ISO recommendations on the best aftermarket knock sensor kit

[Todd]

Hi,
I’d like to know what the best aftermarket knock sensor kit would be? Also, if you can help me with where to mount the knock sensor on the engine.

Thanks,
Todd

[christsay]

I am still early in my journey but I am mounting a knock sensor on the big M10 hole used for the forward starter mount / plenum support arm. The Brise starter doesn't need a forward mount and with a carbon airbag I don't need the plenum support arm either. I am using a M10 - M8 step stud for this.

An alternate location could be to drill and tap one of the pads on the intake side of the block near the head or one of the AC compressor mounts on the exhaust side of the block. Too close to the front of the engine may mean that the timing gear sounds maybe be too pronounced.

-chris

[M3 Frank]

On both of my '02 blocks I drilled the pad between the 3-4 cylinders, next to the head. IIRC there is a similar pad for the S14 block. Many cite that the S14 is too "noisy" but no one has been able to tell me exactly what that is, nor scientific proof.

I use the J&S system, and have for over 25 years along with a Lambda sensor readout. I haven't installed one of these on the S14 because I haven't found a sensor for the calculated frequency(ies).

My question is what frequency does the knock sensors that you are looking at operate? Those frequencies are different in an S14 and say, an '02; or any other engine for that matter. The common Bosch sensor is not frequency specific and it is up to the CPU to determine if there is knock. At least that's what I was told from the engineer at Bosch.

[christsay]

Found a picture of my knock sensor location:
S14 knock sensor by chris tsay, on Flickr

-chris

[christsay]

Someone used the M3 as the test subject for their masters thesis on finding the detonation frequency in a BMW S14 motor:

http://csus-dspace.calstate.edu/bitstream/handle/10211.9/947/Project%20-%20LABVIEW%20FREQUENCY%20ANALYSIS%20-%20Final_csus%20template.pdf

This should help get into the ballpark of the right frequency range to tell your ECU or knock system hardware.

"The detonation frequency of the 2.5 liter BMW S14 engine was determined to be 6,480 Hz as calculated via the LabVIEW application through the use of a Fast Fourier Transform."

He used a Bosch 0261231148, I'm using a Bosch 0261231045 that I'm pretty sure is similar except that the connector is Jr Timer like the other OE sensors used on the S14.

-chris

[stevesingo]

I think I got 6000hz when I tested mine. What ancillaries are attached can effect the frequency, so that might affect the outcome

[M3 Frank]

Interesting.

I calculated the knock frequencies by measurement of the head and piston and then calculating that.

I found that the fundamental frequency for a 2.3 liter stock engine was 5.65 Khz, the length frequency 7.04 Khz, the width frequency 13.20 Khz and the height frequency at 26.41 Khz.

Theoretical of course.

The question that still remains is with the Bosch sensor seen above, how do you just select for the frequencies you want?

Ted

[christsay]

I gotta do some digging on the original vehicle application of the sensor I am using but I believe the Bosch knock sensors are 'unfiltered' and all signal processing is done on the host ECU. I think there do exist knock sensors that are pass filtered for a specific frequency range that is application specific. This allowed for relatively simple electronics to interpret any signal from the sensor as a knock event. Doing this on the ECU is now 'easy' so the sensors just pass through everything and the ECU determines if it's knock or just other noise.

-chris

[Mick]

This thread has some info on Brendan's knock sensor install

https://s14net.vbulletin.net/forum/s14/general-e30-m3-discussion/58162-bearings-fuel-starvation-is-bad

There's probably some additional information on them buried in his build thread

https://s14net.vbulletin.net/forum/s14/topics-of-interest/project-cars/37952-brendan-s-track-car-build

[stevesingo]

I gotta do some digging on the original vehicle application of the sensor I am using but I believe the Bosch knock sensors are 'unfiltered' and all signal processing is done on the host ECU. I think there do exist knock sensors that are pass filtered for a specific frequency range that is application specific. This allowed for relatively simple electronics to interpret any signal from the sensor as a knock event. Doing this on the ECU is now 'easy' so the sensors just pass through everything and the ECU determines if it's knock or just other noise.

