SOHC V6 knock sensor replacement

[2000StreetRod]

knock sensor retard w 87 octane

Since my last knock sensor test I've replaced the entire exhaust system from manifold outlets to tailpipe end, the serpentine belt and the belt tensioner pulley. The graph below is from a brief datalog I performed in the garage with the transmission in Park and using 87 octane fuel with up to 10% of ethanol.

Bold green = engine rpm
Orange = knock sensor retard
Blue = ignition advance
Purple = A/F ratio
As the engine speed rapidly increased from 1,000 rpm the knock sensor retard dropped suddenly at 4,000 rpm from 0 degrees to -6 degrees and stayed there until the throttle was released at 5,400 rpm. As a result of the knock sensor retard the ignition advance dropped from 23 degrees to 18.5 degrees. It should be noted that the A/F ratio was 14.95 at 1,000 rpm when the throttle was mashed. It decreased to 13.75 at 3,000 rpm and then started rapidly increasing to a maximum of 15.97 at 3,200 rpm. That's too lean which corresponds with the knock sensor retard.

From the above I suspect that the richness increase is not able to keep up with the engine speed increase and there actually may be valid detonation detection. It will be interesting to see if I can duplicate the condition after I install a high flow fuel pump and a new fuel rail. I originally observed the knock sensor retard problem when performing uphill WOT runs. I will review those datalogs again but as I recall the A/F ratio was between 12:1 and 13:1 which should be rich enough to prevent detonation.

 

[2000StreetRod]

no retard improvement

I did a brief "hill climb" datalog this morning running the remaining 87 octane with up to 10% ethanol gas in the tank. After replacing the exhaust system, serpentine belt, tensioner and removing the A/C refrigerant manifold there is no improvement in the knock sensor retard.

orange = Air/Fuel ratio
bold yellow = knock sensor retard
green = spark advance
blue = throttle position
purple = rpm

At 2860 rpm the knock sensor retard drops to -2.00 degrees and continues dropping until reaching -7.75 degrees at 3290 rpm with corresponding decreases in total spark advance and power. After 1.5 seconds the retard starts increasing in 0.25 steps until the throttle is released.

bold orange = Air/Fuel ratio
blue = throttle positon
purple = rpm
The A/F ratio plot indicates the only potential lean condition is during the first second of the pull. The max inlet air temperature was only 86 degrees F so if there is actual detonation it is either due to the low grade fuel or the "tip" characteristics. I purposely took about 0.5 seconds to go to WOT in an attempt to minimize "tip" issues. After the pull I drove to the nearest BP station to fill up with premium fuel. I was disappointed to learn that they no longer sell 100% gas fuel. The BP Ultimate 93 octane fuel has up to 10% ethanol and the attendant said that the nearest station with 100% is a competitor about 10 miles away so I purchased a tank of the Ultimate. I'll drive with that for a few days and then perform another "hill climb" datalog.

While driving my normal routes I did not notice any loss of low engine speed torque due to the new high flow exhaust system. The sound level from the exhaust is very sensitive to increasing throttle position (engine load). Driving for quiet exhaust should result in best fuel economy but it is very tempting to mash the throttle for an adrenaline rush.

 

[2000StreetRod]

small improvement with premium fuel

I did a brief "hill climb" datalog this morning running BP Ultimate 93 octane with up to 10% ethanol gas in the tank. There was a small reduction in the maximum knock sensor retard (-6.5 vs -7.75) and it happened at a higher engine speed (3720 vs 3290 rpm). Interestingly, the knock sensor retard advanced from 0 as I opened the throttle plate in .25 degree steps and continued advancing to +1.5 degrees at 3400 rpm during WOT. The rapid drop did not begin until 3530 rpm in 2.0 degree steps.

red = open loop flag
bold yellow = knock sensor retard
green = spark advance
blue = throttle position
purple = rpm
The PCM went to open loop when the load reached 80%.

red = open loop flag
yellow = knock sensor retard
bold orange = Air/Fuel ratio
blue = throttle positon
purple = rpm

 

[2000StreetRod]

