|\ /| | ) ( | | (___) | | ___ | | ( ) | | ) ( | |/ \|
_ (_) | | | | | | (_)
## ###### ## ## ## ###
### # # ## # ## ### # # ####
__ __ \ \ / / \ V / / \ / /^\ \ \/ \/
IMPORTANT: enter the case-INsensitive alphabetic (no numbers) code AND WRITE SOME SHORT summary of changes (below) if you are saving changes. (not required for previewing changes). Wiki-spamming is not tolerated, will be removed, so it does NOT even show up in history. Spammers go away now. Visit Preferences to set your user name Summary of change: '''This page lists important changes to the firmware. Check this before downloading a firmware that is meant for production''' Normally updated by development team only. Unless it's something obvious or minor (eg. spelling), do not edit this page without reading this: * For install-related problems (like needing to upload firmware after bootloader upgrade) use your own MembersPage. * For generic (hopefully reproducible) issues, see IssueReports. * Feature requests moved to GenBoard/UnderDevelopment/FirmWare '''IMPORTANT: when uploading new firmware''' to a controller (even with a motronic55 connector!), to prevent filling cylinders with fuel or possible igncoil or igndriver damage: * '''remove ignition fuses''' * '''remove injector connectors and remove fuelpump relay''' ** or otherwise make sure the fuelpump is off, eg. hidden anti-theft switch ** Upload and verify Firmware ** Upload and verify config, and measure voltage in the injector connectors before powering the fuelpump ** Explanation: when table positions change inside EEPROM (as between 1.0.73 and 1.0.78), the h[0] and h[2] might get corrupted (might fill the cylinders with fuel). * to avoid filling cylinders with fuel (or other problem), '''after firmware upload, always upload config and tables''', reboot, verify, measure, than power ign and inj only if all is well. ** always upload config after firmware upgrade - even an unrelated old config from a very different setup is better than nothing. Do not even try to set everything manually without a config upload. It will take many hours or days, and it will not work ! You have been warned. It is of course recommended to go through all dialogs, but only after uploading a saved config. * '''but first of all: READ EVERYTHING BETWEEN OLD AND NEW REVISION''' and check if any of the changes affects your configuration. Take action if necessary. For example, if you used an 1.0.13 firmware, and upgrading to (eg. 1.0.23), when reading through all changes for intermediate releases, you will see (at 1.0.14) that fuelcut_min_kpa must be adjusted reasonably. If you skip 10 mins to read through, you'll sit in the cold car for hours and wonder why the engine does not start and notice injector pulsewidth is 0. Developers: please update GenBoard/Manual matrix when configuration semantics change. Ask on that page (in the relevant matrix element; make a line if necessary and start to put in links you find) for questions about config variables. '''You can report problems/bugs on this page:''' IssueReports ---- '''Users upgrading to 1.0.36 and above please note:''' There are some issues with the ALS configuration that may not be apparent when upgrading your configs from an earlier version. These are detailed MembersPage/GaborRacz/NewAlsLaunchAndOthers Windows users have an easy solution to solve this issue detaile at MembersPage/PhatBob/ALSConfigFix ---- '''Downloadable firmware releases'''. These are not intended for compilation. See SubVersionSvn if you want to get involved in development. Publish the my_make on your MembersPage if you just want a firmware compiled with some special options. The "Genboard Public Licence" allows modification of firmware (algorithimic changes, additions etc) if the modifications are sent back to authors. Much of the sourcecode was published, and everyone so far who sent modifications got read-write access to the firmware tree. Because of repeated (intentional or unintentional) violations of the copyright, developers decided to apply control over the sourcecode that allows well-intended use but prevents violation * "normal" customers don't (and shouldn't) need to modify firmware * anyone with skills to modify firmware who want to do so: apply for access. See SubVersionSvn . Don't forget to detail your plans ("I implemented / I want to implement ..., concept is ..., any simulations, or PC-testcode, etc...) ** the plans are necessary so the team can cooperate with the applicant to sort the necessary files/functions for which access is needed for the given task. The development team attempts to cooperate. Access to the chosen functions is granted after the development team decides that cooperation is likely to work out. * an NDA is necessary that allows modifications (algorithimic changes, additions etc) for self-use or use in reselled VEMS products, but does not allow carrying the code (or concept) to competitive products. '''Changelog''' - missing tooth trigger installs (N-1 or N-2 with or without camsync) and InputTrigger/AudiTrigger (135+ tooth wheel) users (especially with big-ignadv-window ALS), extrapulse triggers (some honda and homemade triggers) and all VVTI (camshaft-angle-control) and table-switching apps will likely prefer 1.1.x, while few still stay with 1.0.x for now (though this will change) ** 1.1.x always supported 60-2 trigger and supports dual config since 1.1.3x (read below). ** 1.0.x is unlikely to ever support TableSwitch '''IMPORTANT: update VT ini file''' before upgrade (especially if upgrading to a firmware newer than VT) and check validate after upgrade. * Without VT ini update (eg VT older than firmware) don't expect that it works (although in many cases it does work). * minor manual interventions might be needed (that will be automated in some future VT ini). ** Eg after upgrading a simple-coil trigger, if the extrapulse or ls4 config bit (these bits are "don't care" in earlier firmwares) is enabled it will give no spark. ---- '''1.1.97''' * Around when subaru 36-2-2-2 VVTI (camshaft angle control, AVCS) was implemented, the subaru 36-2-2-2 without camsync config setup was hijacked for some time for mitsubishi-evo (a setup made according to bogus specs we received, later remade with a different approach and config). Now from 1.1.97 gifted back to subaru (while factory subaru-s have camsync, which was not effected, it makes sense to run subarus with wasted spark setup without camsync: possible again). * high-current (powerful) unipolar stepper driven from 4 injector outputs. ** We'll think about how this (rarely used, but useful) feature to make the chance minimal that this annoys those who don't use it. Not compiled in by default (and needs special config: stepper+ iac-dual-solenoid configured) * triggerlog is now useful at higher RPM (eg. with 60-2 at >9k RPM), as the primary teeth are "compressed" (eg. only missing tooth is shown at high RPM, not every tooth) * more tolerant if sectrig is lost (or moved out of the allowed window) at high RPM especially if missing tooth primtrig signal is OK (especially useful with doublevanos / VVTI apps). * 360 tooth Nissan trigger might make it in (or possibly that goes to 1.1.98 ?) ---- [http://vems.hu/download/v3/firmware/ 1.1.96] - RELEASED * '''extrapulse triggers''' (with cam position measurement support for camshaft-angle-control) ** like some Honda 30, 30, 30, 10,20, 30, 30, 30, 30, 30, 30, 30, 30 crankwheel ** 120, 120, 40,80, 120, 120, 120 camwheel (typical 6cyl inline truck/powerplant engine) ** other extrapulse trigger examples: 60,120, 180 or 90,90, 180 (the simplest 4-cyl crankwheel). The latter can be configured as 4-1 also. * '''BMW double-vanos''' with 60-2 primary trigger ** 6+extrapulse MUST be connected to sectrig ! ** optional (but recommended): the other cam, 6 evenly placed pulses can be connected to the 3d trigger. The "exhaust cam" table is used for this, regardless of which cam the signal is actually from (is this from the intake cam in the factory app ?). ** Fero(BMW Tököl) needs to swap