Knock wont go away need help

[Mountain_man]

Ok I'm still dealing with issues on my 91 burb. When I datalog using data master I can watch the knock count (about one a second at low rpm) steadly rise. This causes the timing to reduce itself to about 9 or 10 advance instead of the normal 22 or so. Once the rpm is a bit higher (past 3000rpm) the knocks basically stop and it runs great. I'm beside myself as I've replaced many things having to do with other problems but this one wont go away.
The timing is set at 2 and I've tried 0 and some retard and more advance (with timing plug disconnected) yet it continues to do the same thing. Could the distributor be off a tooth and cause these issues? Wiring trouble between distributor and Spark module?

I need some ideas that may help as right now this is all I have till I get another 700r4 trans for diesel.

New parts within last 6 months:
Distributor
Wires
Cap/rotor/plugs
ICM
fuel injectors
Misc wiring redone

This problem has slowly gotten better however will not go away.

Thanks,
Joshua

 

[Fordum]

Something to think about.
There is no such thing as a knock sensor.
That gadget you are reading that is telling the engine that it is knocking, is nothing more than a specialized microphone.
It hears a sound in a certain frequency range, and tells the computer it hears a knock. When what it is hearing is a sound like a knock.

If the engine is actually knocking, then all is well. But if you have a loose washer under an exhaust manifold bolt that is ringing at the right frequency, then the sensor is going to lie to you.

I actually know of one instance where a guy kept bringing his truck in to a friend's shop. The truck ran great. But, whenever he hooked up his old beat-up utility trailer, it just went to hell.
At first, my friend thought it was due to the load, but when he was told that it started even with an empty trailer, he was stumped for a while.
Finally discovered that the knock count was going through the roof when they hooked the trailer up.
The guy had a real POS homemade hitch, and the trailer was not much either. Between the two, they caused so much vibration that the sensor was getting fooled.

In your case, I listen all around the engine for any stray noise, and if you don't hear anything, I would consider a new knock sensor.
But first make sure that one is on tight. If it gets loose, that can cause false readings too.

 

[Mountain_man]

I will recheck the knock sensor however it is also new (huge list of new parts) perhaps it's loose. I will do as you stated and listen for any strange noises.

Any chance an exhaust leak on passenger side could cause a strange knock sound or perhaps causing the engine to actually knock? Just thinking of things up. Thanks for the reply.

 

[Stomis]

Despite the fact that it is a knock sensor your motors not actually knocking. If your motor was knocking you would have a problem somewhere in a rod, bearing, something like that.

The motor is detonating aka pinging.

Now that I've ranted about terminology like I always seem to do onto the discussion.

Make sure your setting the timing with the with the computer wire disconnected.

Change your thermostat and flush your radiator. Pinging and detonation is a side effect of excessive heat. Have you tried super or plus gas and seeing if it still does it?

 

[Mountain_man]

I do of course unplug the timing plug to do the timing. I have tried 89 and 91 octane as well however it doesn't go away. What is standard blm on these things 128? higher lean lower rich? Thanks

 

[Mountain_man]

Ok I think it's all wiring issues. Anyone know where I can get a wiring harness for a 91 suburban? it's a 350 with the 4l80e (computer controlled). Painless has nothing I can find. I checked a few other places but waiting for replies to see if they got them. If you can help I'd appreciate it.

 

[Stomis]

Call Street and Performance. Much better prices than painless.

 

[Madspeed]

Something to think about.
There is no such thing as a knock sensor.
That gadget you are reading that is telling the engine that it is knocking, is nothing more than a specialized microphone.
It hears a sound in a certain frequency range, and tells the computer it hears a knock. When what it is hearing is a sound like a knock.

If the engine is actually knocking, then all is well. But if you have a loose washer under an exhaust manifold bolt that is ringing at the right frequency, then the sensor is going to lie to you.

I actually know of one instance where a guy kept bringing his truck in to a friend's shop. The truck ran great. But, whenever he hooked up his old beat-up utility trailer, it just went to hell.
At first, my friend thought it was due to the load, but when he was told that it started even with an empty trailer, he was stumped for a while.
Finally discovered that the knock count was going through the roof when they hooked the trailer up.
The guy had a real POS homemade hitch, and the trailer was not much either. Between the two, they caused so much vibration that the sensor was getting fooled.

In your case, I listen all around the engine for any stray noise, and if you don't hear anything, I would consider a new knock sensor.
But first make sure that one is on tight. If it gets loose, that can cause false readings too.

Umm who the hell lerned you that that sensor is a "specialized microphone" ? Its a Pressure transducer, A crystal that makes voltage when you put pressure on it ( as simply as I can put it to you )

Next one is How the **** can this trailer send these frequencys through the RUBBER engine mounts? MAYBEE from the exhaust hitting the frame under load. anywho Im not here to beat you up lets help this guy out

Look for the GM Tsb's on this exact problem its a known issue. Its NOT your wiring harness MAKING voltage,

If I come across the info Ill try to rmember and send it off to ya
But your deffinetly going the wrong direction here. you have a bad sensore thats possibly cracked or a noise at that rpm in the engine

 

[Madspeed]

Directly from The manuals


Purpose:
The Knock Sensor (KS) sends a signal to the powertrain control module (PCM) to retard ignition timing during spark knock conditions.

