SVN rev. 635 had injector 100% duty related changes (although earlier versions tested fine earlier, extra conditions were added that must hold to switch off injectors.)
This change apparently affected ver2.x (the hardware before v3.x) injection => engine flooded (1.0.38 version working correctly).
OK solved it myself.\n
uint16_t pwcalc = inj_port->pwcalc;
uint16_t now = etime.now;
// note, that we do not change inj_port.stopat if we do not schedule
// so we don't disturb the turnoff of injector.
if (pwcalc){
uint16_t pwcalc, pwd, stopat;
pwcalc was introduced again/or not initialised correctly. Just removed second one. (overlapping, I guess it was a compiler warning anyway).
Pathch available at http://kotisivu.dnainternet.net/sakrkorh/fw/1.0.44/ ... as it is a bit longer than expected, can you explain a few things ?
- why wasn't it enough to add 1 ifdef so for v2.x the stop_now() (same as old COND_STOPAT()) is used, not the new injector_should_stop(), injector_should_start()
- any parts of the patch that can effect v3.x ?