A way to track knock on Z32

RomChip200 · Post by **RomChip200** » Wed Oct 28, 2009 11:57 pm

Matt,

I modified the consult registers table @EA00 to log some other registers instead the standard ones, with log player.
I can only use the AFR LHS&RHS and AFR learn LHS&RHS fields to report the values faithfully.
On all other fields, you make some post-processing on hex values to make them meaningful (injection times in ms, digital 1-4 with some bit fields, ...).

1) Could you plan an enhancement in Nistune ?: the capability to log some addresses w/o post processing (I need 0x0-0xFF value range in these fields). The purpose is to replace the standard registers @EA00 with some addresses I need to trace. It's a kind of ad-hoc mode. You "just" have to extend the capabilities of the panel "Select Consult registers" with "Select Consult registers or own locations" to trace.

2) With this trick (only 4 bytes fully available: AFR LHS&RHS and AFR learn LHS&RHS fields), I managed to highlight when knock starts to appear. Let me explain:

Here is my registers table @EA00:
Offset 0 1 2 3 4 5 6 7 8 9 A B C D E F

00006A00 14 0A 14 0B 16 4D 16 4E 14 08 14 09 FF FF FF FF .....M.N....ÿÿÿÿ
00006A10 14 0C 15 DA 15 D9 14 07 00 92 14 A7 FF FF 14 28 ...Ú.Ù...’.§ÿÿ.(
00006A20 FF FF FF FF FF FF 15 3C 15 6E 15 6F 14 10 14 E1 ÿÿÿÿÿÿ.<.n.o...á
00006A30 FF FF FF FF 16 14 16 15 16 10 16 11 00 F0 00 F1 ÿÿÿÿ.........ð.ñ
00006A40 FF FF 15 3D 15 70 15 71 FF FF FF FF 14 8C 14 8D ÿÿ.=.p.qÿÿÿÿ.Œ.

Peace · Post by **Peace** » Thu Nov 05, 2009 2:22 pm

Really interesting if we can have knock monitoring inside nistune.

Matt · Post by **Matt** » Thu Nov 05, 2009 6:44 pm

some ECUs have it but i"ve stopped looking at it temporarily because we found the ECU missing actual knock as well as reporting false positives. I will still finish the rest off in due time but not as high priority as it used to be

RomChip200 · Post by **RomChip200** » Fri Nov 06, 2009 2:07 am

Some progress here:

0x1614-1615 is timing_offset2
0x1610-1611 is knock_value

When knock starts to occur, 1611 starts to increase from 0 up to 200 and more. This is light knock here, read below to understand.
The ECU has an additive behavior regarding knock, this is really powerful and well-brained. 1611 and 1610 reflect this behavior.
If knock goes on, 1610 starts to be decreased (1611 continues to evolve, sometimes it restarts from 0) from 0, 255, 254 .... down to 248: 248 is the knock threshold for which, knock_timing_map is definitively entered.
If 1610 is in the 255-249 area and if you manage to perform some strong accelerations w/o producing knock for a time (=safe condition restored), 1610 is incremented ... 253 ... 254 ... 255: knock fully disappeared.
1615 gives you the importance of knock when occurring, when small value (50-150)= light knock, above 200 is stronger knock.
1614 reflects more or less the number of degrees for retard, it's 0 by default:
254 is 1 degree
253 is 2 degrees
...
This last behavior is validonly when the value decreases.

Have a look to the file attached. This is my log with 1610, 1611, 1614 and 14CF. I got some knock at the beginning of my trip.
14CF is knock_inc, has a link with the knock table, not really relevant for me.

But be careful, this is really normal for an engine to get knock sometimes (engine cold, strong accel and low octane, ...), this is why the ECU is adaptative, the regular timing map is only some reference values, the ECU constantly decrease from 1° or 2° in some conditions on a cold engine. This is why the threshold is 248 ... if prolongated knock occurs ==> knock_timing_map

RomChip200 · Post by **RomChip200** » Fri Nov 06, 2009 3:43 am

I confirm the number of degrees of retard is symbolized in 1614.
Line 1751 in file attached: targeted advance was 17°, I got 13° only and 1614 = 251 at this time (255-251= 4° retard)

RomChip200 · Post by **RomChip200** » Thu Nov 19, 2009 8:32 pm

Matt,

Does it help you to go further ?

Bernardd · Post by **Bernardd** » Fri Nov 20, 2009 12:52 am

Rom, I'm not familar with how you got the knock bits into the datalog, please explain for myself and others.

Matt · Post by **Matt** » Fri Nov 20, 2009 12:54 am

yes it will help me. thanks for posting this here. im strained for time at the moment due to family matters. hopefully all gets better soon

bernard - would be for 8 bit

Bernardd · Post by **Bernardd** » Fri Nov 20, 2009 1:20 am

i noticed it was posted in the type 2 section....i hope all is well the family

RomChip200 · Post by **RomChip200** » Fri Nov 20, 2009 2:19 am

Z32_VG30 is Type 2 and my post concerns 300zx, then everything is right, no ?