the intake and exhaust cam signal connections (so the 6+extrapulse goes to the secondary trigger not the 3d trigger). * '''RPM actuator''' ("shiftlight") semantics changed at very low RPM, because the intended hysteresys was non-intuitive for configured 0 or 100 RPM threshold: ** if configured to 0 RPM, then shows running (trigger pulses seen) / not running ** if configured to 100 RPM, then turns on above 200 RPM, turns off below 100 RPM ** if configured > 100 RPM then works as originally (+-100 RPM threshold) * with original InputTrigger/AudiTrigger "Hallgebersignal" is around the crankhome-VR pulse (VT/Tools/play trigger "c270" type: sectrig is every 270 primtrig pulse: every second crankhome-VR masked away by the HALL signal) ** in some installs like the camsync is short and so unfixably offsetted that it cannot mask away one of 2 crankhome-VR pulses ** from 1.1.96, '''if Motorsport/camshaft angle control/third trigger is enabled with auditrigger''', than one sectrig pulse expected every 135 primtrig pulse, and the '''camsync can be connected to the 3d trigger input''' (only available on econoseal pin if requested explicitely, since normally only useful for double-VVTI setups where exhaust cam is also controlled hydraulically like the BMW double-Vanos) ** pleasant side-effect: all 3 triggers show up in the triggerlog ("primtrig", "sectrig", "exhaust cam"). See example: in VemsTune analyze the triggerlog from [http://vems.hu/vemstune/sharingcenter/reports.php?cmd=view&key=Rx3Q7b download zip of this share]. Therefore with this setup it's even easier to diagnose if the crankhome-VR or the cam-HALL signal is faulty. Traditionally possible to diagnose by disconnecting HALL sensor (with cam-HALL inverter, pulling HALL input to GND also needed) ** for now, only tested with divby3 (not with divby27 or other dividers). Keep the 3d trigger disabled unless you really know what you are doing and ready to bench, strobe, etc... Just in case, VemsTune validate will warn if 3d trigger is enabled (usually not intended). ---- '''[http://vems.hu/download/v3/firmware/ 1.1.95] - RELEASED''' * '''improved wheelspeed filtering''' of noisy (prelling) wheelspeed signal (on both wheelspeed inputs) ** Sami provided a suitably noisy log (thanx!) from the mechanical reed-relay of his audi transmission, so we could implement better filtering that handles this case: ** Above 318 (km/h or mph according to calibration) it's considered prelling-noise and filtered. 255-318 is allowed (no overflow to 0), measured speed stays high. 0-255 is displayed normally (good signal was good earlier also). * For 1.1.94 and earlier, wheelspeed dependent launch-mode must be used when wheelspeed is used at all. ** even if all launchRPM values are same (say 4500 RPM) and speed bins are all 255 ** from 1.1.95 the no-wheelspeed dependent mode might work as expected (even if wheelspeed>=1) with just 1-RPM value configured ** it is still recommended to use the wheelspeed dependent launch-mode (which is more flexible anyway) when wheelspeed is used. * SD-card read-retry number bumped up so ** '''some really slow cards that did not work earlier can now be used''' (the standards does not specify max delay) ---- '''[http://vems.hu/download/v3/firmware/ 1.1.94] RELEASED''' * [http://vems.hu/manual/vemstune/help2/htmls/v3/v3_transbrakecreep.html Trans Brake Creep] ** feature was implemented in 1.1.90 but now TRANS_BRAKE_CREEP enabled in default my_make (available to every VEMS user without special request) * logging will restart if SD-card is removed and (usually another) card is inserted. ** drag and rally racers requested this to restart SD-logging within about half minute after SD-card swap (even if engine is not stopped) ** this feature was implemented in 1.1.92 and worked well for SD-card swap, but if the SD-card was removed but another card not inserted (or SD-card logging was enabled with card not installed at all), the reinit attempt disabled interrupts for milliseconds (not microseconds), causing occasional trigger-error. Fixed in 1.1.94 : Now enabling SD-card does not cause trigger-error even if card is not installed (This cannot cause trigger-error in <1.1.92 since it does not try to reinitialize the card again and again, only at startup). * if priming pulse was configured non-0, than actual priming pulse was min 6 msec (range: 6.1 to 31.5 msec) ** new range is 0.1 to 25.5 msec ---- '''1.1.93''' - was never released, now withdrawn, see 1.1.94 notes * [http://www.vems.hu/vt/help/v3/v3_camshaft_angle_control.html exhaust cam control] also supported with 2 solenoid "Vanos" mode (not just with PWM "suzuki"-mode) * in 1.1.92 and 1.1.93 disable SD-card logging if SD-card is not installed or removed (otherwise trigger-err might show up every 32 second). ** If SD-card is enabled, it is recommended to use >=1.1.94 or < 1.1.92 so this cannot effect even if SD-card is unplugged or disconnected ---- '''1.1.92''' - was never released, now withdrawn, see 1.1.94 notes * InputTrigger/SubaruThirtySixMinusTwoMinusTwoMinusTwo supported with '''camshaft control of both left and right camshaft''' ** A (right, intake) connected to secondary trigger input ** and B (left, also intake but connected to "exhaust cam input"). ** please note (if upgrading from a working pre-1.1.92 subaru config) that tooth0 got 30 crankdeg delay to the "4BTDC10" point so: trigger_tooth=02 (instead of 3) and TDC delay=120 deg - verify with strobe ! * suzuki cam position (camshaft angle control) also supported ** both are tested with signals that match scopeshots (and also fluctuated a bit to allow reasonable deviation). Need real-engine tests (experienced installers). * MembersPage/Fero/MitsubishiEVO supported * iac_close_minmap can be set to any reasonable (non-FF) value, will not effect IAC flag * in 1.1.92 and 1.1.93 disable SD-card logging if SD-card is not installed or removed (otherwise trigger-err might show up every 32 second). ** If SD-card is enabled, it is recommended to use >=1.1.94 or < 1.1.92 so this cannot effect even if SD-card is unplugged or disconnected ---- '''1.1.91''' - will not be released * Anytrim controls expanded with speed sensor inputs as trimming source * GPS speed sent out to Round on AIM protocol * in 1.1.91, leave iac_close_minmap at FF (510 kPa default value) - or use 1.1.92 ** otherwise iac flag might be left on in some cases, possibly effecting sparkadv (due to iac retard) ---- '''1.1.90''' - will never be released * Iac can force to specified duty when reached the specified Map (usefull for Boost) * Injector dead time has a simpler strategy * Triggerlog logs the Spark events * Maserati trigger support: MembersPage/OddFireSixCyl * Staging injectors can switch with RPM limits ---- '''1.1.89''' * v3 can now log GPS data to SDcard ** when vemstune notebook is not connected, a GPS can be connected to comm-port (our GPS is 4800 baud, which, unfortunately is different from the standard 19200 baud aim-default baudrate) ** anyway, time information is no longer lacking for logfiles ... this way GPS replaces RTC ** second serial input on V3 also can use for GPS ** for details see: GenBoard/LoggerIntegration/MMC * MembersPage/DamirMuha/MagnetMmarelliCamSensor support ** new sectrig setting: to neglect the offending pulses from 3 (or more) sectrig pulses * the new vemstune (coming out late Febr) syncs superfast to 1.1.89 ** VT recognizes page-data checksums and only downloads changed pages (like rsync). ** But initial config-download is much faster also * Secondary trigger input can use as a stand-alone primary trigger ---- '''[http://vems.hu/download/v3/firmware/ 1.1.88] Withdrawn !''' * withdrawn because compiled with KNOCK_ALTERNATIVE that seems to cause problems (ign advance different than table) under certain circumstances. '''Upgrade to latest released version is highly recommended.''' ** the actual ign-advance is in the log ** if the upgrade is expensive, we'll benchtest any config on request (best to use MembersPage or VemsTune Help to post vemslog or vemscfg) and