| View previous topic :: View next topic |
| Author |
Message |
RC
Joined: 25 Mar 2007
Posts: 2637
Location: Australia
|
 Posted: Tue Jun 12, 2007 12:45 pm Post subject: KNOCK SENSOR LOCATION |
 |
|
Anyone running a knock sensor on a 2.0L?
Where is it mounted?
I am considering locations to mount my knock sensor as part of an EFI conversion. We all know this is a noisy motor and was built before knock sensors became common, so optimum positioning will be somewhat trial and error.
The cold (intake) side reduces the temperature factor but the most suitable mounting boss as I see it is the one lower down the block adjacent to cylinder 2. Therefore my preferance at the moment is at the boss between 2 and 3 higher up the block on the exhaust side, just in front of the starter. I have headers installed and have enough clearance but it won`t be far below the manifold face.
Any input from members who may have experimented with this will be greatly appreciated as will valid comments and opinions from those with technical knowledge on this topic.
Thanks,
Roger |
|
| Back to top |
|
 |
leadfoot
Joined: 11 Dec 2002
Posts: 2222
Location: gOLD cOAST Australia
|
| Posted: Tue Jun 12, 2007 1:35 pm Post subject: |
|
|
Whenever I'm at the dyno they connect one line to a throttle body bolt and one line to an intake manifold mounting bolt.
Leadfoot
_________________
1981 ROW 924 Turbo -
carbon fiber GT mish mash
LS1 conversion in progress... |
|
| Back to top |
|
|
RC
Joined: 25 Mar 2007
Posts: 2637
Location: Australia
|
| Posted: Tue Jun 12, 2007 1:59 pm Post subject: |
|
|
Thanks Leadfoot, thats the sort of info I`m chasing.
Wasn`t sure about the head let alone the manifold. The sound would conduct and may be more prominent there against the background noise. The dyno boys should know.
Thanks for that, hope you`re getting somewhere with your rack. 
Roger |
|
| Back to top |
|
|
leadfoot
Joined: 11 Dec 2002
Posts: 2222
Location: gOLD cOAST Australia
|
| Posted: Tue Jun 12, 2007 2:05 pm Post subject: |
|
|
replaced it on the weekend, had the wheel alignment done... at beaurepairs.... handles really well now...
finished refitting the intercooler pipe which keeps blowing off and finished the air bypass valve hosing, damn near every type of hosing would crumple under vaccum, I now have aluminium tubing in place, just a little bit of tuning at cold start and I'm done.
Leadfoot
_________________
1981 ROW 924 Turbo -
carbon fiber GT mish mash
LS1 conversion in progress... |
|
| Back to top |
|
|
RC
Joined: 25 Mar 2007
Posts: 2637
Location: Australia
|
| Posted: Tue Jun 12, 2007 2:48 pm Post subject: |
|
|
Great to hear that mate!
Probably 1.5K cheaper than the other sharks too!
Sure you can accept IC pipes blowing off as a GOOD PROBLEM 
Cheers |
|
| Back to top |
|
|
RC
Joined: 25 Mar 2007
Posts: 2637
Location: Australia
|
| Posted: Fri Jun 15, 2007 12:02 pm Post subject: |
|
|
KNOCK KNOCK
Anyone there? |
|
| Back to top |
|
|
Tigger937
Joined: 11 Apr 2004
Posts: 919
Location: PCA Milwaukee Region
|
| Posted: Fri Jun 15, 2007 3:01 pm Post subject: |
|
|
Only thing I might be able to add is proper mounting is crucial to sensitivity as is the case with accelerometers. Maybe even more important than location to a certain extent?
_________________
1981 931 (Concours)
1982 931 (Daily Driver)
"Think outside the box" |
|
| Back to top |
|
|
RC
Joined: 25 Mar 2007
Posts: 2637
Location: Australia
|
| Posted: Fri Jun 15, 2007 8:17 pm Post subject: |
|
|
Appreciate your input Tigger937. What you say seems quite logical since they are essentially a piezoelectric microphone.
Most of them have a stud that screws into the motor, like the common GM type I was considering. Others have a hole in the centre like a donut that accepts a bolt.
Now I`m not sure, as I dont fancy drilling out and tapping holes in a trial and error attempt. The donut type will be much easier to experiment with in different locations. Yes, the mounting will be critical. It needs good acoustic transfer to be effective.
Might have to chase up a donut and start with it bolted to the manifold like Leadfoot suggested. What do the dyno techs use?
Any further thoughts welcome.