Desensitized knock sensor tune better

I had a chance to perform another hill pull this morning before it started to rain. Last night I loaded the desensitized knock sensor tune and there is still 3/4 tank of 93 octane fuel. The knock sensor advance prior to the pull (+0.5 deg) was slightly more than with the stock sensitivity knock sensor tune (+0.25 deg). Unlike the stock sensitivity which increased knock sensor advance for the first part of the tune this time the knock sensor advance almost immediately dropped to zero, briefly increased one increment, and then dropped back to zero and stayed there. The max load for both pulls was 86%.

red = open loop flag
bold yellow = knock sensor retard
green = spark advance
blue = throttle position
purple = rpm
The PCM went to open loop when the load reached 82%

red = open loop flag
yellow = knock sensor retard
bold orange = Air/Fuel ratio
blue = throttle positon
purple = rpm

bold green = spark advance
blue = throttle positon
purple = rpm
The spark advance was greater than 14.0 degrees at the beginning of WOT and gradually increased with engine rpm to 17.5 degrees. As a result of the greater spark advance the time from 27.5 to 59.0 mph decreased by 0.28 seconds. My concern is that the knock sensor sensitivity may not be adequate to protect the engine if there is a sudden lean condition due to a malfunction such as a split intank fuel hose. After I burn the rest of this tank of fuel I'll put in a few gallons of 87 octane and do another pull to check for at least some negative knock sensor retard.

Note: the drop in engine rpm from 29.6 secs to 30.5 secs is due to torque converter ratio change. It is very noticeable with the louder exhaust system.

 

[koda2000]

2000StreetRod,

I’m not sure where you’re going with this, and I know from reading your previous posts that you’re certainly well educated in the use of these kinds of diagnostic tools.

I, by no means, claim to be and expert in this area, but knock (or ping) is caused by pre-ignition of the air/fuel mixture before the piston has reached something close to TDC. Granted, my experience has been more with carbureted engines with distributors and w/out computers and sensors, but the principles are the same. The higher the engine's compression, the greater the tendency to induce knock under load. The higher the engine’s rpm more ignition timing advance is required. For example, static timing on an idling engine might be 8 degrees BTDC, but at 3,200 rpm the ignition advance might increase to around 32 degrees BTDC.

Higher octane fuels increase octane by adding stabilizers to the fuel. These stabilizers are designed to allow for higher compression w/out pre-ignition (AKA dieseling). That’s why it’s a complete waste of money (despite what some people believe) to use a higher octane fuel than what your engine needs. Ethanol reduces fuel efficiency and causes fuel system corrosion due to it’s innate tendency to suck moisture/water from the atmosphere, but I guess ethanol is here to stay. The elimination of lead from fuels in the 80's also reduced octane. The lead was there to lubricate cast iron exhaust valve seats, but it also increased stability. The only place I've been able to find 93 octane fuel w/out ethanol is at some BP stations.

With contaminant buildup over time, engine compression can increase and/or you can develop hot-spots in the combustion chamber. If the air/fuel mixture is lean (too much air and not enough fuel) it will increase cylinder temperature which can also cause pre-ignition. Distributors used mechanical weights and vacuum advance to control ignition timing based on rpm and engine demand, based on vacuum. It worked okay, but sensors, fuel injection and computers work much better, which has led to 100,000 mile tune-up frequencies (rather than every 10,000 miles) and increased power and fuel economy while decreasing emissions.

A knock sensor literally hears/feels the knock and sends a signal to the PCM, which in turn retards ignition timing (making the spark plugs fire later, rather than sooner) to try to prevent pre-ignition. The computer can also vary the injector pulse-width to richen the A/F mixture. This is done to reduce cylinder temp. This is much better than using a carburetor and a distributor, but it’s still not perfect.

In the near future, DI (direct injection) and VVT (variable valve timing) will make engines adapt to conditions even more easily and will probably make engine management systems of the late 20th century as primitive as carbureted engines with distributors.

 

[2000StreetRod]

knock sensor related goal

. . . Higher octane fuels increase octane by adding stabilizers to the fuel. These stabilizers are designed to allow for higher compression w/out pre-ignition (AKA dieseling). That’s why it’s a complete waste of money (despite what some people believe) to use a higher octane fuel than what your engine needs. . .