report if the test shows it is effected, and upgrade the config to latest release, test that too, to make upgrade a simple "connect and upload" type 4 minute operation. * '''Rover K series 36-1-1-1''' type MembersPage/YasecElise trigger supported ** from "primary trigger" dialog upload configlet "4cyl 36-1 no camsync", press validate. Set number of tooth to 32 (not 35). Press burn. You just configured "elise trigger". From tooth_wheel=20 (decimal 32) the firmware will know it's not 36-1 but elise type trigger. However, 2011-01-24 vemstune configcriteria rules are not yet prepared for this special case, and might fire false warning (so reftooth 0 18 0 18 is good). ** or use "rover-k_lotus-elise_e324" configlet ** "e324" type generates this pattern (Tools/Play trigger) * compiled with KNOCK_ALTERNATIVE defined (so the knock amplitude threshold is configurable per loadsite). * the wheelspeed (non-PS2) version also sends fuel consumption over serial to AfreshTiny/AimDisplay if set up properly (eg. wheelspeed must be calibrated) * PID boost overflow bug (injected after 1.1.81) fixed. * 4+1 (coiltype 2 pins on crank + 1 pulse on cam) worked on bench but RPM=0 on engine (even though inputs were perfect in triggerlog), [http://vems.hu/files/Sanci/ vemslog files here] ** we don't know if auditrigger divby1 workaround works (because the starter on the engine failed just when we wanted to try this). Investigating the issue '''[http://vems.hu/download/v3/firmware/ 1.1.87]''' - release candidate * max 1.1.85 recommended for now. '''If you really need 1.1.87 (use nonps2 and), use newest VemsTune''' * '''see SafetyMode''' - just in case: be prepared to set ECU calibrations / prohibit manually (one time operation. Should be automatic in new vemstune, but better be prepared than surprised). * we had a few positive reports with VT NIGHTLY VemsTune-Install-2010-12-22-commModel.exe (including triggerlog), but ... '''1.1.86''' - unreleased, internal version '''[http://www.vems.hu/download/v3/firmware/ 1.1.85]''' - '''released''' * auditrigger '''divby2 changed to divbyANY''' ** toothrel_missing=01 for divby2 ** toothrel_missing=19 for divby26 ** toothrel_missing=1a for divby27 ** with auditrigger 11000 RPM could be reached earlier (in benchtests) using divby27. Not 25500 RPM as with lower toothcounts, but sufficient. ** note: auditrigger divby3 (which is used in hundreds of 135 tooth audi engines) maintains compatibility '''[http://www.vems.hu/download/v3/firmware/ 1.1.84]''' - '''not released. Benchtests first before uploading to the real thing''' * anti theft iButton function, see ButtonImmobilizer for details (also related to '''anti-flooding SafetyMode''' in 1.1.83) * Wheelspeed measures faster (lower time-constant used in internal software filter) ** '''second wheelspeed input implemented''', see: InputTrigger/WheelSpeed * New LCD page (page 10) ** I have been running/driving on 1.1.84 all day today, it is running well. I cannot test speedsensor response speed (my car is too slow!) until I install in race car this weekend. KevinBlack ** false-positive trigger-"errors" under certain conditions. Operation not effected. '''[http://www.vems.hu/download/v3/firmware/ 1.1.83]''' - was used for VIP-testing. '''Will not be released''' * '''anti-flooding SafetyMode implemented for firmware upload.''' * config/table switch modifications, see [http://www.vems.hu/manual/vemstune/help2/htmls/ VemsTune help ] for more details [http://www.vems.hu/download/v3/firmware/experimental 1.1.81] - '''Released''' * got many positive feedback. ** upgrade properly (always upload config after upgrade - even an unrelated old config from a very different setup is better than nothing). Do not even try to set everything manually. It will not work ! * easily configurable injection angle * better trigger-error indication ** we'll work for more graceful handling of trigger errors (related to injection-angle) ** some false-positive trigger-error got introduced in 1.1.80 when the new sectrig window was reworked, not fully fixed even in 1.1.81. It only effected the reported flag, not actual operation. Sorry about the inconvenience. Only report such false-positive if also happens in 1.1.84 or newer. * VVT (variable valve timing) reimplemented (not using anything from the old code). Only 1 control loop now, but target table reserved for 2nd control loop too. ** only for missing-tooth+camsync trigger ! Eg. coiltype or subarutrigger not (yet) supported [http://www.vems.hu/download/v3/firmware/experimental 1.1.80] - NOT released (1.1.81 or newer is recommended) * easily configurable '''injection end-angle''' (angular position when the injector output switches off) '''in function of RPM''' ** mostly matters at low load loadsites with aggressive cam engines. With a port-injected audi 20-VT (with cyl1 injector output in same position as cyl1 ignition output) smoothest idle and best throttle-reaction was achieved around 300 ATDC = 420 BTDC (was not very sensitive to precise adjustment, a window of at least +-30 degrees) ** secondary-trigger need to be configured as cam-sync for inj-angle to work properly. Without camsync (eg. with a missing tooth wheel) after startup, the chance for the sweet half is 50% at best. [http://www.vems.hu/download/v3/firmware/ 1.1.79] - '''will not be released''' * better low-RPM cranking behavior. Display calculated RPM (reading not stick to 100) well below 100 RPM. Also lowered min RPM for expected behaviour output actuation. Ign and inj outputs off below that. 28-70 RPM for most triggers (if signal is reasonably good). The real minimum RPM also depends on cylinder count, HW setup and trigger type and setup. ** the price is +1/2 crankrot to get initial sync in most situations. * '''config12=20''' (bit5=1) means wheelspeed is sent in km/h in the AIM packet ** config12=00 means wheelspeed is sent in 0.1 km/h (VEMS internal value *10) as in the [http://www.vems.hu/files/MembersPage/NanassyPeter/AIM_support/AIM-ECU%20protocol.pdf aim specs] . Fero says that unfortunately the AIM display does not follow the AIM specs, so AIM display shows 10 times higher than the sent value interpreted in 0.1km/h ** to enable/disable this divby10 one needs to upload small config12=20 configfile manually, or use 2010-06-28 or newer VT (or apply http://www.vems.hu/files/pacsa/aim_wheelspeed_patch.zip ini files to recent VT) * known bug: no TPS update under certain circumstances (harder to reproduce in real engine, easier on bench) [http://www.vems.hu/download/v3/firmware/ 1.1.78] - '''release candidate''' * note: coiltype withOUT camsync (c008 distributer) that didn't sync earlier also passed tests * '''stop the coolant-fan during cranking''' if RPM < cranking threshold * Subaru 36-2-2-2 support with camsync InputTrigger/SubaruThirtySixMinusTwoMinusTwoMinusTwo ** the code was backported to experimental/1.1.76 (see 1.1.76 below) * InputTrigger/TriggerLog changed so '''secondary trigger pulse timestamps''' (if secondary trigger is enabled) are also sent through the serial port to help installing and verifying camsync+crankwheel setups (like the subarutrigger that has 2 low-toothcount VR wheels). ** '''CRC also added''' so higher than 19200 baud can be used => not just for cranking log, but multiple thousand RPM can be logged as well. 19200 baud limits triggerlogging a 60-2 wheel (with or without camsync) to 900 RPM, and the 135 tooth auditrigger (45 teeth logged) to appr 1200 RPM. (ManmcB08mde40) 115200 baud allows much higher RPM (with 135 tooth auditrigger triggerlog is not recommended above 7000 RPM anyway). ** to support the sectrig timestamp and CRC, the format is incompatible with old format. The old tools will not work (mde40 can be captured with brayterm, but uptodate VemsTune needed to analyze the file or VT can also read directly from the controller). Actually the format was changed twice, now it allows other type of timestamps for future extensions (like ignition events) ** for the curious: mde40 output has 0x7E marked frames. Frames need unescape (0x7D,0x5D => 0x7D and 0x7D,0x5E => 0x7E). Each frame has N* primtrig + <optional sectrig> + 1byte type + 2 byte CRC (+ 0x7E frame-marker). With type=2 the sectrig is there (2 byte big endian timestamp in 4 usec resolution), while type=1 lacks sectrig timestamp. ** VemsTune Tools / Record, Analize Triggerlog is being prepared for displaying the additional data (from captured file or real-time data from controller) ** if the missing tooth advanced filter (primary trigger bit3=1) is enabled with auditrigger (c270 pulsetrain), than div-by-27 is applied instead of divby3. With this it's internally not a 90+1 any longer but a 10+1 trigger with tooth width=72 degrees, another_trigger_tooth=02 and h[1]=00 08 06 04 02 ... See 11500 RPM auditrigger log on InputTrigger/AudiTrigger (while most triggers have 25500 RPM limit, traditionally the 135 tooth auditrigger had sub-10k RPM limit, now beyond 10k RPM). [http://www.vems.hu/download/v3/firmware/experimental/ 1.1.77] - will not be released * unfortunately with the original HALL-dirac implementation missing-tooth + camsync did not syncronise at all (fixed in 1.1.78 experimental 2010-06-21) * using matching vemstune ini-s is absolutely essential. The new functions will cause serious problems with random values (think about injection and ignition trims). Review all the related dialogs carefully ! * '''individual cylinder fuel pulsewidth adjustments''' ** '''multiplicative:''' most effect at high load. Greater or less than 100% allowed. Can be used to compensate slight deviations in flowrate, or cool the 5th cyl of a BMW I6 a bit (the hottest and most sensitive of all, closest to detonation point) ** '''additive:''' Most effect at low load / idle. Before tuning a new engine, set it to 52 usec so individual shortening can be applied too. (After upgrade use 0-12 usec for compatibility) * '''individual cylinder ignition sparkadvance trimming:''' Decreases spark advance for each cylinder with a given amount (default 0). * tach_div=1..5 can be used for non 50% duty for special tachometer gauges (some Nissan?). ** Lower than 50% normally, but you can invert the output. ** Output active and inactive for N and 32-N units of time ** It follows that for example tach_div=4 means 4/32 = 12.5% duty * '''up to 3 anytrim inputs''', that can apply adjustment to '''fuelpw, lambdatarget, boosttarget, boostpwm and ignadv'''. Turn the feature off unless you really know what you are doing. * '''Trailing spark support for any trigger''' (except fiatstilo/subaru "shortgaptrigger"), '''usually InputTrigger/RotaryTrigger as in Mazda Wankel engines''' ** reftooth=128 means time from the same tooth as the previous event, just with some delay ("trailing spark"). The exact delay is configurable in function of RPM. See InputTrigger/RotaryTrigger * '''HALL dirac filtering'''. config12 bit7 and bit6 ** Highly experimental. Can filter very short spikes on HALL trigger input (usually from capacitive coupled noise from ignition wires - also depends on harness arrangement). ** only enable it for low toothcount HALL trigger, if the pulse after the chosen edge is min 10 crankdegree long (or min 180 usec) and if you have mysterious trigger errors that you cannot get rid of, even with properly shielded trigger cable. ** '''Must be disabled for VR trigger'''. Useless with VR trigger anyway (the LM1815 makes a constant pulse anyway if the signal goes below 0.0 V ... and that must not be filtered out for the trigger to work). ** like for any VR, '''HALL-dirac filter must be disabled in both primtrig and sectrig with InputTrigger/ShortGapTrigger ''', InputTrigger/SubaruTrigger, (MembersPage/FiatStilo fiatstilo and subarutrigger). ** config12=80 bit7=1 : HALL-dirac filtering for primary trigger ** config12=40 bit6=1 : HALL-dirac filtering for secondary trigger (only very short pulses can be filtered on sectrig, even a 3-5 usec pulse might be accepted as a normal pulse) ** config12=C0 HALL-dirac filtering for both trigger * warning: the old hybridRPM based blending is gone (will not be supported in future: the new method is easier to understand and tune) ** If one uses MAP sensor at all, he can set up kPa based switching point for lambdatarget, ignadv, etc... Eg MAP used for table lookup above 120 kPa, TPS used below. Keep threshold=0 kPa unless you know what you are doing ** there is no such single switchpoint threshold for VE, because mapcorr is used for fuel pulsewidth with adjustable bins * bugfix: iac stepper sync ("initial step-dance") was started too late (not at powerup but during cranking) for quite some time (also in 1.1.75 !). Stepper iac users will likely prefer 1.1.78 or newer. '''[http://www.vems.hu/download/v3/firmware/experimental/ 1.1.76]''' - not released * subaru 6+7 (toothcounting implementation) and subaru 36-2-2-2 and the newest secondary-triggerlog (normally available from 1.1.78) was backported on top of the stable 1.1.75 ** this 1.1.76 is only for those subarus, no advantage at all for any other triggers * note: Jorgen had a private 1.1.76 with HALL-dirac for his problematic noisy 24+1 nissan which is not included in this. Any new installs that really need HALL-dirac filter can only use 1.1.78 or newer. '''[http://www.vems.hu/download/v3/firmware/ 1.1.75]''' - released when 2010-05-xx VemsTune is available for it. * CRC protected runtime comm for the "D" command. (firmware upgrade was protected earlier too) '''[http://www.vems.hu/download/v3/firmware/ 1.1.74] - released''' * more precise tach output frequency even at very high RPM ** up to 36000 output pulses per minute, that is 18000 RPM for "4 cyl" tachometer (which can be used even for 5-6-8-12 or other cyl count, just adjust the tach divider) * '''LCD updates are synchronized to LCD refresh''' (when lambda changes from 1.00 to 0.99, 1.99 should not be intermittently displayed). Disable "forced goto" in LCD menu for higher refresh rate * cranking pulsewidth shown after powerup (even before trigger) '''[http://www.vems.hu/download/v3/firmware/ 1.1.73] - not released''' * improved ignition when trigger signal is extremely noisy during cranking ** at least for missing tooth trigger - not much can be done for "coiltype" trigger of course. '''[http://www.vems.hu/download/v3/firmware/ 1.1.72] - not released''' * InputTrigger/SubaruTrigger implemented. Unlike with 1.0.x, configured as coiltype (not missing tooth type). primary_trigger=A3 ** (or AB, as the missing tooth advanced filter can be enabled and configured for this, even though derived from the coiltype family. Than we trig from the short "normal tooth", max 100% and neglect the long "missing tooth" eg. min 125% ) ** the weird subaru multipulse camsync requires another_trigger_tooth=03 and secondary_trigger=59 (bit6=1) ** for the simpler fiatstilo campulse, another_trigger_tooth=01 applies and either secondary_trigger=59 or 19 should work. * InputTrigger/LanciaCosworth (which was "coiltype" before, theoretically not changed, must also be retested.) * some small other changes, eg. warninglight sent instead of warmup (warmup light on when warmup_enrichment > 100% is calculated in vemstune) * for launch, flamebit (a config bit) enables launch_enrich fadeout (partial sparkcut is good, but partial launch_enrichment will be normally disabled by most tuners to avoid exhaust bang and flame effects) '''[http://www.vems.hu/download/v3/firmware/ 1.1.71] - not released. For experts only''' * als_iacpos;// iac position when ALS active (0 disables) * communication protection against unauthorized (write / read) access ** (perhaps only in 1.1.71 : not certain if this will be available in the default compile after 1.1.71 ) ** only applied if config_ext.pwcommlock0 == DE ( decimal 222 ) ** than serialnum dependent password is needed for communication. Request password for the given serialnr when ordering. ** Use GenBoard/Firmware/BootLoaderLoopback to upload an old firmware if you accidentally locked yourself out (that's why this is for experts only). ** no need to reenter password if reconnecting; Only needed after reboot (might be asked for only once on a certain notebook: if VemsTune saves it on the notebook harddrive and sends it automatically after that) '''[http://www.vems.hu/download/v3/firmware/ 1.1.70] - RELEASED''' * important scheduler priority changes (effecting ign and lambda_target lookup) * This also means that '''1.1.6x will never be released. Any 1.1.6x MUST be upgraded - with the usual care of course.''' * new VemsTune required '''[http://www.vems.hu/download/v3/firmware/ 1.1.69]''' - NOT RELEASED * '''anytrim: 17 entry array for compensation (78-178 % range for fuel injector pulsewidth)'''. Useful for '''dynamic barometric BARO compensation''', LPG temperature and pressure compensation, even for nonlinear EGT enrichment or ambient referenced Fuel Pressure Regulator compensation. ** config_ext.anytrim1 bit7:4 F: disabled. 