knock retard--hard starting?

I am occasionally getting up to 10* of knock retard upon startup--motor runs for about a second, stumbles and dies. When the motor does start, it will stumble for a second and then recover as the retard is eliminated. In an effort to diagnose this and separate real from false knock, I have run a Xylene mix, I have tried two seperate PCM's, KM's, a couple of knock sensors and I currently have the sensor removed from the block (circuit is still connected however). I even set max. retard at -0- for all RPM less than 1600 via Tunercat--still showing various degrees of retard upon startup--how can this be? I am stumped....anyone experienced this before? HELP!! Incidentally, the second PCM seemed to perform well for awhile (maybe 15 starts or so) and then seemed to develop the knock retard issue just like the first PCM--not sure if this was just coincidence, or if it "learned" this retard behavior over the course of the 15 starts.

I've never seen my stock engine start w/o knock counts having gone up. I think the lowest I've ever seen it was 15xx.

I'm pretty sure everyone else sees "knock" at startup too. Whether it's real or false from the noise of the starter, etc., I donno...

Some retard at startup might be somewhat normal, but in my case there is so much that it is preventing the motor from starting. On my "clean" starts, my logs show no retard whatsoever, effective timing of around 21* BTDC.

I guess it's possible it could be burst knock retard... which looks for suden large differences in airflow and preemptively retards the timing.

There are separate tables and constants for this in TunerCat. Try disabling it and see if that fixes it.

Almost looks like the Opti pulses might be "off", causing the wrong timing on SOME starts. Have you checked the Opti harness connector for corrosion? Have you checked the PCM for intermittent DTC 16 and/or 36?

If you have the KS hanging loose by the harness wire, what are you doing for a ground? If you don't ground the sensor, the PCM will assume it isn't even there and use programmed default retard values.

my buddy 94 did the exact thing you are describing and we have narrowed out everything except opti. hes going to change that and see what happens. but i bet its opti.

97Z-M6--I have been reluctant to blame the Opti because it has a new cap and rotor and a real clean reluctor wheel, and runs excellent otherwise. Logically however, this is the only other component that I have not really put to the test that has anything to do with timing, so...I will give it a shot.

Injuneer--the KS is grounded with a wire. Opti connectors should all be OK, this stuff is new. Have not thrown any codes.

JSK333--that's an interesting point regarding the burst knock...I have assumed the max. retard vs. RPM values would override all other knock parameters/tables in Tunercat, but maybe I am making an incorrect assumption.