Done

 

updated post47

 

hesitation is gone, but timing pull still occurs
Do I need to address the RPM Delta table?

RPM Delta Table = accelerator position Delta table? wouldn't be the first time cobb made a mistake.

"Acceleration Ignition Map - Ignition timing can also be corrected for a rapid change in the throttle opening rate for a smoother transition into wide open throttle."

 

Did you ever resolve this?

I'm starting to think that there really is quite a bit more to the timing strategy.

Have you tried disconnecting the knock sensor and resetting the memory?

 

Still happening.
This is on a new engine installed. I have thought about resetting the ecs again. I think the rpm delta uses the knock tables perticularly the increments and decrements and max. I made some minor adjustments but have not tested it.

 

Iat.

 

talk to the hand, or rather might as well be ...

 

I just wasn't sure if I had mentioned it here previously.
My hands are good listeners.

 

It was stated numerous times by various people just like your maf number comment in the Greddy compressor map thread

 

This actually makes some sense

 

No - the RPM delta table is the RPM delta table.

Obviously, RPM delta is affected by throttle opening angle and rate, but it is not the same thing.

 

I have been seeing logs that look just like mine, similar setup, and it has no timing pull, similar intake temps, g/s, and load. Maybe this only effects newish/healthy motors,

 

There is a knock learning strategy in the ecu and it looks just like what is implemented in every other Denso DBW ECU, right down to the values of certain parameters.
I'm not good enough with disassembly to understand exactly what, if anything, it is doing. However it doesn't seem like Cobb understood it well when they stopped working on the RX8, or for that matter even today on the Mazdaspeed cars.
So, there could be a knock learning table in ram affecting timing advance, and it wouldn't show up under the knock retard pid.

 

The routines are present.
Unfortunately, the Renesis knock sensor cannot actually hear detonation, so it doesn't apply any retard based on actual timing events.
I have analog datalogs of the sensor and the only thing it can hear is a bad motor mount.

 

Analog datalogs?
It is a linear knock sensor, if you didn't see a signal then the sensor was broken or you didn't measure it correctly. More likely you thought it was a resonant knock sensor.

It is a microphone, and the response is linear throughout the audible range. If you can hear it, the knock sensor certainly can.

The Ecu filters knock from noise.

 

You clearly have no applicable practical experience with regard to the RX8.

 

This is how the knock retard value gets converted from a float to an unsigned integer for transmission over CAN.


This shows why no one ever logs knock retard.

I'm sure Team can explain the rest.

 

So are u saying that the knock senser is pulling timming, but datalogs does not register it??

 

I'm saying:
- We can't log rough or fine knock learning corrections.
- Turns out, we can't actually log knock retard either.
- The assertion that the knock sensor cannot hear detonation is ludicrous.

Linear/flat response knock sensors hear everything. The ECU then has to use filters and FFT analysis to discern knock from other noise. You can't just look at a scope trace and determine what the knock signal is, though you could with the old resonant knock sensors.

It is difficult for me to say exactly what the KR system is doing in practice. Racing Beat says the system is too sensitive. They may have been monitoring knock correction parameters in ram directly through the AUD port or mode 23.

You are seeing timing pulled, and it seems to me that knock correction is a potential cause that has been discounted by some for dubious reasons.

 

Let me rephrase the question

Is it Possible that the system is in fact picking up knock, then reducing timing, but the knock value we see in Datalogs will remain 0????

 

Yes, there are three ways that knock can retard spark. We can't log any of them.

 

Very interesting discussion.
I know zero ... but I'm lurking thanks guys

 

thank you!!!

 

Absolutely incorrect.

Start your log and take a medium-sized Crescent wrench and lightly tap the left-side motor mount.
Observe your logs.

 

Oltman is looking at code, perhaps we can log major events but not minor events, but minor events can still pull timing?

If thats wrong, why dont u tell us exactly whats happening?