-chris

Correct.

The Motec SKM manual has some goon info on the operating principals.

http://www.motec.com/filedownload.php/SKM%20and%20OKM%20User%20Manual.pdf?docid=2811

[M3 Frank]

So, what I gather from Brendan's install is that his ignition system accounts for that. I didn't get that it was a J&S system, unless it is a "new" version that is programmable.

I have had two J&S systems from the '90s in my Tii and Turbo. The knock sensor is a one frequency (GM) sensor, but I think it is a different frequency that what is needed on the 2.3Liter S14 or the 2.5 Liter S14.

I did a lot of research on the Bosch sensor when they came out through the Motorsports division. After a lot of e-mailing back and forth, I gathered that the ECU was responsible for identifying the frequency, so not for me.

I would like to have another J&S that would work in the M3.

[JimK]

Someone used the M3 as the test subject for their masters thesis on finding the detonation frequency in a BMW S14 motor:
http://csus-dspace.calstate.edu/bits...20template.pdf

This should help get into the ballpark of the right frequency range to tell your ECU or knock system hardware.

"The detonation frequency of the 2.5 liter BMW S14 engine was determined to be 6,480 Hz as calculated via the LabVIEW application through the use of a Fast Fourier Transform."

He used a Bosch 0261231148, I'm using a Bosch 0261231045 that I'm pretty sure is similar except that the connector is Jr Timer like the other OE sensors used on the S14.

-chris

The paper is unsigned and so taken as preliminary. It has errors, for example it is wrongly cites that the Motronic tables for load indication are percentage of throttle when throttle position is not capable to be read by the OEM throttle switch. The switch can only signify throttle closed or throttle wide open. Load is percentage of the air door swing in the AFM and is a 0-5 volt signal. The air flow curve for the S14 can be read from the chip and created. Done so using the techniques described in the Porsche information. I haven't studied the paper completely for errors.
I use a Bosch knock sensor, they are all similar in performance per Haltech and set it up using the methods in the Haltech Elite ECU manual on a chasis dyno. My results are knock at 6330 Hz and the software is set to observe from -10deg BTDC to 60deg ATDC. There is a background noise table built to ignore noise below the engine knock threshold, created from no knock datalogs while on the dyno. The Elite software uses Fast Fourier Transform method of analysis also.
So far the engine is holding together!

[Ironhead]

Someone used the M3 as the test subject for their masters thesis on finding the detonation frequency in a BMW S14 motor:

http://csus-dspace.calstate.edu/bitstream/handle/10211.9/947/Project%20-%20LABVIEW%20FREQUENCY%20ANALYSIS%20-%20Final_csus%20template.pdf

This should help get into the ballpark of the right frequency range to tell your ECU or knock system hardware.

"The detonation frequency of the 2.5 liter BMW S14 engine was determined to be 6,480 Hz as calculated via the LabVIEW application through the use of a Fast Fourier Transform."

He used a Bosch 0261231148, I'm using a Bosch 0261231045 that I'm pretty sure is similar except that the connector is Jr Timer like the other OE sensors used on the S14.

-chris

The writer of that thesis used to be fairly active on this forum....not surprising I guess....but for the life of me I cannot remember his forum name. I don't think he has been active here for some time though.

I just mention it because I would imagine he has further insights on the topic...

[M3 Frank]

I would like to calculate the theoretical knock for the 2.5 liter vs. the 2.3 liter, as neither is even close to the theoretical.

Assuming the head is the same, what is different? Stroke? Con Rod length? piston shape?

Thanks

[hardtailer]

Stroke and bore are different, heads ever so slightly too. I seem to remember that Jake has shown a picture comparison of that in the past.
Conrods are the same.