Ignition timing issue

[raddy]

Hi all. Recently I had some time so installed resistor to get out knock sensor DTC. After that I immediatelly noticed that map tracing is following map as it should and I didnt get any timing dip anymore. By this, I have to say that some of my previous statements regarding timing issue wasnt really true, as they was result of incorrect input. As much as strange it looks for me, ECU really retard timing in certain map area with this DTC. Anyway, Im aplogise for any confusion....

[louiswun]

It is glad that you have solve your problem !

[raddy]

yes, issue is solved, I also had to tune AFR after resitor mod, as AFR goes noticeably leaner by about 1 point at the boost. Probably ist result of changed timing. What was confused me before, was fact that timing retard was present only in certain small area of map (transition area) and was folowing map at off boost and at full boost...

[MIL-Tec]

Nice, maybe I'll make that mod aswell. I had pretty bad clutch slips a while back because my timing jumps suddenly about 4 degrees. Boost is now set to 1.0bar with RE85 and I didn't have any slipping during track day. However I had some problems with blown boost piping... Fixed it and we'll see how well it lasts in one piece

[raddy]

Can anyone confim me that knock sensor DTC can cause also AFR enrichment beside of timing retard (on the ca18det, even if knock maps are same as the main maps)?

[PL]

Hmmm. Good question. That's one I'll have to look into.

PL

[Matt]

Good question. Well I was fixing knock today for another ECU so tried this out

When disconnecting the knock sensor, the knock sensor DTC gets raised and the timing gets pulled out

If I then reconnect the knock sensor the timing gradually goes back to where it should, but DTC code still remains until the vehicle is powered off/on again

You can see there is no effect on the injection time so it does not change the AFRs during knock sensor disappeared condition

[Matt]

You will also see on knock reconnect the 'knock' flag in the log is showing for a short period of time and then it disappears. ECU stays locked on the knock maps for the remainder of the time the ECU is running

[raddy]

thanks Matt for test, thats clear a lot....from the log it looks that knock timing retard is not based on the knock map, but its rather standalone function of ecu, as timing get back to previous values even if knock map is still used... in the future I will take care about knock sensor DTC....
By the way, will be possible to perform similar bench test but instead of disconnect knock sensor try to simulate knock signal (if thats possible) to see how ecu reacts?

[Matt]

Hi Raddy

When I developed the firmware for knock sensing tests I was unable to repeat knock fast enough to trick the ECU into thinking the knock was real to simulate the results. I'm unsure at this time how to 'fake' knock input to the ECU as to pull timing

I had a signal generator on the R32 ECU bypassing the knock sensor board filtered input. However even a quick burst of pulses (ie say 1/2 second) was enough to flatten timing to 2 degrees, so I need a better emulation device but not sure where to start with that