Ignition timing issue

[raddy]

by my experience, it looks like this:

- horizontal movement is causing this isue only if TP scale have bigger steps (in given area), this is typical for transitional boost area. If is horizontal movement present in area of smooth TP steps, it follow map values exactly. This is typical for low load/cruising area of map. So its not only about slowing horizontal movement down, but also about to use as smooth TP scale as possible while is horizontal movement major. As soon as this is not possible due to limited resolution of map, you should try to completely discard horizontal movement and use only vertical movement in the same TP column, right after max. torque area, where is TP at maximum and obviously doesnt increase a lot until limiter, what enable to sue same TP column-this is how I do it. In my case horizontal movement ends at about 4500rpm, then it continue only vertically. In pure vertical movement is consult timing same as map value. In my case, Im using fixed timing during vertical movement, but this is not condition...

louiswun: you got my point
MIL-Tec: if you can make me log on the 4th gear, full throttle from 2000rpm to your choosen rpm limit, I can try to adjust TP scale of your map for full boost timing with fixed value. You can always play with TP scale while replaying log file to see changes of matracing movement....Also, from your posted log, off boost timing is corresponding with map, also, it is good to set IDLE TPS MAX for real TP value thas is reached at idle.... Area of pumping throttle a bit (like 02:13s) are not usable for timing issue, as there you closed throttle, IDLE switch is ON and different timing table is used there. If you make me another log at 4th gear crusing at cca 70-80kmh, with some medium/full throttle areas to reach turbo spool and begin of max boost, we can check it for off boost and transitional TP scale setting to get as stable timing as possible...

In my case I was almost completely depress this timing dips expect on point in transitional area, where it drop to about 16deg (target is 20deg), but its comes back to the map value in about 500rpm, so its not noticed on the road...To delete this I would need map with something like 32 columns lol....

please note that this description was tested only with ca18det and MECD0007 ECU, knock sensor completelly disconnected.

[MIL-Tec]

Hmm... If I understood correctly, after ~5000rpm horizonal movement on the map would stop (full throttle).

Sounds non-ideal and certainly not like I was expecting it should be done. I think the movements should go something like in the picture attached below. Right now I have the ignition load scale adjusted linearly. The steepness of the curves would depend on how well the turbo wakes up, and vice versa how much air goes trough when throttle is almost closed. But if that is how you think it has to be done, I don't really have a choise do I?

But in the log files, I get these late advances around 3000 rpm aswell...

I have adjusted the timing to be 15 deg at idle when "TPS-idle" -flag is lit, and the more the revs the more the advance. I've read the manuals

So this would be my understanding of ideal maptrace behavior:

[louiswun]

horizontal movement is causing this isue only if TP scale have bigger steps (in given area)

Let say my transitional boost area is from TP30 to TP70, and that is from 0 bar to 1 bar
You say to smooth the TP steps, do you mean the TP scale in the boost transition area like this
30 35 40 45 50 55 60 65 70 (each step increasing 5, I think it is "smaller step")
And this will make the map tracer walk through 9 columns
The horizontal movement of the map tracer look move faster

If the TP scale like this
30 40 50 60 70
And this will make the map tracer walk through 5 columns only
The horizontal movement of the map tracer look move slower

Both way also have smooth TP scale,
Which one works better ?

[raddy]

well, here is my example, please note that this map is only example and let say that in our imaginary engine is half boost reached at about 3000rpm and full boost at about 3500rpm, timing for turbo spool 20deg, timing for full boost 16deg... Here is how I tend do configure both rpm and TP scales to reach this kind of maptracing movement. In my example it is full throttle from crusing speed at about 2500rpm. This movement use only horizontal or vertical way opposite to above example where it crosses columns in both ways. My example generate lot less interpolating effect. Generally, Im aiming for situation, where each circle contains same/very similar values with as smooth as possible transition between them... in picture example2 is my adjustments of timing valôues to example map to reach my goal... this is just quick example, but this is how I do it with good result to ignition interpolation issue.
example:


example2

[MIL-Tec]

Okay, thanks. I will see if that makes any difference to the reported consult values.

[MIL-Tec]

Ok, I did some tweaks and now I got the actual timing around 15 deg's (says consult). Could you confirm is the consult display safe way to monitor actual timing?

The behavior of maptrace was pretty much spot on as raddy said. I did the 4th gear pulls on a freeway (I'll attach the logs here). No AFR in the second log... sorry
I advanced the maps quite a bit for the first pull, but still got 8deg at 4000rpm. After a bit tweaking I got it better.

But now afterwards when I look at the logs I note it might be using the knock maps afterall since "Knock Timing Map" and "Knock Fuel Map" is lit green on the lefthand menu.
I watch my logs usually with my desktop computer and do the tweaking in car with my laptop.

