Any experience with a failure of a knock sensor?
 

I was driving up a long grade with the cruise control engaged; about 3,100 rpm in fifth gear, and the Check Engine light came on. I checked the engine error codes and it came up with "1132", which is the number 2 knock sensor. The car runs normally with no loss of power and the distributor drive belt is OK and properly timed. I reset the error code and it came back once in several days since the first experience.

Has anyone experienced a failure with a knock sensor?

I guess these are like any other sensor, they will fail at some point, but is the failure mode?

Thanks for any help...

Keith

What year and mileage of your car? What fuel are you using? Also how do you reset your engine codes (the only known way is if you have the Bosch Hammer or the dealer did this for you). Disconnecting the battery will not reset fault codes. Also if the best gas you can find is not high enough octane rating then the knocking will be repeated under those conditions you mentioned sensing that happening. I don't know how robust knock sensors are, have not heard of them failing.

The car is a 1991 C2 Targa 3.6L US destination and resides in Southern California and has 104K miles. The engine codes can be checked and reset with the ignition switch and the throttle pedal. Ignition on, pedal to WOT for 5 seconds, back to idle stop, one very long flash of the “CHECK ENGINE” light and then the codes begin. To reset or clear the existing codes, hold throttle down for 1 minute and it will clear the codes.

In my case the code “1132” which is the #2 Knock Sensor, appears now and then. It normally shows up after a quick WOT/near Redline shift from 1st to 2nd gear. Before it was happening under low rpm and climbing a hill in temp-o-matic (Cruise Control) engaged.

Here in California, we can normally only get 91 Octane. I have had this car for over three years and the “CHECK ENGINE” light started in with the “1132” in the past few months. So, it isn’t a bad load of fuel or that the 91 Octane is insufficient in number.

If I leave the code in, and the “CHECK ENGINE” illuminated for a few miles, the “CHECK ENGINE” goes out.

I question when does a “Knock Sensor” or “CHECK ENGINE 1132” come up? I don’t think it is because the sensor detected a knock. I believe it is because the sensor is reporting an anomalous condition with the sensor itself.

I read some interesting posts about a loss of signal or corroded contacts at the harness junction that could cause this problem. I may tackle it this weekend.

Has anyone heard of a failure of the sensor itself?

Thanks, Keith

I have a 1993 RSA which is the same DME OBD1 system controller that is in the 1989-1994 964 series cars. I was told by the technicians and experts (PCA Tech committee Joel Reiser and Bruce Andersen) that the only way clear fault codes (check engine light fault announcements) accurately is to use the Bosch Hammer (KTS300) or AKA Porsche Factory tool 9288. The method of the ignition on, accelerator pedal is only an emergency way of reading the codes (for which you need a translator table for various possible faults anyway). The codes are then stored in the memory accumulator which cannot be cleared without factory service-just so not anyone can mess with the car to the point of no return! I would venture to guess at this point that your codes are still there because they were never cleared. As to your question of whether knock sensors ever fail we'll have to dissect how they work and their weaknesses of design if they have them. I think there is a website on how gadgets work (like oxygen sensors). I'll have to look for knock sensors and how they work. In principle I'd venture that there is a crystal which vibrates and sends out an electrical pulse (piezoelectric?) so this crystal in the sensor can detect knock and send the message to the CPU for dialing back the ignition timing.

I found some basics to start. The Knock Sensors from Bosch are Piezo ceramic sensor elements.

From cartoday.com:

