SP3 N03 SPK:LOG:NOKCTS

[UNEEK]

Hi,
I have been using the knock count logging but was getting some strange results. So i have put it on the test bench and found that the RPM and MAP are incorrect. The RPM is offset by 400RPM and a MAP of 80kPa logs as 90kPa. This makes it difficut to use this to adjust the spark maps. In the below example the engine is at 3288RPM which is using a window of 150RPM so it should be logging to 3200RPM not 3600RPM.

RPM vs MAP Off 2_sm.JPG (136.88 KiB) Viewed 204 times

RPM vs MAP Off 1_sm.JPG (110.21 KiB) Viewed 204 times

One error is X axis on the screen, it should be 20,30,40,50,60,70,80,90,100 but reads 20,30,40,60,80,85,90,95,100. This is why there is a 10kPa MAP shift


The other errors are in the bin so I checked the code and found the errors...



First for RPM...

Code: Select all

LAAE7:   ldx   #L8552         ; Logging 6400RPM To 9600RPM
      ldaa   *L006C         ;  RPM
      brset   *L003D,#0x08,LAAF5   ; Branch If RPM Is Greater Than 6400RPM
                     ; Else...
      ldx   #L84CB         ; Logging - 400RPM To 6400RPM
      ldaa   *L006C         ; RPM
                     ;
LAAF5:   tab               ; Keep A Copy
      anda   #0x0F         ; Keep The Last 4bits
      cmpa   L8203      ; Knock Count Logging -  RPM Window
      bls   LAB07         ; Branch If Inside Upper Part Of Window
                     ; Else...
      adda   L8203         ; Add The Window And Now Check Below
      cmpa   #0x10      ;
      bcs   LAB3D         ; Branch If Not Inside The Window
                     ; Else...               
      addb   L8203         ; Make It Positive
   
   
LAB07:   lsrb               ; \
      lsrb               ; |
      lsrb               ; | 400RPM breakpoints
      lsrb               ; /
      ldaa   #0x09         ; 9 MAP Points
      mul               ;
      abx               ; Add To The Table Offset



The low RPM table starts at 400RPM and not zero RPM, so this needs to be subbed off at the start.




The other error is logging at 100kPa causes the 20kPa cell to be populated.

MAP Not 100kPa 1_sm.JPG (88.07 KiB) Viewed 204 times

Code: Select all

ldaa      *L005A            ; MAP
      tab               ;
      anda   #0x1F         ; Keep The Last Last 5bits
      cmpa   L8204      ; Knock Count Logging -  MAP Window
      bls   LAB23         ;
                     ; Else...
      adda   L8204         ; Knock Count Logging -  MAP Window
      cmpa   #0x20      ;
      bcs   LAB3D         ;
                     ; Else...
      addb   L8204         ; Knock Count Logging -  MAP Window


After the 'addb L8204' instruction the code rolls past 0xFF if the MAP is greater than 0xFF minus L8204. This causes the roll to 20kPa so need to check for that and modify the value.





I am really hoping Ken is able to fix these up and release a new version, i use knock logging heaps and its caused hours of frustration trying to work out whats going on.

Thanks

[Hooter]

Thanks for that UNEEK, Im sorry I havent been here fro awhile and spotted that, its a buggar, as there is only so many hours in a day, lame excuse I know.......but I have sent this to Ken to check out.
Regards
Al

[Hooter]

Hey Uneek, Can you email me your dongle user number to tpiconn@space.net.au, and Ken will do you new update disc.
Regards
Al Gibbs

[VPSteve]

I have just read this for the first time and thought, 'that sounds familiar'. My SP3 has been in the car for about 4 or 5 years, and I've often wondered how it could log in areas that the engine wouldn't normally see. Would there be a series of dongle numbers that are effected by this or a batch of SP3's perhaps? If this is the case I would love to have it corrected if only for the sake of knowing it's right. I am happy with the product and it works fine, and have had the engine professionally tuned and logged AFR with wideband and all that.
Steve.

[UNEEK]

I think its in all SP3 variants (N03, 5, 6 etc) from what I could tell. I dont know if there are different versions of SP3 though so maybe not all are affected. I dont use it anymore after I found the closed loop boost control was broken too.