Operation:
The knock sensor is a piezoelectric device. A ceramic crystal vibrates with engine knock. The vibration of the crystal produces a voltage signal which is relayed to the control module where the signal is interpreted by the control module and used to calculate necessary ignition timing adjustments.

When the knock sensor senses no more knocking the PCM advances timing in small increments. This allows the control module to maintain maximum timing advance under various conditions.

Location:
The knock sensor is located in the engine block.


Circuit Description:
The Code 43 circuit consists of a knock sensor with one wire that goes directly to the PCM. There are two Code 43 checks performed by the PCM. One check consists of monitoring CKT 496 for a voltage that is more than .04 volt and less than 4.6 volts.

If voltage is either too high or too low for 16 or more seconds, Code 43 will set. Once engine temperature reaches 87.5°C, MAP is over 81 kPa, and engine speed is less than 3200 rpm, the PCM will perform a self check. This self check will advance the timing until it receives a knock signal. If no knock signal is received during two consecutive tests, Code 43 will set.

Test Description:
Numbers below refer to circled numbers on the diagnostic chart.

1. The first test is to determine if thew system is functioning at the present time.
2. Test two determines the state of the 5 volt reference voltage applied to the knock sensor circuit.

Diagnostic Aids:
The PCM applies 5 volts to CKT 496. A 3900 ohm resistor in the knock sensors reduces the voltage to about 2.5 volts. When knock occurs, the knock sensor produces a small AC voltage that rides on top of the 2.5 volts already applied. An AC voltage monitor, in the PCM, is able to read this signal as knock and incrementally retard spark.

If the ESC system checks OK, but detonation is the complaint, refer to Detonation/Spark Knock/Pings See: Testing and Inspection\Symptom Related Diagnostic Procedures\Detonation/ Spark Knock

 

[Fordum]

Umm who the hell lerned you that that sensor is a "specialized microphone" ? Its a Pressure transducer, A crystal that makes voltage when you put pressure on it ( as simply as I can put it to you )

That would probably be Dr. Robert C. Jones. He was my electronics instructor when I got my first degree.

Here is the definition for a microphone:

micro·phone (mī′krə fōn′)
noun
an instrument containing a transducer that converts the mechanical energy of sound waves into an electric signal, used in telephony, radio, sound amplification, etc.

Remember, while I have lots of interests, my primary degrees are in electronics. So, to me, a transducer IS a microphone. Just a specialized one.

The only point I was trying to make, and I made it poorly, is that a "knock sensor" is just a name.
It does not magically look into an engine and see detonation.
It just listens for a certain range of sound frequencies and if it hears them, it sends back the knock signal.

Actually, of course the vibrations at the correct frequency create an AC signal that the computer interprets as detonation, but I was trying to keep it as nontechnical as possible.

Also, remember, I go a long ways back.
Back before everybody had scan tools, there used to be a test of the knock sensor. You watched the timing, and tapped the block with a small hammer while the engine was running.
If the timing shifted, the sensor was working.

Your hitting the block with a hammer is NOT knock, pinging, or predetonation, but the sensor thinks it is.

As for the trailer, I agree, it doesn't seem likely.
I had been knowing this guy for a long time. He was a good mechanic.

He called me and told me about this truck that was whipping his butt. Asked me if I could ride over one weekend and help him out.

We made arraignments to meet the next weekend.
Before then, the guy brought his truck in again. They test drove it, ran awful.
At this point, the guy mentions that it only did it when the trailer was hooked up.

They tried it.
Sure enough, unhook the trailer, ran fine. Hook it up, bad.
My guy thought maybe it was some kind of a load problem even though that did not make sense.
He tied something, I think an old engine block, to the guy's truck and had him drag it up and down the dirt road.

Truck ran fine.
Finally, in sheer desperation, he took it to a welder shop and had them replace that POS hitch.
Problem solved.

He called me back and told me I did not have to come over.
Told me his best guess.
I do not know if he ever saw the timing change or not.
I was just glad I did not have to drive 40 miles one way to help him out and could go fishing instead.

Anyway, my overall point is, any vibration at the correct frequency will be read by the computer as a knock.
And, I have seen cases where it was definitely caused by a loose bracket, or something besides a knock.
But the trailer was the only one I could remember off hand.

It seemed that he had checked all the obvious possibilitys, and I wanted to let him know about the not so obvious one.

anywho Im not here to beat you up lets help this guy out

Don't sweat it, never thought you were. Its just a case of different specialties conflicting.

 

[Mountain_man]

so a shorted ignition wire couldn't cause the below 0.4 or above 4.96 condition? in order to cause an incorrectly adjusted spark timing? Triggering a knock (ping)? I'm just trying to figure this out as once in a while the thing will run like a striped ape (perfect) and the rest of the time it wont. Thanks for the advice too.

 

[Fordum]

You say you changed the knock sensor.
Have you still got the old one?
If so, here is an off the wall idea.
Hook a ground wire to the old sensor, take the sensor wire off the new one and put it on the old one.
Then wrap it in foam or shield it from vibrations and put it somewhere where it won't get knocked around.

Then monitor the engine for a little while.

If you still see a knock signal, then we are looking at the wrong place. Its not the engine, its something fooling the knock system.
If not, then that points to the engine actually knocking.

I don't think you can leave it unhooked for long, since according to what was posted here, the computer is going to start changing the timing to create a knock after a while.
But, if you have everything setup, and run it until the signal shows up, turn off the engine, swap the sensor wire and restart, you should be able to tell in a minute or so what it is going to do.