knock sensor

[tony99cobra]

Hello all,

I have a question about knock sensors...

I data logged at wide open throttle the timing and knock sensor 1 activity:

RPM: Total spark advance:
2155 18.75
2575 18.75
2955 15.50
3143 16.25
3612 22.25
4019 20.00
4469 20.00
4947 20.00
5690 23.00
5886 23.75
6058 23.00
6188 22.50

I data logged knock sensor 1 to see perhaps why my car may be retarding spark at 2-3.5K (I am acutually not sure its retarding anything and it may be normal, I here no knock/ping). It makes no sense to me. There is voltage coming from the knock sensor all the time. The curve follows the rpm's and there is no unusual fluctuation at 2-3.5K.

rpm knock sensor (A/D counts)
2000 72
2500 95
3000 136
3500 213
4000 266
4500 327
5000 376
5500 398
6000 482

The louder the engine, the higher the counts. Is the knock sensor sort of an open circuit that would be inhibited if knock occurs or fluctuate more, so that there would be dips/peaks in the volts if knock occurred, because I don't see any of that.

I emailed the data looger computer support and this is what they said...

"I would have to say you are picking up some noise on the knock sensors and I think it is normal engine noises as well as the loud exhaust. Raspy exhausts will cause issues like this too. Other than shutting them off there isn't much you can do. I have played with them on a Mach1 and it wasn't till I shut it all off that the timing would do what I command."

He makes it sound like engine noise could be retarding my timing... I think probably not. The car is a 1999 mustang cobra with quite a few modifications but naturally aspirated.

 

[patprimmer]

Try the test again with a much higher octane fuel and possibly a third time with slightly lower octane. See what that does to the timing. Changes will prove knock sensor input.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[mcgyvr]

Knock sensors can definitely be fooled by odd engine noises.

http://www.eng-tips.com/viewthread.cfm?qid=217731&page=1

 

[mattsooty]

tony99cobra

Im sorry to say but you're measurements are all in vain.

In terms of the knock sensor it will always be transmitting some signal, as you know this signal is the vibration response of the cylinder head and will obviously change with varying engine conditions.

Further than that signal from a knock sensor is not a digital knock/no-knock, instead the signal is passed through filter & evaluation circuits before being compared to threshold values within the ECU - which then judges whether knock has occured or not and retards the spark accordingly.

The fact that you are seeing various different spark values at different speeds is perfectly normal and, unless you know what the corrected base timing is (which will be some function of a 3d look up table), you will never know if the ECU is retarding the spark due to knock.

Why are you so worried about knocking anyway?

One of the best pieces of knock detection equipment is the human ear, you can buy all sorts of fancy kit to plug into the knock sensor and then listen away till your hearts content, listening out for that tell tale crackling sound.

I have had brilliant results, on the dyno, with a length of of copper piping flattened at one end, drilled through and bolted to the knock sensor location with the other end in the control corridor - the knocking is very easy to distinguish.

MS

 

[tony99cobra]

Mattsooty,

"Why are you so worried about knocking anyway?"

I want to advance the timing under wide open throttle above and beyond the numbers above. From what I understand the optimum total timing advance on this car is to have it hit 26 degrees. The problem is I'm worried about pre-detonation. The air:fuel ratio is dialed in at 12.8 to 13 right now. In the past before tuning, when the car was richer (A:F 12)it would ping mildly with 87 octane (an attendent put in 87 rather than 93 once by accident). 91 is minumum as per Ford. I run 93 octane now and wanted to see of I had any knocking now before advancing the timing further. I may have a hard time hearing the engine since the exhaust is pretty loud and air flow across the intake is easily audible. I guess I have never really heard knocking, does anyone have a video with sound of it?

I guess the way to do it would be advance the timing, listen carefully and relog the timing advance and see if the ECU truely advances it or not.

 

[TheBlacksmith]

When I taught marine EFI, I likened the knock circuit to having teenagers. They talk all the time and you ignore them unless you hear words like sex, drugs, or pregnant - then you pay attention. Knock sensors hear lots of noise, but only pay attention to knock frequency range noise.

 

[patprimmer]

Tony

Once you hear knock, you can't mistake it. It sounds like a woodpecker with a hammer going at your piston tops at 1/2 engine speed.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[santorta]

Tony,

Regarding the Knock input signal as matt said, the knock signal noise signal increases with the RPM, and the ECU algorithm filters this noise, and by having a adaptive gain vs RPM lookup table, we will be able to identify the knock and the correction will automatically take place.

Why are you trying to advance the timing? In today's algorithm which has active knock system, the ECU will look for the knock signal and advance the spark timing by 1deg, and if it detects a mild knock it will retard the spark by few degrees. This operation will go in cycle to ensure that the engine is operating close to MBT at all times.

The spark data that you are looking at may be a base engine spark timing that is used for initial calculation, and then the ECU algorithm updates from that value.

I am pretty sure that a 1999 Mustang will have this Active knock control strategy.

Santhosh Arasan