GM ESC modules and matched knock sensors.

[dfarr67]

This is an EBL setup, and the ping has been an issue for some time. I dabble in tuning so don't really know it well. I was thinking last night that the learns is doing its thing, I've cut PE/SA to 0, I've taken 5 deg out universally and although it clears up most of the knock table it will still ping in a certain area even with 0 timing. I've filled up it up with better octane and the same thing. So I'm thinking it's going lean under heavy throttle and options are PE, AE, or VE adjustments. This is a 383 sbc with modern alum heads and a version of tpi- so I'm also thinking the tuned ram is really doing something.

[1project2many]

In order to stop kock the KS system will reduce spark advance when knock is detected then advance spark back to the ECM desired timing while knock is not occurring. If knock occurrs at any time then spark advance will again be reduced. In order to ensure effectiveness I would look at the kock and spark advance during knock to ensure the minimum allowed advance is not too high. If spark advance cannot be reduced enough then knocking will not stop. I would then look at the initial retard when knock occurrs and the rate at which the system returns advance.

And I wouldn't be afraid to add MEK. :)

[dfarr67]

Capture1.jpgCapture2.jpg

[Tom H]

I have been looking at the knock module from various LT1 and V6 applications in the late '90s. The circuit is housed in a plastic carrier with eight contacts. Inside there is a hybrid circuit (I believe it is thin film). On the circuit are a number of capacitors (one looks to be a larger polarized tantalum cap), LASER trimmed resistors and two ICs. One of the integrated circuits is a soic 8 pin manufactured by philips (marked NE532D dual op-amp [edit]) the other has a higher pin count (perhaps 16 pin) and is not packaged. The second chip has no markings because it is not packaged. I believe the part is made with solder bumps and is mounted directly on the substrate.

I find it interesting that when manufactured, the hybrid is mounted in the top of the plastic housing with some very low density elastomer. This material is very soft and I believe is used as a sound deadener to prevent flexing of the substrate from becoming microphonic. That is to say that the module could under some circumstance pick up vibrations and turn them into unwanted signal.

I plan to look at this module in the future & try to find it's operation, but for now this is all I have found.

One question for the group... knock is when the compression cycle heats the mixture to the point where it ignites before the spark (pre-ignition). How does retarding the spark seek to cure this? I can see how making the mixture richer might cool things. Overall retarding the timing would cool things when knock isn't present. Perhaps someone could explain?

-Tom

[brian617]

One question for the group... knock is when the compression cycle heats the mixture to the point where it ignites before the spark (pre-ignition). How does retarding the spark seek to cure this? I can see how making the mixture richer might cool things. Overall retarding the timing would cool things when knock isn't present. Perhaps someone could explain?

-Tom

Pre-ignition (ignition before spark) and "Knock" Detonation (after ignition) are two different events.

[dfarr67]

I think the trick is to prevent rather than recover. I did order the sensor, I just found it odd that the same style and engine displacement demanded a different part number.

[NomakeWan]

I did order the sensor, I just found it odd that the same style and engine displacement demanded a different part number.

The sensor will be a different part number depending on whether it's OBDI or OBDII, because the OBDI version contains a resistor inline that the PCM uses to determine whether the sensor is physically present or not (but that can become damaged and fail without necessarily causing the sensor to fail, tripping a DTC when the system is in fact still functioning as designed). The OBDII version has no such resistor because the PCM determines whether the sensors are present based on their actual output signal.

The modules all have totally different part numbers because every single one is calibrated differently due to changes in packaging between various different engine configurations and vehicle layouts, which cause changes in the audio profile that need to be filtered in different ways. It's this part that is so tricky when you start making changes to the system, and why creating a replacement module would be the only "correct" way to fix the problem of not correcting detecting knock once you've created an engine that doesn't fit the stock GM profiles.

[Tom H]

Pre-ignition (ignition before spark) and "Knock" Detonation (after ignition) are two different events.

I understand a little more now... with knock the plug ignites the mixture and a flame front starts from it's tip. Normal ignition the front continues, pressure builds and all is good. With Knock what happens next. Could someone point to a good link that describes all this??

-Tom

[LeMarky Dissod]

Pre-ignition is when the electrical spark is not the start of the flame front. It often but not quite always leads to knock. Even if it somehow is not knock is still bad because pressure builds before the piston is ready to convert the pressure to useful work.

Knock is a sudden explosion - detonation - rather than a smooth controlled burn. Instead of building, pressure spikes so fast it's like a hammer striking the piston or the cylinder or the valves.

Depending on the amount of air & fuel involved in the knock event as well as when it occurs, has an effect on its acoustics, but it always sounds like some kind of metal on metal.