Knock sensing
This is often seen as the glamour part of ECU control, because it has relieved the engine designer of the biggest stumbling block in petrol engine design, namely the need for conservatism when selecting compression ratios, ignition timing, or even lean mixtures. These are some of the factors that can cause engine knock, a condition where the last part of the mixture, the end-gas, ignites spontaneously, before the flame-front reaches it. This speeds up the combustion so that it does not proceed at normal speeds of around 30 metres/sec, but can reach speeds of up to 2 000 metres/ sec. The resulting pressure increase causes damage to pistons, gaskets and valves, and in addition, the shock wave creates a noise.
The knock sensor uses a piezo-electric element, ie one that emits a voltage when if suffers a change of shape. This enables it to respond to the shock waves in the cylinder block resulting from the engine knock, and the resulting signal is used by the CPU to retard the ignition timing for a number of power strokes, only to advance it again if the knock does not recur. The placement position of the sensor on the engine block is critical, because it has to be able to respond to knock in each of the cylinders, especially at high engine load and high-speed conditions, where an inaudible form of knock often occurs. As a rule, four-cylinder engines have one sensor, whereas bigger engines have at least two sensors.
Link: http://www.cartoday.com/content/technical/jake_venter/2001/electronic_control_5.asp

From Bosch:Knock Sensors Piezo ceramic sensor element
Standard and custom design for connector and pigtail version Non-resonant type
Knock control for: Lower emissions Better fuel economy
More power
Link: http://www.boschusa.com/AutoOrigEquip/Sensors/KnockTempSensors

After all that, I understand on a basic level how they work, but the application of how the OBDI works in conjunction is the real question. When a “CHECK ENGINE” light is illuminated, what caused it to set a fault? It is not because the engine knocked. When the engine knocks the motronic engine management retards the ignition timing. The fault is a loss of signal or voltage. Either a high restance connection, or a open sensor, or a fractured sensor. I wonder what the static test values the sensor are at rest?

The OBD1 must be able to tell that the sensor is connected some how...

Keith

My suggestion is take it to someone (TRE cup in LA) or the dealer who can read out and reset the codes for you. Then you can decide if the problems need pulling the motor to get at the deeply buried knock sensors to see if they are really bad.

Pull the motor? I can see these by removing the duct that runs from the fan shroud to the heater fan. If I can’t reach them by this access, I’ll pull the fan/alternator and for sure get to them with my elbow bent backward. I don’t believe I would need to pull the engine.

I have a theory: I think there are two residences for faults. The DME and the Central Electronics. The faults that are engine related to the DME only reports that a fault exists to the Central Electronics and the check engine light is triggered, while the air-bag and other HVAC faults are kept in the Central Electronic.

I am going to a Dealer tomorrow to have the codes reset for a problem I induced when I was resoldering the clock terminals. I will reset the engine codes using the throttle pedal before I turn it over to the service folks and ask them specifically what codes were set, and then to reset all to null.

One other suggestion is to buy a tankful of racing gas or aviation fuel from the tracks nearby and see if knocking occurs by driving in high gear uphill (classical way of inducing knock). The check engine light should not come on since you've taken the fuel quality (octane #) factor out of the equation.

Your being able to access the sensors by removing the fan/alternator is a good point. I didn't look at the workshop manual for where they were located and just assumed that with post 1990 engines the accessibility of most repairs are becoming more like the Turbo models where nothing can be reached without pulling the engine.

Also, I recently had to replace my AFS (air flow sensor) at only 12000 miles after the car started to send out the various fault codes with the oxygen sensor, air flow meter, and possibly the knock sensor. So I replaced the oxygen sensor ($140), engine ran fine then acted up again....it was the pernicious intermittent problem! Then only after the AFS was replaced ($450) did I get a reliable running car again (cross my fingers-only had chance to put 50 miles on it before winter and salts came). Although after forcing the car with the faulty AFS to limp home during the earlier test drives (where the check engine light came on following rough running and no power with accompanying backfiring and knock no doubt!) must not have been good for the car, it had the effect of making the once intermittent problem become a continuous one which was solved with a new AFS. (visit my post to get the details).

Good luck and let us know what the problem really is. I don't think it is the knock sensors since they appear to be pretty simple and reliable.

Hey tihih,

I just confirmed my theory on the OBD1 having two fault registers or fault storage locations. I just got back from the P-Dealer. I had the errors codes reset with the Bosch Hammer. Before I went in, I reset the error codes with the ignition on and using the throttle pedal. I asked the Service Advisor to check and see if there were any other codes other than the one that I couldn’t clear. The one I couldn’t clear was the air-bag fault. I induced that error when I had the clock out and started the car and moved it to let my wife out of the driveway. I’ll never do that again. He said I only had the air-bag fault code. Nothing else.

