i have Tephra mod v5 patched onto my ecu, and enabled to SES light to flash whenever it sees more than 4 counts of knock..

recently, ive seen the SES light start flashing during two occasions, most downshifts from 3-2, and whenever i rev the motor in neutral. (i first noticed this when i was at the drag strip and did the 2-step)

i did a quick log earlier, with the car in neutral, reving it to 5500. from 2500 up i get knock counts ranging from 4-11.

im curious as to if its just seeing "noise" and thinking its knock, or if its something of a bigger problem. what does everyone else think?

 

Did you ever figure out what was wrong because my car is doing the same thing.

 

what is your AFR?!?!?!? is you are running the corret AF and u still having knocks is just a fake knock, u will be good if you rise the value on the knock sensor.

 

I haven't logged it yet because my AEM wideband isn't logging to EVO scan for some reason. The funny thing is when I ease into the throttle it won't knock.

 

i hope you figure it out because i am having somewhat of the same problem.

 

Bump

 

Make yourself a pair of det-cans as shown here and then listen for yourself, that way you can be 100% sure if it is or isn't phantom knock.

@ javicracer, good AFR does NOT mean you will get no knock! You can still get knock due to the timing being too advanced or a bad batch of fuel. Both of these problems will still give you the correct AFR, but can still cause unsafe levels of knock.

 

I havent run into this, but if you log "knock voltage", and "knock base", you will see if its real or false knock. (I might have it backwards) but if "knock voltage" exceeds the "knock base" its real knock.

If you have a non vendor map, post up your fuel / timing maps, and a log, and we can take a look...

 

How does that make it real knock?? The knock sensor is basically a microphone bolted to the block. If it detects a frequency within the range defined as knock it shows up as 1 knock count.

Yes, there are lots of things that can create this frequency (exhaust hitting part of the car for instance) that will cause phantom knock, but you cant just say if voltage is less than knock base it's real. All the sensor is, is a microphone. It is set to detect certain frequencies, and when it picks up these frequcenies is think it's knock, even if it comes from elsewhere.

The only REAL way to be 100% is to use a pair of det-cans and listen for yourself. That way there is NO confusion.

 

I agree w/ u that the sensor is just a micrphone bolted to the block, and I believe that det-cans are the best way to check. I was just trying to give the OP an alternate method, incase he has no det-cans, or unable to make them. I have read this is the "next best" method, although I have had no knock/flase knock issues, to test it out... just trying to help out...

 

Fair enough mate, but I am genuinely curious as to how it shows it's real knock if the knock voltage exceeds the knock base.

Does that mean if it's false knock the knock voltage will be less than the knock base?? How can the voltage be different for real & false knock?

Surely when the knock sensor picks up knock it sends a binary 1 (high voltage) for every knock count it sees to the ECU, and if no knock is seen then a binary 0 (low voltage or 0v) will be being sent to the ECU.

So how can it determine between real & false knock, it should send the same voltage/signal to the ECU regardless of what it is, as long as it's in the right frequency.

I am genuinely interested in how that shows if it's real or false knock, because the sensor just sends 1 signal for knock and then no signal for no knock, it won't have a 3rd value it can send for false knock.

 

You get piston slap if free revving or in decel or if the engine is lightly loaded.

The signal from piston slap looks like knock on an oscilloscope. Program your ECU to ignore it.