Forged pistions and knock sensors

[CSXT802]

Has anyone had a problem with forged pistons setting of the knock sensor.? I have JE forged pistons in my engine and my knock voltage will dance around at idle. The voltage will swing around from .02v to .28v .

[Directconnection]

wrap your knok sensor in that pipe thread tape/sealant. Worked for Gus years ago.

[CSXT802]

wrap your knok sensor in that pipe thread tape/sealant. Worked for Gus years ago.

tryed that, it made no differance.

[ShelGame]

Get a custom cal and have them bump up the sensor voltage threshold...

[Directconnection]

I would guess the sensor may be doing it's job and I'd investigate before partially disabling it if it were me.

[ShelGame]

I wasn't suggesting it be dis-abled. The knock sensor is really just a microphone (of sorts). It turns mechanical vibrations into an electrical signal. The ECU cannot tell if the mechanical vibrations it picks up are due to real knock, or just piston slap. The sensor has to be calibrated for each RPM (at lower RPM, the simple mechanical noise is lower than at high RPM) to tell the computer what is an acceptable voltage level.

If forged pistons make more noise, the sensor needs to be re-calibrated.

Actually, anything that will make more mechanical noise will need to have the knock sensor re-calibrated - to ensure the ECU doesn't pull timing when it doesn't really need to. Anything you do to increase the power output will increase the mechanical noise generated by the engine.

The opposite is also true. If you switch to rollers from sliders, there's a chance you'll get detonation that won't set off the knock retard due to the rollers being quieter than the sliders.

[Directconnection]

The opposite is also true. If you switch to rollers from sliders, there's a chance you'll get detonation that won't set off the knock retard due to the rollers being quieter than the sliders.

How can this be? If detonation is putting out "x" db's or whatever it is in vibrational oscillations (wtf is this?) it is still "x" no matter what the lesser "y" is below it.

[CSXT802]

Well I hate to say this, but I unpluged the dam thing. and guess what.? The car has never run better. Of course I would not be running this way if I were not spraying meth.

[ShelGame]

How can this be? If detonation is putting out "x" db's or whatever it is in vibrational oscillations (wtf is this?) it is still "x" no matter what the lesser "y" is below it.

The sensor just reads volts. If the voltage comes from normal mechanical noise or detonation/knock - it doesn't know. The computer has to be calibrated to a certain voltage level for each RPM. Above that voltage is considered knock, below is considered normal noise. So, if you switch from sliders to rollers, there's a chance that mild knock will come in under the voltage limit, and not set off the knock retard.

Basically, anything you do to the engine that changes the "normal" mechanical noise will affect the operation of the knock sensor. And, it can go either way...

[Frank]

Not quite Rob... before the computer reads the signal, there is a audio band pass filter setup to pass a certain range of frequencies that matches up with the knock experienced at that particular location. It will cut out high and low frequencies, so say a hammer hitting the intake at 5hz wont trigger knock.


Frank

[ShelGame]

Not quite Rob... before the computer reads the signal, there is a audio band pass filter setup to pass a certain range of frequencies that matches up with the knock experienced at that particular location. It will cut out high and low frequencies, so say a hammer hitting the intake at 5hz wont trigger knock.


Frank

5Hz, no. But the valvetrain operates in the same frequency range as the cylinder firing. So, changes there can affect the knock sensor. The pistons reciprocate at twice the frequency of the cylinder firing, so changes there can also affect the knock sensor...

[puppet]

Does knock have a specific frequency (hz) and can the sensor be calibrated to just that pitch?

[turbovanman²]

My knock sensor cal needs tweaking, my A/C compressor sets it off. Every time it cycles on and off, the SES light flashes, and I can feel the van surge as the timing is pulled out, doh!

[ShelGame]

My knock sensor cal needs tweaking, my A/C compressor sets it off. Every time it cycles on and off, the SES light flashes, and I can feel the van surge as the timing is pulled out, doh!

No kidding? It might be actual knock as the load changes when the A/C compressor kickes on. I assume it happens when cruising, in vacuum, at lower RPM?

[GLHSKEN]

good discussion

[CSXT802]

I'm glad I started this topic. This is some good stuff. I have had the knock problem for a couple years and did not know it. I should have bought a scan tool when I built the motor and went to the 3 bar cal.

[Frank]

Does knock have a specific frequency (hz) and can the sensor be calibrated to just that pitch?

5Hz, no. But the valvetrain operates in the same frequency range as the cylinder firing. So, changes there can affect the knock sensor. The pistons reciprocate at twice the frequency of the cylinder firing, so changes there can also affect the knock sensor...

Well valvetrain, piston slap, detonation, cam noise, etc, etc, all operate at their own frequencies. However, when each of these hits (for lack of better word), it transfers mechanical energy into the surrounding area. It has to do with resonance fequencies of materials. For example. Take a metal plate, hit it with a hammer, note the sound. Take the same plate, weld strips of metal on it, or even cut out sections. Hit it with a hammer, note the sound. It is different. It is because you have changed the natural response.

Most mechanical noise operates below 1kHz on our motors. Response of metal structures like our engines are due to a hammer/knock is much higher in frequency of around 10-20kHz. Alot of the wideband knock sensors are roughly tuned to your bore size of the block, but then fine tuned based on location. So another wards, if you mount it on the head, it will translate to slighly offset frequency - it could translate alot differently if you didnt have two flat surfaces that were completely smashed together tight. Same with the intake, it is going to sound differently. The benifit of the intake though is that its hollow and has thin metal and I doubt it will have much energy loss like the head. The best location is obviously the block. However I suspect they didnt like the block because of the rod to stroke ratio being bad.

With that said, mechanical noise like valves wont be picked up and doesnt fall within the mechanical frequency ranges of pistons and detonation, HOWEVER Rob is right in that it can be picked up. The valves themselves slamming into the head are similar to knock as the sensor has to be calibrated to pick up some of this energy translation thru the head, etc. However this typically tends to be minimal from what I read because the bore response of the block is what really matters and most of the energy from the valves creates frequency responses that are just to the head and dont translate alot of energy into the block.


Frank

[gkcooper]

Very interesting thread. Does the frequency change because of the piston material and/or metalurgy? Are you running the larger wrist pins? Maybe you have some pin slap?

Greg K. Cooper
Huntsville, Alabama

[Frank]

In theory, but measurable after the filter is put in? Doubtful.


Frank

[turbovanman²]

No kidding? It might be actual knock as the load changes when the A/C compressor kickes on. I assume it happens when cruising, in vacuum, at lower RPM?

All the time, idle, cruise. Its the funniest thing, lol! I just unplugged it for now,

I know my engine is noiser with forged but the old cal, the knock sensor didn't go off until heavy detonation, Rob's cal the sensor is a bit sensitive, no biggie really.

Knock sensors are calibrated for frequenzy's, the 4.3L V6 around the 92 and up era use's to sensors, both have different Hz frequency's and if you measure the ohm,s, both different.
I know when testing GM's for knock sensor issues, we hit the block with a hammer and watch the computer react, kinda neat,