Engine runs, but config needs review.
Engine Type
- SR-20 DET, Nissan GTI-R, Pulsar in USA
"Nissantrigger 360+4" with 2,2,2,8 window-lengths
4cyl
- secondary trigger: 360 pulse
- HW divider was NOT used, simply connected the 360 pulses to sectrig HALL input.
- primary trigger: 4 pulse
- the length of pulses on this engine is something like 2,2,2,8 which DIFFERS from the original specs (the wheel we got and the patent) the 360 tooth trigger was made for originally . Primtrig polarity also differs !
- So we were lucky that we could still get 1.1.x firmware to sync: even without sectrig divby4 HW divider, we had to enable the "divby 4" flag in 1.1.98 config which meaned "sync to a primtrig tooth that is longer or equal >= 3 sectrig pulses" even though we did not use divider on sec
- the sectrig divider config setting is not used from 1.1.99, secignore is used instead (see below)
- was triggerlog captured with primtrig rising or falling ? (write rising or falling in the filename next time)
- we know now that primtrig=rising edge must be used to get an even signal (the only good setting).
Most likely edge related
- Either the choice of primary trigger edge (we know now: primtrig must be rising)
- or the assumption (from available descriptions and patent) at the time of implementation was wrong
- even with a real disk, the direction of rotation can be wrong and the polarity of nissan HALL was uncertain (actually wrong in the orig spec we received).
primtrig scopeshot reveiled that rising edge must be used, the primtrig=hi window length must be measured
Now that proper specs (eg. primtrig polarity) was given for Nissan [1.1.99 2012-03-22] has firmware mod to apply for all known types of 360 sectrig-pulse nissan triggers.
- WBO2 heater also good now (wbo2 heater didn't always start with 2012-03-15 experimental 1.1.99)
- vemslog running upto 9800 RPM: [1]
- polarity
- primtrig=rising must be configured of course
- firmware always count sectrig pulses when primtrig=high
- secignore to adjust window-toothcount threshold.
- recommended threshold values: >4, >9, >13 (13 was used on bench signalgen, eg. in above vemslog), so start with secignore=4 for this engine
- because max window can be 8 for this engine, 12 or 16 (for different Nissan wheels)
- in safe mode (eg prohibit byte screwed or ibutton not touched), corr.air shows the detected peak window-length+128 to aid the user count window-length without scope (VemsTune developers might make a wizard for it). Useful to find the good setting for an unknown engine/trigger wheel, or verify for an otherwise known engine/trigger wheel.
Important: ignition output might have slipped one entry in 1.1.99 with the secignore mod (strobe would show immediately)
- easy to compensate: VT, ignition outputs visual dialog, click shift up once, then burn
C103 consideration sectrig was traditionally low pulsecount, and C103=1nF might be an artificial HW frequency limit for 360 pulses
- emil suggested a firmware mod to allow neglecting nissan360 sectrig above certain (possibly adjustable) RPM.
- The c103=1nF with 2.7k pullup is appr 3 usec time constant, and signal peak amplitude at 19+21usec (8000 RPM) might not always reach the required high-threshold
- we definitely had to remove C103 in the benchtest tester v3 when testing firmware with the nissan 360 pulse sectrig (although that is triggered from a signalgen, not optical sensor).
- it is recommend to request sectrig c103=100 or 220pF in order comment for new nissan360 installs but that firmware mod would allow old boxes to work without HW change.