Ignition timing issue

[MIL-Tec]

Hello,

I'm having problems setting up my CA18DET ignition. I'm trying to advance my ignition, map is asking for ~20deg but consult says only 14.
At full throttle I'm stuck at 17 degrees. As I increase my throttle, ignition advance goes down (see picture).

I wonder why the reported value and the timing map value is not equal?
Knock map is the same as my primary map and AFR is about 12.5 when accelerating.

[PL]

Check that it's not using the Knock table values.

Go to Knock IGN map and adjust values where maptraces is. Does reported IGN value change?

PL

[MIL-Tec]

Well, actually I haven't touched the maps yet... but I'm wondering is the map value only some arbitrary number or is it supposed to be degrees?
In the picture above Nistune is showing "filtered values".

Knock table is exactly the same as the primary timing map.

[raddy]

difference between map values and consult values can be caused by interpolation of map cells. Try simple test: set complete last 4 columns for example to 14 deg (to be safe), than you will find that consult value will be equal to map value (if map offset is correct). While ecu maptracing travel to the right, it will interpolate final timing from 4 cellls allways. If there are bigger steps in TP scale and/or between map values of 2 columns, you will allways get interpolated result. If maptracing travel over same values, final timing will be equal to map value....

[MIL-Tec]

Yes I'm aware of the interpolation. At the picture I posted above, in the end of the graph map is asking for about 16deg but consult display says 11. So there is something funky going on. If you take a look at the TPS vs. ignition timing you note that the graphs are almost like the same trace inverted. That's what is puzzling me.

[Matt]

Make the timing maps (main/knock) the same. ECU takes values from both and uses an average. Strange but gone through the code a number of times and uses them together

[raddy]

he stated that primary and knock maps are same....
anyway, as there are different values in last 2 columns, there will be interpolation. set culomns values to same number and test then....

[MIL-Tec]

Okay, here is the same log and same maps loaded. This time it's showing the knock table also.
As I increase throttle, hilighted cells move away from where the arrow is. (See log graph: TPS vs. ignition advance).
So I think I should be having around 20deg ignition timing, am I right?

When TPS is under 1,5 Volts I get about the right timing from the maps. After that timing is going way too late... Maybe I have missed some basic setup?
This kind of behavior would be understandable with the fuel maps... but cam position is so closely monitored with the CAS so I think the timing tables and the actual timing should be withing 0-3 degrees or something like that. Right now you see about 6 degrees distortion between these two (@ 3800rpm).

So is that consult display reliable tool anyway to monitor ignition etc.?

Right now I'm away from the car so I can't make any further tests. But I have the maps and the logs in my laptop so I can inspect them all I want.

[raddy]

mate, its still interpolation effect, this is strange behavior of ca18det ecu. once again, reason is map tracing movement between columns with bigger TP scale step between them. you really should test of filling cells with same value to see difference....anyway, in practice, there is no need to care too much about values in timing map in transition area where is interpolation effect biggest, usualy is OK to set correct timing for off boost area, then correct timing for full boost area and make as possible as smooth transition in map values between them, in area of turbo spool. This works quite well and you can easily reach both off boost timing target and also full boost timing target. To make it easier for full boost I tend to scale TP to get maptracing move only vertically as soon as full boost is reached, with fixed constant timing. This will simplify set of timig, as interpolation will be depressed, and map values will exactly corrspond with consult timing....

[MIL-Tec]

Okay, thanks for the advice. It just seems pretty strange to have 11 deg ignition without having such numbers anywhere near the used cell in timing map. Maybe I need to tune first the full boost area. I was just hoping to get better turbo response if I had the transition area figured out right. Going full boost and setting timings right needs some arrangements to be done (to hear possible detonation), so I was concentrating at the 0bar range for now.

[Matt]

Even still 11 is a fair bit out from those values. The interpolation is on the four selected cells

Can you post your BIN here and I'll repeat on the bench to see if I get the same with knock sensors stubbed

Even if you have knock maps the same if the ECU determines detonation then it will start pulling timing in addition to the knock maps. One thing you can look at in the logs is the knock count on the trace maps to see if any knock counts were registered by the software (assuming your CA18 board has knock count firmware in it... its been in them for about two years at least)

[raddy]

I got some experience regarding mentioned issue in ca18det, EU version (MECD0007) where is no active code for knock sensor, also tested it with knock sensor completelly disconected, and its still present. For me it looks that one condition for interpolating are 4 cells, but it looks that also specially fast movement between columns causing this issue, specially if there are bigger steps in TP scale. If movement is only vertical (same column is used), consult values become same/very close to the map values. This can be also visible in low load areas, where is TP scale quite smooth and where is no fast movement through columns. In this case is also consult timing same as map values...
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....

[MIL-Tec]

The interpolation is on the four selected cells

That's what I tought aswell.

I have the 0.10.13R version of nistune software, but I have no clue what version number the ECU is. Should that knock count button be available on the leftside menu if I had it on my ECU? Or do I need to upgrade my PC software to 0.10.19?

Then question about feedback bits, I have the feedback bits set like this:

Bit 0:
Bit 1:
Bit 2:
O2 Sensor:
Bit 4:
Bit 5: X
Bit 6:
Bit 7: X

Is any of those bits a switch for the knock sensor feedback? I will not be using the knock sensor at all since it detects turbo sounds etc to knocking. If someone knows what those bits actually do, would be a good software enhancement to write there what they're for.

Here is my .BIN and a .CSV log, the driving was just normal cruising and checking out the timing advance (note gas pedal pumps and ignition timing around 1:53 and 3:10 )

[Matt]

No switch for feedback bits I'm aware of

If you open up the Knock Warning panel from the menus in versions 0.10.19 knock count will be greyed out of now available. The 0.10.13 didnt grey this out but I've added improvements to grey out anything not used in later versions

[louiswun]

reason is map tracing movement between columns with bigger TP scale step between them

Hi Raddy,
I have the exact same problem like MIL-Tec
do you mean ...
to adjust the TP scale to use less columns while in boost transition area?
ie : to slow down map tracing horizontal movement through columns