0=trim fuel pulsewidth. *** 1..E values reserved for later when lambdatarget, sparkadv, boosttarget will also be option. Only use F or 0 now. ** config_ext.anytrim1 bit3:0 : *** 0..7: analog input=mcp3208 ch0..7 Eg 115 kPa absolute sensor connected for BARO-comp. *** 8: MAP (calibrated, 0-1024 kPa). In naturally aspirated, alpha-N mode the internal (or EC18/6 connected external) MAP sensor can also be used for BARO. 115 kPa sensor or other can be used. *** 9-15: 700 kPa pressure sensor (slope=778 kPa / 5V) connected to mcp3208 ch1..7 (ch0 not available for this). Pressure-MAP is calculated internally, giving the differential_pressure (across injectors). Useful when (eg. limitation of some LPG setups) ambient-referenced or absolute pressure sensor used, instead of (MAP referenced) differential pressure sensor. Most recommended fuel pressure setup (in this order): * '''Fuel pressure REGULATOR referenced to MAP''' ** as almost always, by the way. Unfortunately some LPG kits are ambient referenced * '''Measure fuel pressure relative to MAP''' (with a 2-nipple differential pressure sensor) ** than use the simple 0..7 mode. ** this is the easiest measurement to give an early warning when fuel pressure drops (because fuelpump cannot handle the flow, fuel filter dirty or fuel pressure regulator reference hose forgotten or blown off by boost) * '''Measure fuel pressure relative to ambient''' ** either with a differential pressure sensor (with reference nipple = ambient) or "gauge" type sensor * '''Measure fuel pressure relative to vacuum''' (with an absolute 700 kPa sensor) ** same mode (9..15) as above, just the 17 point curve X axis is not 45-781 kPa but -45..681kPa. This mode is NOT recommended. If you really insist, subtract 100 kPa in your head. '''[http://www.vems.hu/download/v3/firmware/ 1.1.68]''' - not released * config_ext.iac_conf2 bit 4 enables wheel_err warning light for 1.6 second ** trigger error light in vemstune lights up for 1.6 second regardless of this config bit * minor change: accel retard (if non-0) was "sticking" longer (after the accel enrichment decayed) '''[http://www.vems.hu/download/v3/firmware/ 1.1.67]''' - experimental * see BoostControl : consolidated the plurality of strategies '''[http://www.vems.hu/download/v3/firmware/ 1.1.66]''' - will not be released * cranking and afterstart changed (again). Max value is 1694 (+1594%) to allow E85 in extremely cold (possibly without batchfire-all-injectors during cranking). This is on top of warmup enrichment (and possibly higher cranking VE than normal VE). ** write the old enrichments to paper, and look up the real injector pulsewidth in logs as well - take note. ** set the enrichment values in config so the real injector pulsewidth is as desired. ** there is currently no automatic upgrade of the iac values. '''[http://www.vems.hu/download/v3/firmware/ 1.1.65]''' - not released * minor iac and boostalt changes (no intended change in semantics) '''[http://www.vems.hu/download/v3/firmware/ 1.1.64]''' * afterstart enrichment change: got smaller and smoother. ** note that for first few events, higher than configured afterstart enrichment is applied (like '''cranking enrichment, afterstart is also ON TOP OF WARMUP ENRICHMENT''' - so rather small values are best, often around 10..20% but this varies). Afterstart enrichment decayes to configured value in about 30 engine cycles. Than it continues to decay, becomes 0 after the configured number of cycles. Always '''consult pulsewidth in your log''' when tweaking to verify that your understanding of settings matches actual behavior! ** also remember that cold-start behavior depends on cranking, afterstart and iac-duty. Usually not trivial to tell what to change and what direction, so experiments are needed (use pencil or a laptop to take notes when your pen freezes ;-). Surprisingly in very cold, some engines start better with relatively low iac-duty. '''[http://www.vems.hu/download/v3/firmware/ 1.1.63]''' - withdrawn. * many changes, like '''cranking and afterstart table separated''' ** for E85 the cranking fuel amount climbs steeply as CLT drops around 27C, because mostly just the gasoline part burns during cranking. The alcohol part also burns right after the engine has started. ** afterstart enrichment does not need the excessive amount of E85 like cranking * '''INJ STAGE2 now has an additive offset''' (+-2048 usec) to be able to smooth fuel quantity around the transition (something is very wrong if more than +-300 usec deviation needed: likely the relative size of staged inj). ** Inj staging also has some (fixed) hysteresys now. * more indication flags: ** Realtime byte 47 bit0 is N2O active ** Realtime byte 47 bit4 is INJ STAGE2 active * throw away old copy if you downloaded on 2009-11-25 and re-download: ** The PS2 version was actually too big to fit when 1.1.63 was uploaded for some testers before release. WARNING beep/light output was disabled on 2009-11-26 and vems.hex that actually fits was put into the zip. '''[http://www.vems.hu/download/v3/firmware/ 1.1.62]''' PS2 version released. (tested on more cars than the non-PS2 = wheelspeed version). When VemsTune is released for 1.1.63, from than on 1.1.63 will become the supported version (and 1.1.62 not). * Under VemsTune defaultfirmwareconfigs directory copy 1.1.61 to 1.1.62 before uploading config (saved from 1.1.x or 1.0.x) to assist you. Review everything before starting. set Primary trigger/tooth width and refernce tooth table. Patience and lots of care needed ! * tachout inactivated (so gauge should not flicker) when RPM=0 * old GM and '''kickstart motorbike trigger''': for coiltype trigger if trigtooth=0 and ign_tdcdelay <= 42 degrees (normally 10-15 deg in this special case), and ignadv > ign_tdcdelay, we allow spark timing from prev tooth (not just dwell timing). ** For precise cranking you place coiltype trigger 15..10 BTDC ** during cranking, if (eg.) ign_tdcdelay=10 degrees and crank_advance=7..10 degrees (too close) dwell will start immediate from the trigger pulse, and spark might be slightly late (timing dwell from a long distance from prev tooth during cranking would be bad idea because of the high relative RPM variations) ** ign_tdcdelay in the 20-50 degrees range not recommended, but should work '''[http://www.vems.hu/download/v3/firmware/ 1.1.61]''' testing * decrease dwell at high RPM for a target of DWELL_MIN_INACTIVE time (defined as 400 usec). Automatically detects if same ign output channel is configured in neighbor events (distributer), 288 deg apart (2/5 of wasted spark 5 cyl), 360 deg (normal wasted spark) or other reftooth distance * the advanced '''conditions for misc1 and misc2 MUST be set''' (do not forget to set them, otherwise these outputs could be effected in an unexpected way) * for the '''PS2''' MY_CONF += -D EXPENSIVE_MISC1OUT -D EXPENSIVE_MISC2OUT ** note: -D KNOCK_ALTERNATIVE is not available in this. PS2 keyboard, time-based launch are available of course. * for the '''NON-PS2''' MY_CONF += -D EXPENSIVE_MISC1OUT -D EXPENSIVE_MISC2OUT -D KNOCK_ALTERNATIVE ** so wheelspeed, wheelspeed-based launch and the experimental knock-threshold-map is