TIA,
Roger
_________________
World`s quickest 924 2L slushbox
| Allan @ DTA wrote: | | I have no issue with superchargers, they are for guys who want to drive a car rather than talk about horsepower with their baseball cap on backwards |
|
|
| Back to top |
|
|
Min
Joined: 04 Nov 2002
Posts: 2368
Location: Vernon, British Columbia, Canada
|
| Posted: Sat Jun 16, 2007 2:23 am Post subject: |
|
|
Personally, having a knock sensor to mount myself, I've been hoping I can mount it to the hole right beside cylinder 3 on the exhaust side, there is a stud hole there very close to were the head meets the block. And cylinder 3 will knock before any of the others do. There unfortunately is not enough room in that spot for my knock sensor due to where exhaust manifold is mounted.
Min
_________________
Custom means it didn't come from a box.
1980 n/a with EDIS and Megasquirt II Injection. 7 different colors and counting. |
|
| Back to top |
|
|
Peter_in_AU
Joined: 29 Jul 2001
Posts: 2743
Location: Sydney, Australia
|
| Posted: Sat Jun 16, 2007 10:47 am Post subject: |
|
|
Silicon Chip magazine published a knock sensor kit in April 1996 based on the GM/Commodore sensor and the filter board inside the Delco ECU. I think it was just a warning light but it might be useful to find the article. Could be useful in working out a good sensor location.
Obviously your major problem will be if you choose a bad location for the sensor then the ECU will keep retarding the ignition when it doesn't need to or won't when it does.
I'd suggest a long research session on google. This whole question will have been explained, discussed and argued about many times before.
_________________
1979 924 (Gone to a better place)
1974 Lotus 7 S4 "Big Valve" Twin-cam (waiting)
1982 924 (As featured on Wikipedia)
Learn to love your multimeter and may the search be with you |
|
| Back to top |
|
|
RC
Joined: 25 Mar 2007
Posts: 2637
Location: Australia
|
| Posted: Sat Jun 16, 2007 12:09 pm Post subject: |
|
|
Min, thanks for that suggestion. It probably is a more sensitive location.
Although I have headers that give more access it would be a PITA at best and may not last long with the heat. Not even sure if it would even fit with the stock manifold on. Thats why I was leaning towards the other boss slightly forward and lower, but that may comprimise performance. Dont know myself , so was hoping someone had already played around with this.
Thinking now I will be better off getting a bolt on donut type and try these locations and see if I can quantify the results.
Cool Peter, get the feeling we are heading up the F3 side by side with this. 
Been a keen reader of SC and EA and have this circuit too. Quite a clever, yet simple design with the hi/lo pass filters and the schmitt triger input from the ignition. Was looking at taking the analog signal and interfacing it to MS.
I accept your advice, as I can see a lot more research is needed. Still have little idea how it will go on our engines. It was too optimistic of me to hope someone could save me the time.
Good weekend to all
Roger
_________________
World`s quickest 924 2L slushbox
| Allan @ DTA wrote: | | I have no issue with superchargers, they are for guys who want to drive a car rather than talk about horsepower with their baseball cap on backwards |
|
|
| Back to top |
|
|
Tigger937
Joined: 11 Apr 2004
Posts: 919
Location: PCA Milwaukee Region
|
| Posted: Sat Jun 16, 2007 7:06 pm Post subject: |
|
|
| RC wrote: | Although I have headers that give more access it would be a PITA at best and may not last long with the heat. Not even sure if it would even fit with the stock manifold on.
Thinking now I will be better off getting a bolt on donut type and try these locations and see if I can quantify the results.
I accept your advice, as I can see a lot more research is needed. Still have little idea how it will go on our engines. It was too optimistic of me to hope someone could save me the time. |
| Peter_in_AU wrote: | Silicon Chip magazine published a knock sensor kit in April 1996 based on the GM/Commodore sensor and the filter board inside the Delco ECU. I think it was just a warning light but it might be useful to find the article. Could be useful in working out a good sensor location.
Obviously your major problem will be if you choose a bad location for the sensor then the ECU will keep retarding the ignition when it doesn't need to or won't when it does. |
Couple of things. First, why the need to mount on the EM side with all that heat? I don't understand why the intake side wouldn't be just as effective, like just under the IM near #3. As far as knock sensing is concerned, what side you mount on shouldn't make any difference due to symmetry. Am I missing something here?
Second, every engine is going to have a different spectral signature, therefore you can't expect a given transducer/detector circuit to have the same sensitivity on a V8 vs. V6 vs. rotary vs. IL4, for example. (Engines of similar construction, type and make should be relatively similar though.) To get around this problem of 1 knock sensor (transducer and detector) for all, one needs a sensor specifically tuned to our engines, i.e., one that focuses on that part of our unique spectrum that produces the greatest change during a knock condition while filtering out all the extraneous noise.
_________________
1981 931 (Concours)
1982 931 (Daily Driver)
"Think outside the box" |
|
| Back to top |
|
|
leadfoot
Joined: 11 Dec 2002
Posts: 2222
Location: gOLD cOAST Australia
|
| Posted: Sat Jun 16, 2007 7:48 pm Post subject: |
|
|
Yep, even if you do find a good location, the readings may all be wrong...
