors...resonant and non-resonant.

Something off the web:
The knock sensor consists of an electric coil that is wound around two ceramic rods with a magnet in the center. Engine vibration makes the ceramic rods vibrate. This disturbs the coils magnetic field and alters the current passing through the coil. This disturbance returns to the computer as a signal pattern that the computer analyzes. The computer then determines whether the vibrations are characteristic of engine knocking.
 
  #9  
Old 10-15-02, 02:32 PM
Gomer-techimo
Visiting Guest
Posts: n/a
Stevo2:
That was my understanding of how the sensor functioned.

Here's the latest... there are no aftermarket stores that carry that part. The dealership wants $58 for the knock sensor pigtail. $58!!!!!!! Goodness! Any idea why the price is so steep? I can get the sensor alone for $40.

Should I still pursue this pigtail deal? I would love to be able to hook into the ECM and witness what is occuring. That is not going to happen though. I am lacking the tool and they are not cheap either. Unfortunately I am displaced from my hometown. I worked at a garage there for 2 years and could probably get taken care of there... here I am unsure of where to take it to. I thought the dealership would handle it, but they seemed to really care less. I am also trying to remedy this at the least amount of expense.
 
  #10  
Old 10-15-02, 03:04 PM
Member
Join Date: Sep 2002
Location: Colorado
Posts: 431
Gomer, I'm in the middle of working on my cars and I have continuous parts problems also. Maybe the reason for the high cost is that the connector pin/s are gold plated. I have seen this change on the web for some cars. Let me look a round a bit. I'll get back to you later.
 
  #11  
Old 10-15-02, 06:22 PM
Joe_F
Visiting Guest
Posts: n/a
Check any jobber that sells AC Delco or any good aftermarket parts counter with a good buyer's guide. By thumbing through, they can find you something.

The sensor can be pulled off the shelf by the parts store and you guys can thumb through the buyer's guides and find a replacement. Bound to be out there.
 
  #12  
Old 10-15-02, 09:14 PM
Member
Join Date: Sep 2002
Location: Colorado
Posts: 431
Joe has a good point. Before thinking about the connector replacement try this.

1. Pull the connector off and flush it out with a good electronic contact cleaner. Also flush the end of the knock sensor.

2. Relieve the stress on the wiring and harness at that area so that is has slack to move.

Then see what happens.

Find out what the GM part number is for that connector/pigtail and post it. I'm still curious about availabilty/price and would like to check some other sources.
 
  #13  
Old 10-15-02, 09:16 PM
Joe_F
Visiting Guest
Posts: n/a
Simply e-mail AC Delco and ask .

Their pigtails start with PT in the part #s.
 
  #14  
Old 10-15-02, 10:54 PM
Member
Join Date: Sep 2002
Location: Colorado
Posts: 431
With all due respect Joe, I Emailed AC Delco once and never will again. They told me to contact a local auto parts dealer who already couldnt answer the application question. I dont care to go around in circles. Complacency is everywhere! If you want something done you must do it yourself.
 
  #15  
Old 10-16-02, 04:00 AM
Joe_F
Visiting Guest
Posts: n/a
You didn't ask them properly .

I simply send those e-mails back with a, "That's not the answer I need reply". Or, "What are we going to do about this situation???"

It does filter down to the right person, or gets into a supervisor's hands .
 
  #16  
Old 10-16-02, 06:26 AM
Joe_F
Visiting Guest
Posts: n/a
Thumbs up

The connector required is a GM #12126456, or an AC Delco PT727 if you wind up needing it.

I found it in a jiffy. Used on a ton of GM vehicles. Can't see why the dealer couldn't find it. Right there in group 2.383 of the parts catalog for this Lumina
 
  #17  
Old 10-16-02, 11:16 PM
knuckles
Visiting Guest
Posts: n/a
gmpartsdirect.com has that pigtail for $29.27 plus shipping.
 
  #18  
Old 10-17-02, 01:15 AM
Member
Join Date: Sep 2002
Location: Colorado
Posts: 431
Now this is what I call teamwork! Knuckles...www.gmpartsdirect.com is a great place to deal with and Joe you came through shining again with that part number! I knew we could do it! Joe I'll post a recent war story about parts and maybe you'll see why I have little patience for poor Tech Support and their equally poor supervision. Regards to All-Stevo
 
  #19  
Old 10-17-02, 03:57 AM
Joe_F
Visiting Guest
Posts: n/a
This pigtail is listed by GM as "high temperature", hence the 58 dollar list price . Must be of a special material.

Stevo: Post your problem on a separate post and let's see what you have. I too, come across "trouble" companies, but my NY persistence usually wears them down .
 
  #20  
Old 10-18-02, 10:11 PM
Gomer-techimo
Visiting Guest
Posts: n/a
Thanks for all the input guys... we will beat this thing yet. Work has been rough these last few days so I haven't touched... I'll be under it tomorrow to see if I can get some slack in that wire, give it a good visual inspection etc.