available. '''[http://www.vems.hu/download/v3/firmware/ 1.1.60]''' not released. * '''low toothcount wheels like 4-1 (m031) and c004, c008 supported, without camsync too''' For 4-1, use h[1]=00 06 04 02 00 00 00 00 wether with or without camsync ("camsyncpretender" inside). * lambdatarget table can be read-written directly too - without loss of precision (or with the traditional ugly hack required for megatune lacking 1/x conversion function: some values cannot be represented this way) '''[http://www.vems.hu/download/v3/firmware/ 1.1.59]''' released for testing.. Seems quite good so far. We expect to release this (or newer with no new functions just minor changes and bugfixes) * '''dwell_limit: if dwell is configured too high''' (at actual dwell base value, RPM, VBATT and MAP) it will be '''automatically limited to allow appr 400 usec coil inactive time''' (similar as it has been in 1.0.x for ages). Most significant for distributer ignition 5..8 cyl. '''[http://www.vems.hu/download/v3/firmware/ 1.1.58]''' released for testing. * supports low toothcount missing tooth wheels like 4-1 crankwheel without camsync (the simplest type of trigger that allows DIS ignition). See 1.1.60 note * supports low toothcount "coil-type" wheels like 2..3..4 tooth on crank without camsync (with distributer of course, as this is not sufficient for DIS). * supports InputTrigger/LanciaCosworth, good at very high RPM and also seems good at cranking initial syncup (required special code for this trigger) * Unsure about the very special fiatstilo/subaru-trigger though. Benchtesting is essential before using with such trigger. '''[http://www.vems.hu/download/v3/firmware/ 1.1.56]''' released for experts only. Seems quite good though. * launch light (flag) is on when launch button is on (not when launch function is on): better indiction, no need to wait before launch-RPM is hit ** smoother decay of launch-retard as RPM goes down or MAP goes up. This provides some regulation: way too high boost during launch can be prevented. * most recent VemsTune needed. Menus are arranged for ergonomy. Firmware menu is under tools (rarely used anyway) * upload [http://vems.hu/files/MembersPage/MarcellGal/Audi/etc/Boost1.1.53_tests/ working config] before uploading your (eg. 1.1.27) config. This way the missing variables are more likely to be initialized to sane values, and less likely to cause headache and problems. * We found that we need '''low map_dwell=03''' or lower (possibly 01 or 00) because '''too high (>4.5msec) dwell''' due to high MAP on a turbo-audi '''could froze the controller (repeatably)''' in a case (due to harness, ground and supply-related issue, not software) with an audi 3B harness modified for DIS ignition. * 2 (none-missing, even spaced) or 3-1 (missing-toothed) cranktooth without camsync are unfortunately not yet supported in 1.1.x (some missing sparks at certain RPM), but we hope to support them later. Use 1.0.xx in the meantime. '''[http://www.vems.hu/download/v3/firmware/ 1.1.54]''' not released * Note: 1.1.53 uses a different lambda (runtime vars, offset=10) value above 1,0, (eg. not compliant with 20.9 reading at 20.9% O2). Higher resolution around 1.0 though. Likely we're forking 1.1.53 and 1.1.54 vemstune.ini and vemsv3.ini to make it work with both. * PS2 version is more thoroughly tested ** non-PS2 (wheelspeed) version will see some changes later around wheelspeed-launchRPM (and launch-flag). You can (should?) '''continuously activate launch''': RPM-limit will be adjusted automatically according to table as wheelspeed goes up * requires [http://www.vems.hu/download/v3gui/ VemsTune 2009-04-24] or newer. No known megatune yet. Andrey, can you publish your 1.1.5x vemsv3.ini for megatune ? Pls, make a Megatune usable for this firmwares! Race season is ON. MembersPage/GintsK * '''time dependent launccontrol RPM''' (PS2 enabled firmware only): '''after''' launch button is released and '''X time passed''' (X=first entry in speed/time table, 64usec unit), there is '''no enrichment and no retard''', but RPM-limit (partial ign-cut in the softrpm-control-range) can be specified in 8 time => RPM steps. ** also tested with shiftcut and launch on same (shared) analog (activation) inputs. ** Set all 8 entries to revlimit or to 25500 RPM if not using this feature. ** launch will show as active in runtime flags even '''after X time/speed is passed''' (X=first entry in speed/time table), but there is '''no enrichment and no retard'''. Launch flag means that RPMlimit is controlled according to launch. ** wheelspeed dependent launchcontrol RPM-limit (non-PS2 firmware only, where PS2-CLK is used for wheelspeed freq 0/5V signal). You will want to use the wheelspeed input from a non-driven wheel. '''1.1.52''' testing only * almost seemed fine, but MembersPage/MarcellGal/PowerAndTraction engine was flooded with fuel. Not during upgrade, but after it seemed to be running well. After removing sparkplugs the engine runs well again. Under investigation (if it's misconfiguration, eg. injector staging related, or. sg. else). We could not reproduce it either on the bench or on car. * The first 1.1.x firmware that supports coil-type trigger without camsync (eg. distributer) '''1.1.51''' *MAT-retard (MAT, TPS). *MembersPage/Andrey/Cranking *MembersPage/Andrey/KnockEGTRich *MembersPage/Andrey/FuelFilm *Gearboost use same RefDC as PID, so BoostDC table could be used for altboost out when speedsensor FW used. '''[http://www.vems.hu/download/v3/firmware/ 1.1.50]''' - testing only * coolant range changed (in 1.1.4x) to -100..+155C . See GenBoard/Manual/Config/CoolantTemperature * [http://vems.hu/download/v3gui/ vemstune 2009-03-05] or more recent version needed for upgrade and tuning => vemsv3.ini boost (and other) dialogs updated ** upload a [http://www.vems.hu/files/MembersPage/poweredracing/VolvoS605cyl/v3.3_n001855-2009.03.10-01.46.52.vemscfg working config] before uploading your (eg. 1.1.27) config. This way the missing variables are more likely to be initialized to sane values, and less likely to cause headache and problems. ** lotsof new variables when upgrading from 1.1.27, lotsof room for error. Take care and review everything, preferrably on the bench (don't forget the staging, n2o, MAP filter, configurable coolant bins in warmup enrichment, and other dialogs in vemstune). Apply caution. Fuelpump off, injectors off, igncoils no supply etc... when upgrading, until all dialogs and variables are verified. * the '''standard firmware also has 16x14 VE table''' (the good old and practical 12x12 VE table is not supported any longer). ** '''my_make.16x14/vems.hex also has 16x14 VE size, but has no PS2KEYBOARD support'''. It supports wheelspeed input (0/5V HALL sensor connected to PS2 clock signal, aka "INT6") *'''If you use speedsensor (PS2 undefined) and boostcontroller''', another calcs used (named gearboost), and it cannot be used with altboost at the same time, because it use ref table as refpos for PID, will be fixed at 1.1.51 * boost_channel is always the output actuator of the PID boost. ** boost PID P value internally multiplied by 4, so change P to quarter its previous value. Earlier P=255 had to be used for many setups - if yours too, start with P=64..70 or so. ** warning: '''boostalt simple_closedloop uses TPS scaling array for different purpose''' (not for boosttarget, but scaling output DC). Therefore when using boostalt simple_closedloop, PID boost will not use TPS scaling (assume 100%). Simpleboost closed loop works by taking refdc value from rpm/map table, downscales it if 8kpa over boosttarget graph and then scales refdc directly by tpsboost graph. BoostSW is for downscaling whole refdc table. Key here is directly scale boostDC according TPS values and not by the target. Boost target is there for limit boost if refDC table is badly configured and also helps to tune that table when you see that target start * '''boostalt has 12x12 table and separate output''' config_ext.boostalt_channel so it can '''operate simultaneously with the PID boost'''. Boostalt