It would be alot simpler to just your car to a dyno and test it at full load with equipment designed to suit all engines...
If you want to pursue this further there were suggestions that using a stethoscope provides a good means to test for knock, I would suggest various locations can then be tested to ascertain noise vs frequency vs sensitivity...
Leadfoot
_________________
1981 ROW 924 Turbo -
carbon fiber GT mish mash
LS1 conversion in progress... |
|
| Back to top |
|
|
RC
Joined: 25 Mar 2007
Posts: 2637
Location: Australia
|
| Posted: Sat Jun 16, 2007 9:45 pm Post subject: |
|
|
Welcome to this thread Tigger937.
You wrote:
| Quote: | Couple of things. First, why the need to mount on the EM side with all that heat? I don't understand why the intake side wouldn't be just as effective, like just under the IM near #3. As far as knock sensing is concerned, what side you mount on shouldn't make any difference due to symmetry. Am I missing something here?
|
There is no need to mount on the hot side and I would much prefer to have it on the intake side as was stated in my first post. From my very limited understanding on this topic I believe that that sensor is better placed close to the source of the knock, ie the combustion chamber; while at the same time being distant from other noise such as the valve train especially, pumps, gears, and indeed, any exhaust noise.
Trust you`re not looking at a bare block, since the boss adjacent to #3 just under the IM takes the top engine mount bolt, after it passes through an integral sleeve approximately an inch deep. The only viable option here is using a donut shaped sensor that would probably disintegrate when torqued up.
Therefore the next best location on that side is the boss next to #2 at about the bottom of the stroke. Personally I feel this may be too low for optimum pickup. However I dont know, hence this topic. Also as Min pointed out knock would theoretically occur in #3 or possibly #4 first since they run hotter.
Tigger937 wrote:
| Quote: | | Second, every engine is going to have a different spectral signature, therefore you can't expect a given transducer/detector circuit to have the same sensitivity on a V8 vs. V6 vs. rotary vs. IL4, for example. (Engines of similar construction, type and make should be relatively similar though.) To get around this problem of 1 knock sensor (transducer and detector) for all, one needs a sensor specifically tuned to our engines, i.e., one that focuses on that part of our unique spectrum that produces the greatest change during a knock condition while filtering out all the extraneous noise. |
Knock sensors are the automotive equivalent to a piezoelectric microphone. Same thing just heavy duty, more temperature tolerant, and of course, more expensive. In fact in the Silicon Chip (magazine) article that Peter-in-Au mentioned that is exactly what was used as an alternative to an automotive unit.
While there will undoubtedly be some variations between individual units they are all basicly similar and have a flat linear frequency response far wider than the spectrum where knock occurs, typically between 1 and 8 KHz.
It is the associated electronics that "tunes" the sensor to any particular frequency, either by altering R/C parameters or digitally, after it is first amplified. In other words, its all in the black box not the sensor.
The audiable knock frequency can to some degre be calculated by applying the formula:
Frequency (Hz) =900/pi x cylinder radius(metres)
This translates to about 6.6KHz for our engines. IMHO I feel this may be a little high as it also depends on combustion chamber design, plug location and other factors.
Other noise can to a limited extent be minimised by incorporating a schmitt triger (electronic switch) that only turns on during a timimg interval after the spark occurs, during which knocking may be present.
Considering all the above there is still no perfect solution to eliminating knock, and especially with our engines will involve much trial and error to reduce it to a safe level when running high compressions or boost.
So really the best option is leave your engine as Porsche made it, tune it to the book, and use the reccomended fuel.
Roger
_________________
World`s quickest 924 2L slushbox
| Allan @ DTA wrote: | | I have no issue with superchargers, they are for guys who want to drive a car rather than talk about horsepower with their baseball cap on backwards |
|
|
| Back to top |
|
|
Min
Joined: 04 Nov 2002
Posts: 2368
Location: Vernon, British Columbia, Canada
|
| Posted: Sat Jun 16, 2007 11:11 pm Post subject: |
|
|
| Tigger937 wrote: | | Couple of things. First, why the need to mount on the EM side with all that heat? I don't understand why the intake side wouldn't be just as effective, like just under the IM near #3. As far as knock sensing is concerned, what side you mount on shouldn't make any difference due to symmetry. Am I missing something here? |
Conviently located mounting hole.
| Tigger937 wrote: | | Second, every engine is going to have a different spectral signature |
Apparently the frequency of the knock is simply based on the bore of the cylinder, that is all.
Min
_________________
Custom means it didn't come from a box.
1980 n/a with EDIS and Megasquirt II Injection. 7 different colors and counting. |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
924 FAQ (Technical)