I think everything in your post is valid except for the above in some cases. Many electronic control units periodically test the fuel quality by advancing the ignition timing until knock is detected. The test results are used to optimize the timing for the fuel in the vehicle - usually a compromise between fuel economy and performance. During more than one of my custom tune drives (not WOT) I observed a knock sensor retard of as much as +2.0 degrees running high octane fuel. If you'll note during the 93 octane hill climb the retard actually advanced during the first 6.5 seconds of WOT to take advantage of the higher quality fuel. Many manufacturers specify lower octane requirements to sell their vehicles knowing that their smart ECUs will take advantage of the capabilities of whatever fuel is purchased. With those systems higher octane results in better fuel economy and performance.

My knock sensor related goal is to get the best performance possible without endangering the engine. The specified fuel for my 9.7:1 compression ratio SOHC V6 is 87 octane. It is extremely unlikely with an IAT of 84 degrees F and 93 octane fuel that detonation is present especially when it wasn't detected for the first 6.5 seconds of the WOT pull. I will research the source code to try to determine if it the retard is the result of some other parameter besides knock.

 

[koda2000]

I think everything in your post is valid except for the above in some cases. Many electronic control units periodically test the fuel quality by advancing the ignition timing until knock is detected. The test results are used to optimize the timing for the fuel in the vehicle - usually a compromise between fuel economy and performance. During more than one of my custom tune drives (not WOT) I observed a knock sensor retard of as much as +2.0 degrees running high octane fuel. If you'll note during the 93 octane hill climb the retard actually advanced during the first 6.5 seconds of WOT to take advantage of the higher quality fuel. Many manufacturers specify lower octane requirements to sell their vehicles knowing that their smart ECUs will take advantage of the capabilities of whatever fuel is purchased. With those systems higher octane results in better fuel economy and performance.

My knock sensor related goal is to get the best performance possible without endangering the engine. The specified fuel for my 9.7:1 compression ratio SOHC V6 is 87 octane. It is extremely unlikely with an IAT of 84 degrees F and 93 octane fuel that detonation is present especially when it wasn't detected for the first 6.5 seconds of the WOT pull. I will research the source code to try to determine if it the retard is the result of some other parameter besides knock.

I accept your hypothesis regarding the computer advancing timing to take advantage of higher octane fuels. It makes sense to me that the PCM is possible of doing this. I wonder if the use of more expensive grades of fuel vs economy is worth the higher price. I kinda doubt it, but if performance is your goal, it might be worth it to you. I miss the days when you could buy 101 octane at Hess stations.

 

[kevin p]

Hi, can anyone point out where the electrical connector for the knock sensor is positioned. On one of the pictures, it shows a new one with locating studs on it, so I presume it is mounted on a bracket of some kind. I want to check the outputs, but I don't want to take the inlet manifold off just to test it.
Thanks Kev.

 

[2000StreetRod]

knock sensor connector location

Hi, can anyone point out where the electrical connector for the knock sensor is positioned. On one of the pictures, it shows a new one with locating studs on it, so I presume it is mounted on a bracket of some kind. I want to check the outputs, but I don't want to take the inlet manifold off just to test it.
Thanks Kev.

On my USA 2000 Sport the connector is above the right valve cover slightly inboard of the PCV port. The bracket that the connector mounts to is the same bracket that supports the metal tube section of the upper radiator hose.

 

[2000StreetRod]

puzzling results

This morning I tried another pull after loading the custom tune with the knock sensor turned off. Even though the vehicle total weight was less than before by one forth tank of fuel my 27.5 mph to 59.0 mph time was 0.25 seconds greater than the reduced knock sensor pull. At the start of the pull the IAT this morning was 74 degrees compared to 80 degrees before. The average ATF temperature during the pull was 111 degrees vs 119 degrees before. The engine coolant temperature (192 degrees) was the same for both pulls. The air/fuel ratio characteristics were nearly identical as were the knock sensor retard and the overall ignition advance. The hill gets steeper with distance so I thought maybe I started a little farther up the hill. If so, I assumed that the max load would be greater for a steeper hill but the max load was identical (86%) for both pulls with the average the same (83%). One thing I didn't note were wind conditions. This morning the wind was calm and I don't know about last time. Colder air is denser but calculations at my altitude (760 ft) only result in a 1% density increase. 0.25 seconds is a 3% increase in time. The results are puzzling.

 

[kevin p]

Thanks for the pics streetrod. My bracket is empty of a connector!!! With the manifolds still on, I just can't see if there is a knock sensor in place. Will look again 2moz. Mine is an January 97 build 4.0 sohc, and can only presume there isn't one. Thanks for your time.