can be configured in 4 different operation mode (or strategy) coded in boost_conf bit1:0 ** 00: '''boostalt relative lookup'''. Boost reference table lookup ( RPM, MAP - boost), where MAP - boost range is -155 .. +100 kPa (for 0..0xFF vemsv3.ini update needed so it does not show confusing 0..510 in this mode). Solenoid inactive when MAP < boost_minpressure. Boosttarget is calculated the same way (from same config variables) as for the PID boost. Relative mode is useful for wastegate: eg. bottom can be controlled with PID and the top with a table-lookup, or the other way around (although most setups connect the 2 solenoids in parallel electrically, and opposite pneumatically). Highly flexible - versatile. ** 01: '''boostalt absolute lookup'''. Boost reference table lookup ( RPM, MAP ), MAP range is 0 .. +510 kPa (for 0..0xFF). boost_minpressure is neglected. ** absolute mode is useful for water injection or other PWM output (can be used eg. for a gauge or an analog output). ** 10: boostalt "simpleboost" closed loop - for compatibility with existing installs that utilize "boost alternate" ** 11: boostalt "simpleboost" open loop. Boost reference table lookup ( RPM, analin) where analin is an mcp3208 analog input (like a 0..5V potentiometer) '''[http://www.vems.hu/download/v3/firmware/ 1.1.48]''' - testing only. AlphaN new strategy added (upgrade possible with 2009-01-08 VemsTune or newer) * ''to use AlphaN without MAP sensor'': ** hybrid_rpm_m=FE or lower, but not 0 (so the load is sent in position 4-5 byte of the realtime vars, not MAP) ** important: fill mapcorr table with all 00 (hint: at least set the 4 corners). * ''new mapcorr (boost map correction) strategy'' available for AlphaN with a MAP sensor (eg. charged + ITB): ** hybrid_rpm_a=FF ** hybrid_rpm_m=FF ** config13 bit5=0 to lookup ign and lambda according to kpa (even at low RPM). ** config13 bit2=1 for no automatic "alltime" load multiplication (yes, this is no mistake: because you do not want load multiplication below 100 kPa!) ** important: '''fill mapcorr table''' (i/I comm commands). All 100 means neutral. With this configuration *MAP/100 is also applied above 100 kPa, automatically, so table stays flat and tuning becomes very easy. Maybe a small slope is a good idea to start with: 100 @ 100 kPa (below 100 kPa the table is not used at all) and 110 @ 320 kPa. If you need more than +10% at high kPa, that suggests the fuel system is not up to the job. * the old "hyprid tps and kpa smooth gradient" strategy is still available: which uses TPS for lookup RPM<hybrid_rpm_a and MAP when RPM>hybrid_rpm_m and interpolation in between. However, at high gear, boost might appear at rather low RPM (especially in a streetcar). The mapcorr strategy is also easier to tune ** hybrid_rpm_a=14 (2000 RPM for example) ** hybrid_rpm_m=1E (3000 RPM for example) ** config13 bit5=0 to lookup ign and lambda according to kpa (even at low RPM). On the countrary bit5=1 means load (tps or kpa or mix depending on RPM, hybrid_rpm_a, hybrid_rpm_m relation) is used, the only option for a normally aspirated engine without MAP sensor. ** config13 bit2=1 for no automatic load multiplication '''[http://www.vems.hu/download/v3/firmware/ 1.1.47]''' - release candidate (still testing) * misconfiguration of inj_staging_conf (FF instead of 00) can cause symptom like: only first element of h[0] (injector output table) actuates (injects). * in general extreme care is needed for the relatively new n2o options and matcorr ** VemsTune is being updated to warn about certain misconfigurations. * partial fuelcut (special case when overrun_fuelcut - overrun_fuelresume = 10 or 11 that is ** RPM difference = 1000 RPM is configured : higher TPS => higher RPM-range for the partial fuelcut ** when RPM difference = 1100 RPM is configured : when TPS > iac_tps_threshold, partial_fuelcut=min( ft(TPS), fr(RPM) ). This means higher TPS can never increase fuelcut. '''[http://www.vems.hu/download/v3/firmware/ 1.1.46]''' - will not be released * partial fuelcut (special case when overrun_fuelcut - overrun_fuelresume = 10 that is 1000 RPM configured).) TPS pushes the 1024 RPM partial fuelcut window up by +4000 RPM / 24% TPS * 3 new options for ego, freeze corr for openloop, minkpa, reactivation delay. * boost mat and egt (added?) to boostpage * cranking pulsewidth tenpoint curve '''[http://www.vems.hu/download/v3/firmware/ 1.1.42]''' - will not be released * '''tooth_wheel_twidth endianness swapped''' (to allow user friedly interface with megatune's constraints). Check primary trigger helppage, '''AND CHECK YOUR SETTING, CAUSES UPGRADE PROBLEM!''' VemsTune will consider this automatically during upgrade: new settings are calculated by: ** tooth_wheel_twidth[0]=tooth_wheel_twidth2 (upper byte) ** tooth_wheel_twidth[1]=tooth_wheel_twidth1 (lower byte) * '''starts up in 19200 baud''' (not 9600, so megatune must be configured 19200,8n1). Note if 1.1.42(or newer) already in board, upload any (newer or older) firmware with command "tools\megaloader.exe firmware\vems.hex -wtE -b 19200 -c com1" ("-b 19200" added) and if config.lcd_c0 bit2=1 open and close MT before upgrading firmware to disable AIM stream. * config.lcd_c0 bit2=1 will start AIM stream automatically for round (with 2008-09 firmware) or aim display ** aim stream stops for Man command, or megatune A command or 'o' (config read) command, so it should not confuse MegaTune * Cosworth initial syncup sequence fixed. Hopefully no more backfire during cranking * for any test report, mention if 16x14 or 12x12. Although technically not justified, for marketing, we are migrating towards 16x14. Perhaps 12x12 support will be dropped at some point. Watch injector injgroups, does the h[0] first entry squirt with 12x12 firmware ? (we've seen in not squirt once, for 16x14 we haven't seen that) * problem: apparently with PID-boost the boost(RPM) table is not considered. Workaround: raising map-target offset. Under investigation: likely related to the gear-dependent boost code. '''[http://www.vems.hu/download/v3/firmware/ 1.1.41]''' will not be released. (Fero.uses it on racecar with success. not sure if 12x12 or 16x14). No config change. * ego correction works even in the very special case when partial fuelcut is configured: overrun_fuelcut - overrun_fuelresume = 10 (that is 1000 RPM). * shiftcut is disabled when launch input is active (so partial clutch will not accidentally activate shiftcut retard) * Make sure MAP filter configured properly. Andrey's starting values MAP samples = 15, Triggers num = 1 for 4cyl, MAP samples = 10, Triggers num = 2 for 6cyl. '''[http://www.vems.hu/download/v3/firmware/ 1.1.40]''' - not released yet, only for bench testing. Many new variables (vemsv3.ini must be used with megatune), upgrading (even from 1.1.39) is quite problematic * '''launch disables shiftcut''' (even if not configured to same analog input). After launch inactivation, shiftcut rearming is needed (not to interfere with launch, even if clutch partially engaged) ** '''shiftcut_time=FF''' (max, 1020 msec) means '''shiftcut holds as long as input is kept active''' (but min 150msec. Good for practicing, or if gearshift 1=>2 takes much more time than 2=>3 and 3=>4, like in case of short-first-gear long-second-gear audi gearboxes) * special case handled for 5cyl engine with 60-2 and camsync (does not wait indefinitely for nonexisting tooth 10+24+24 and 11+24+24) * '''Stall-O2 feature''' in runtime vars (position 16 and 17) sending O2 (percentage * 2560, so 20.95% is 53632 ) instead of PW when engine is stalled but LSU4 is on) - makes '''mde02 WB calibration procedure much-much simpler''' for those without LCD. Use StallO2 MT gauge, but not PW gauge. '''[http://www.vems.hu/download/v3/firmware/ 1.1.38]''' - not released yet, only for testing. * smooth come out of partial fuelcut idle if TPS pressed * iacref_matslope added * internal (scheduler) changes that might break things '''1.1.37''' is placeholder. * if we find problem with 1.1.38, than we patch the 1.1.38 features to 