I tried something, and I dont understand what happened. My bin I'm using had around 9* of advance in pe mode @ 3200 rpm. Made a short run, and it had 8.5* knock retard. I then burned new chip with 3* @ 3200 and made same run. Had 8.2 * knock retard at same rpm/lv8. I guess I'm wondering why with 6* less timing, I had the same retard with the same conditions. Anyone have any good ideas?? Is it possible the knock sensor hears something, and freaks out????

 

Sometimes U have to reduce the timing before the rpm the sensor picked up the knock.
Example:
3200rpm = knock
reduce the timing at 3000rpms to see if that helps.
njdaewoo

 

I agree..

knock is the result of something that happened just before (and is possibly continuing). If you overheated your combustion chamber from 2,400-3,000 by being to lean with to much advance the knock will come in at say 3,200 for example and you'll get the counts there.

 

I took the timing out on the whole pe curve, not just at 3200 rpm. The 5.5d i took out @3200 would have been about 4.75 out @3000, and about 4 out @2800 etc. It was along a curve, not just like taking it out with a switch. Remember, this is the pe timing table, not the timing map for the engine, so it will work along the same lines as a curve. I know the knock sensor is picking up detonation and has a certain lag time. Im also aware of something prior to the actual knock starting the whole process. what Im not getting, is the fact I took out alot of timing, and it made no difference in the amount of retard. And i wasnt hitting the knock retard limit, i checked . That would have explained it, but of course, that would have been way to easy.
Thanks
Bob

 

If you were to do some real fast data logging, and include a cam synch signal you'd see there is very very very little lag. You can pinpoint which cylinder is the culprit, with the right software, ie J+S Electronics uses the oem GM sensors in their $600 unit.

There have been some interesting blurbs on the GMECM list about the GM Factory Detonation detection strategies in the LeMans C5s.

If changing the timing has no effect, try adding fuel.

If still no change, might be false knock.

 

I have a graph showing knock and the timing retard on my GN using the 148 ECM. Although not a 3rd gen ecm, the opeartion is probably similar. I don't have a way to post pictures, so if anyone is interested in seeing this graph, or maybe someone can post it here for me , email me.

Dave

gnjones231@mchsi.com

 

FWIW, I think I found my prob. I was going to" desensitize" my knock sensor with a brass fitting, and tried to remove it. The d*mn thing is in so tight I rounded the hex part, using a 6 pt snapon socket, and a 18" flex handle. It wouldnt budge, and i was worrying about wringing it off. Now, I guess Ill just put it on the other side, in that drain, and extend the wire. Thanks to everyone for their input.
Bob

 

I haven't ran into a knock sensor yet I didn't haev to remove with a monkey wrench. I doubt thats your issue.

 

FWIT,

I was having a similiar problem with the *lag* time using Craigs software. Not sure if all software acts the same, but I was able to log MUCH, MUCH quicker and more accurate with the program running in pure dos rather than windows.

You could visually see the difference and also the logs were much more accurate.

Might want to try that.