I checked out that part GM #12126456 at the gmpartsdirect.com website, it listed it as a coolant sensor connector. I emailed them and asked them what the # was and they suggested an 12102621 which just shows up as "connector."

we'll see what a day of crawling around under the car yields.

Thanks again all =)
 
  #21  
Old 10-19-02, 06:41 AM
Joe_F
Visiting Guest
Posts: n/a
The number I provided is what GM lists as a knock sensor connector for that vehicle.

It is listed as "high temperature", hence the 58 dollar list price.

I believe the coolant sensor connector for your vehicle is a standard issue Weatherpack connector...no where near 58 bucks .

Double check both part #s with any GM dealer to be sure of the application and price. I think you'll find I'm right (or GM's parts catalog is wrong)

Can also go to acdelco.com and pull up an image of the PT # I provided and see if it looks like the part you need.

Do a little homework and let us know.
 
  #22  
Old 10-19-02, 08:17 AM
Gomer-techimo
Visiting Guest
Posts: n/a
haven't crawled under the car yet... but I have done some sleuthing.

Called a chevy dealer in my home town (I am home today). He was very helpful. Asked him about the knock sensor pigtail and he gave me the 12102621 part number. The less expensive ($10) part number. I asked him about this and mentioned the pt727 # and he said that was listed as well as "high temp" like you said. He said the 12102621 should work fine. And for $40 less I am going to give it a shot. I am pretty confident the problem lies in that lead, but not 100% sure. I will give the 12102621 a shot and see what happens (found it a local parts counter)
 
  #23  
Old 10-19-02, 03:24 PM
Gomer-techimo
Visiting Guest
Posts: n/a
Threw the 12102621 on as well as a new knock sensor. Soldered, heat shrinked and extended the lead. Took the car for a short (15 mile) test drive and no CE light. Lot more zip and power as well. I think it was retarding the timing almost constantly.

regarding the oil leak:
redneck
"Sounds like steveo has some input on your elec. problem. As for the oil--all 3.1's do this--look directly under the throttle body. There is a plug where the distributer used to be on these motors. That is where your oil leak is at. You have to remove the throttle body and fuel lines to get at it. Remove hold down bolt and pop up the plug--there is an O ring on it--don't bother replacing O ring just silicone it real good and put back together!!"

this is indeed where the oil has been leaking from. Is it possible to do this repair without yanking the throttle body and fuel rail? I lack the tools and perhaps the talent to go pulling stuff off the top half of the engine like. Seems like I can make too much go wrong. It looks like this plug might be accessible without pulling the throttle body etc. Anyone else with experience on this matter?
 
  #24  
Old 10-19-02, 04:45 PM
Member
Join Date: Dec 2001
Posts: 2,538
it is not just a plug it is your oil pump drive and is to long to remove with throttle body installed, however it is not all that difficult to remove throttle body, i usually install a paper distributer gasket on it as what would be on a old chevy v8 it seals up alot better than the oring it has on it, if you do decide to do the job yourself just make sure the drive is fully seated against the block sometimes they will not drop right into the oil pump when reinstalling and will stick up about 1/4 inch.
 
  #25  
Old 10-21-02, 07:50 AM
redneck
Visiting Guest
Posts: n/a
Gomer, Bejay is right, there is a shaft and gear on that plug! You must remove at a minimum the throttle body. It is a very easy job--I have removed my throttle body twice and have not changed the gasket. The hardest part of the job is actually getting the plug to "pop" out. Once you find it, you will see what you will need to remove to get to it.
 
  #26  
Old 10-21-02, 02:13 PM
Gomer-techimo
Visiting Guest
Posts: n/a
redneck: I see the plug. I was concerned with pulling the fuel rail. Pulling the throttle body doesn't scare me too much. Any tricks to popping that plug out?

As for the knock sensor/connector. I have put 250 miles on the car since saturday and it drives like a champ. A lot more power/smoother acceleration etc. It must have graudually failed over the last 20K or so, because I'd have noticed had the power dropped off suddenly.
 
  #27  
Old 10-21-02, 04:58 PM
Member
Join Date: Dec 2001
Posts: 2,538
you dont have to pull the fuel rail, that would require upper intake removal, just remove the 2 gas lines going to it and the hold down clamp on them so you can lay them out of the way,
a pair of pliers on the top of the oil pump drive usually helps to loosen it by turning it back and forth while lifting up, once you have enough room to get under it with a prybar it should pop right out.
 
  #28  
Old 10-23-02, 06:50 AM
redneck
Visiting Guest
Posts: n/a
Again Bejay hit it--you may have to work at it awhile from different angles and trying different tools--but you will get it to "pop" out!
 
  #29  
Old 10-24-02, 12:20 AM
knuckles
Visiting Guest
Posts: n/a
Thumbs up

OTC sells a removal tool for this job, but I've yet to need it. You can buy the tool from www.ontool.com if yours just won't budge.
 
Reply

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Thread Tools
Search this Thread
Display Modes