The procedure for clearing the error codes, “CHECK ENGINE” light, is as follows:

1. Ignition Switch to “ON” position (Do Not Crank the Engine)
2. Throttle Pedal to Wide Open Throttle “WOT”
3. Release after about five seconds.
4. Read Codes: You will see a long flash of the “CHECK ENGINE” light indicating the beginning of the sequence, and then the codes will blink. An error code will have four sets of blinks with a pause in between each set. Long Flash (meaning the code is about to start) then the blink count, then a pause, then the second number in the set and so on. For my #2 knock sensor, the code is “1,1,3,2”, or blink, pause, blink, pause, blink- blink- blink, pause, blink- blink. I believe that you can have multiples of the same code also.

5. Depress the Throttle Pedal for three seconds and you will get the second error code if there is one.
6. Keep advancing until you run through the list of error codes until you get to “1,1” which means the end of the list.
7. To clear the error codes, depress the throttle Pedal WOT for 1 minute and then release.
8. The codes should be reset. I tried to clear the codes after reading the first code without advancing to the end and the codes would not reset. I guess the software designers wanted you to read all the codes before you dump them.
9. If you have no error codes and you perform the above check, all you will see a 1,5 meaning no error codes. There is an implied 0,0 on both the end of codes and no error code sets. (blink, pause, blink-blink-blink-blink-blink) = 1500

There is a listing in the Porsche Service Manual

First number means “DME/LH”
Second Digit means “Constant fault = 1, and 2= Loose Contact”
Third and Forth Digit mean the actual fault.

11 = Supply Voltage
12 = Idle Speed Contact Ground Short
13 = Full Load Contact
14 = Engine Temperature Sensor 2
15 = Idle Speed Contact Break
21 = Air Flow Sensor
22 = Idle Speed Control Activation
23 = Oxygen Regulation Stop
24 = Oxygen Sensor
25 = Intake Air Sensor
31 = Knock Sensor 1
32 = Knock Sensor 2
33 = Control Unit (Knock computer)
34 = Hall Signal
41 = Control Unit Faulty
43 = Tank Venting Valve
44 = Resonance Plate
45= Check Engine Warning Lamp
51-56= Injection Valve – Injection Control Unit

The Service Manual does a fair amount of troubleshooting beyond this; it instructs one to each connector and gives the values that should be present.

I conclude that any of the above faults can be cleared using the above method.

I will post after I finish my exploratory on my Knock Sensor problem.

Keith

Hi Keith,
Thanks for the explanations and it makes good sense now looking back that I did get codes 1124 and 1121 as well as 1115 (or really 0110, 0150) which I misinterpreted to be idle speed contact break. Anyway I broke the bank and got the factory tool to get the damn codes in plain text and have faith in my diagnosis of these electrical problems.

Tom

Hi Tom,

I checked out my knock sensor #2 on the right bank. I removed the duct between the fan shroud and the heater blower and was able to get a 13mm socket on the sensor attachment bolt with a single universal joint and a 16" x 1/4 drive extension. I’m very sure that I could change the #2 sensor with little difficulty. The cable comes right out in that same area and is very accessible. When I traced the cable out to the harness connector, it was right there above the fan. The connector is bright green. When I disconnected it from its mate, I found that the #1 terminal of three and was corroded. I cleaned up both halves with little effort and then rinsed. Time will tell, but I think I nailed it...

The #1 sensor is a little more involved. If only they had manufactured the cooling shroud that covers the whole top of the engine with an access hole to loosen and tighten the attachment bolt, I could change it as well. If I had to change #1 sensor, I would pull the Fan/Alternator to get at it.

Keith

Good job Keith,
If there's corrosion then there's reason to believe the voltage readings would be off and the ECU would interpret a problem with the sensor and dial back the ignition as a precaution.

Tom

the plug for knock sensor 2 is right under the engine lid grille, so it can get wet and corroded. the other one is way back by #3 cylinder, out of the rain. i've cleaned those plugs with good results.