Anyway, at least I got rid of the late advances... Actual timings are still pretty conservative.
I'll have a better look with the timing maps at weekend. Currently I'm using 1bar boost... hopefully I'll have MAP sensor attached soon to my Nistune.

[Matt]

If the knock timing maps are green that is what the ECU is using at that time. It will also record which maps its using in the logs

[raddy]

Im glad that my experience helps....I can say that with further tweaking that way you will be able to reach any values that you want, maybe except small dip somewhere for short time, as in my case....
Can I ask you what part number of ECU you are using, regarding knock sensing? But anyway, I have really strong feeling that its enough to have same main and knock maps, doesn matter if only main or both are in use, interpolation issue will be still same. You should be really sure that you actually doesnt have real knock/or unplug knock sensor, this can shows behaviour of ecu in way of using knock maps.... In my case knock maps arent used at all, but I have also knock sensor disconected.

[MIL-Tec]

ECU number is still unknown. I have my old cylinderblock so maybe I just take another knock sensor from there and tie it somewhere so it won't detect anything.

[raddy]

Hi MIL-Tec, here is your map where I made adjustments in the main ignition map, in TP and RPM scale and also in map values. I adjusted area of turbo spool and maximum boost based on your 2nd log. Now map values are more consitnent in critical areas and also maptracing movement is optimized to be more or less only horizontal or only vertical. This should increase timing during fast transition area (TP 30-52), where you had too big drop that will affect off-boost/zero boost response of engine. Than I stabilized maptracing in area of turbo spool (TP 57-87) to get stable, slightly increased ignition there and then fixed max. boost area to the last column.
Please note, that I didnt changed knock map, so you can compare changes and copy afterwards. Also please note that adjustment will very probably slightly change timing, so be carefull while testing..... If you will aim for more timing generally (as current map is conservative) you can simultaneously increase map values to keep consistency....

[Matt]

Ok, I did some tweaks and now I got the actual timing around 15 deg's (says consult). Could you confirm is the consult display safe way to monitor actual timing?

Sorry for the delays but I have to split code development and support into separate parts of the week

There are various places that the CA18 code gets the timing and records it. The final timing value is dependent on if using main timing maps or idle timing maps. In case of idle then Neutral switch is also used. Temperature comes into play also (warmup timing table conditionally used). From my reverse engineering exercises and checking of table accesses the timing values always reflect the currently accessed table (after interpolation)

Just been going over repeating on the bench. Bench is a controlled environment. It will not have any knock and the conditions I put into the bench are the same as those repeated but it wont have the same history necessarily (ie crank, startup, warm temperature etc)

Consult shows in these maps that the reported value is the same as in the maps (36F05 image MEC-001). Your run is pulling out timing. See image attached

ECU number is still unknown

See top of the screen in Nistune (ECU ID). 36Fxx JDM and 39Fxx EDM

By the way, Matt, it is possible to clearly state situation regarding knock sensing and timing pull for ca18det ECU, as until now I can say that EU code doesnt react to knock at all...but it looks that other ECU variatns have this feature....

Is your ECU reporting knock (in the knock reporting panel) but there is no timing taken away? I'm not too familiar with the differences between the various CA ECUs. We mainly use JDM ones here (36F00, 36F05...) but do program up quite a few 39F00 EDM (MEC-D007) for customers in Europe

From 4th gear log I can see around 6:54 going up in load/rpm the timing is lower than shown in the maps. However repeat the same here and it matches (same temp, TPS value etc)

[raddy]

Hi Matt, thanks for detailed answer. You wrote that on the recorded log is pulling out timing, probably as knock detection feature... but if you look where is actualy timinig different, there is no chance for knock as so low rpm and without boost, Im strongly doubt that there was real knock. Also, why its changed after rescale of TP/RPM? It is very interesting that your bench test timing exactly follow map values, can I kindly ask you if will be possible to repeat same test with 39F00 EDM (MEC-D007) image for comparison? In my case is knock sensor completely disconnected, I have related DTC error present, so it shouldnt affect anything. However, I have experience that also another ecus of this kind with same behavior...

[MIL-Tec]

Edit: false alarm, some parts were just making very "knocky" noises on the floor
I did some further adjustments and seems that map and consult is about 5 degrees out in transition area. I copied the main map to knock map but didn't get rid of that problem

[Matt]

ECU will pull timing in addition to knock maps (if it switches over to them)

[PL]

Did you say you have a knock sensor DTC???? If so then that would explain your problem. Most (all??) Nissan ECU's will pull extra timing when they have a knock sensor DTC.

PL