1.1.36 firmware so it can be released 2008-08-16 '''[http://vems.hu/download/v3/firmware/ 1.1.35]''' for bench testing only (Fero uses on engine with 12-1 trigger). [http://vems.hu/download/v3/firmware/vemsv3.ini vemsv3.ini] still needs some more work and testing, eg. the nice 2D warmup wizard was disabled in this, and the new boost settings are temporarily in the shiftcut dialog) * supports [http://shop.vems.hu/catalog/product_info.php?products_id=131 internal logging] to [http://www.vems.hu/download/SD/v3SDcard.htm SDcard] * shiftcut input thresholds for picky sequential gearboxes: ** shiftcut_potlow: activates below this lo-threshold (19.6mV unit, use 95 for a normal input switch that is appr 1.9V threshold ) ** shiftcut_pothi: activates above this hi-threshold (19.6mV, use 255 with normal input switch so it never activates) * shiftcut TPSthreshold: shiftcut will not activate below this * shiftcut igncut (0..100%) not necessarily 100% as earlier * shiftcut retard (0..60 deg ATDC) * boosttarget decrease above adjustable MAT * boosttarget decrease above adjustable EGT * '''partial fuelcut idle''', if overrun_fuelresume + 1000 == overrun_fuelcut (0..100% in the 1000 RPM range if TPS <= tps_thresh. Smoothly adding power if TPS pressed NOT applied before 1.1.37) ** with an ALS engine with always-open throttle, this is smoother on the environment and the exhaust than an igncut based idle '''1.1.34''' do NOT use * dropping VE learning. PC based VE analysis is much better now, no use for this smartness in firmware * power shiftcut (partial igncut and retard configurable, not just full igncut) * overrun fuelcut and fuelresume not injector 0 or 100%, but partial (even dropping) is possible * more testing on 16x14 tablesize which is likely to become recommended in future. (12x12 might be finally not supported in megatune) * more testing with InputTrigger/MultiToothNoneMissing (eg c010 5cyl 5 crankpulse + camsync) * RS232 port alternative use: either sending AIM-protocol data (to several rounds or aliens) or talking to MT. But not simultaneously. 19200baud (always, even in MT) considered '''[http://www.vems.hu/download/v3/firmware/ 1.1.33] for testing new features on the bench'''. Although a bug that made outputs (including ign) misbehave in 1.1.28-1.1.32 was fixed, prefer 1.1.27 unless testing the new features (like switch-table). * MAT enrichment is not firmware-upload time any * ... other changes ? '''1.1.32''' - do NOT use * '''very important to read all changes between versions you upgrade from-to. Notably a very important change in 1.1.30: missing-tooth wheel users likely need to adjust tooth_wheel_twidth2 (labelled "width of missing tooth" in MegaTune), likely to 00'''. This (width of tooth!!!) effects spark timing. This is very important. * solenoid PWM iac compensates with measured VBATT, so it's somewhat easier to tune with a Bosch alternator where VBATT depends on RPM at idle (unwanted positive feedback). ** Solenoid iac duty=0 (unpower iac) after pump_on_mintime (same behavior as fuelpump), and only apply cranking iac position after cranking starts * '''config.ign_dwellmap''' configurable MAP-based dwell scaleup (VBATT based is automatic) ** 00 no dwell extension at all (neither VBATT nor MAP based): ign_dwell14 is used ** 01 only VBATT based dwell extension (max *2, when VBATT <= 7V) ** values 02..FF : VBATT based extension is automatic. Slope of MAP-based dwell-extension is configured with this variable. MAP-based extension is max *1.4 (+40%), while total scaleup (vbattmultiplier * mapmultiplier) is max *2 *** dwell is extended by (map - 70kPa) * ign_dwellmap/4096 but no more than +0.4 (40% extension that is *1.4). Examples: *** ign_dwellmap=0A (decimal 10) max dwell-extension +40% reached at 234 kPa because (234 kPa - 70 kPa) * 10 /4096 = 0.4 *** ign_dwellmap=14 (decimal 20) max dwell-extension +40% reached at 152 kPa because (152 kPa - 70 kPa) * 20 /4096 = 0.4 *** higher value means higher slope (so max dwell extension reached earlier). Take care not to overload the ignition output or the coil (use lower ign_dwell14 so you only get extended dwell when really necessary). * '''config.iac2load''' see GenBoard/UnderDevelopment/AlphaNIAC ** When rpm < hybrid_rpm_m the iac.position is added to load before table-lookup (for all tables: VE, ign, lambda). Basically proportional to the iac/throttle ratio. Formula: iac.position * config.iac2load /256 / 8 is added to load, that is max 31 if iac.position=FF and iac2load=FF. iac2load=FF is throttle/8 size (area) iac, that would be extremely big. Probably iac2load=06 or similar values could be more realistic (iac 341 times smaller area than throttle). Because of the throttle shape (special at low-TPS) you might need some experiment and logging to tune it * the P and D terms in PID boost became asymmetric. The main difference between simple-boost and PID boost was this (simple-boost was asymetric and therefore easier to tune) * switch-activated '''full-table-and-config-swap''' see TableSwitch. Expert feature, must upload valid config+tables to both storage (2 completely different permanent storage area inside the ECU) separately. Same ignition and injection outputs must be used in same positions (otherwise => risk of blowing ignition fuse and flooding cylinders with gasoline. Therefore this does not allow additional injector-staging-hack features, besides the existing staging functionality). Configuration of storage0 (eg. confswitch=C3) can allow to change to storage1. If you want to be able to switch back to storage0 without reboot, config of storage1 (eg. confswitch=D3) must allow to change to storage0. MegaTune does not know about which storage it reads/writes (closing megatune windows before switching to alternate storage seems wise). Useful for bi-fuel, or special different tables like N2O or similar. Switching can also depend on misc1 condition (TPS, RPM, MAP). After switching, the misc1 condition of the new storage will be used (which can be slightly different, eg. to allow hysteresis). For most applications, rpm < 100 (engine off) is recommended, or at least rpm < 1500 (idle). For N2O, the condition will be different of course. ** confswitch=00 disables switching ** confswitch=C3 does load_variables(1) if mcp3208 ch3 goes low ** confswitch=D3 does load_variables(0) if mcp3208 ch3 goes high ** confswitch=83 does load_variables(1) if mcp3208 ch3 goes low AND misc1 condition is true ** confswitch=93 does load_variables(0) if mcp3208 ch3 goes high AND misc1 condition is false ** any mcp3208 analog input channel (mcp3208ch 0..7) can be chosen with confswitch bit 0:3 ** use it with known-good config manually uploaded to storage1 first (mcS1 command). When tweaking from megatune, verify the result after (both storage0 and storage1) * compiled with BOOSTGAUGE (linear analog voltage output on an internal s259 channel for audi dash boost input) '''1.1.30''' do NOT use any 1.1.28 - 1.1.32 ! * low-RPM improvements for 8 ( InputTrigger/LanciaCosworth ) and 10+1 ( InputTrigger/MultiToothNoneMissing ) triggers ** 4-1 trigger supported (a 4 cyl VW Golf running on it) since 1.1.29 * '''IMPORTANT: tooth_wheel_twidth2 must be 0 when normal tooth < 64 crankdeg (as usual). tooth_wheel_twidth2 is NOT the width of the missing tooth (any more''', as annotated in megatune), but always the upper byte of the normal toothwidth. Normal toothwidth is in '''0.25 crankdeg resolution''' (no change here). So for a 60-2 the toothwidth is 6 deg ( in 0.25 deg unit that is decimal 24=0x0018) ** tooth_wheel_twidth1=18 (24 in megatune) ** tooth_wheel_twidth2=00 * for a 10+1 where a tooth is 72 degrees (decimal 288=0x0120), ** tooth_wheel_twidth1=20 (32 in megatune) ** tooth_wheel_twidth2=01 * so how will the code know the width of missing tooth for a missing tooth wheel ? If the "special N-1" is enabled (primary_trigger bit4=1) than 2xnormal tooth. Otherwise (N-2 assumed) 3x normal tooth. '''1.1.27 Released''' [http://www.vems.hu/download/v3/firmware/v3_firmware_1.1.27.zip v3_firmware_1.1.27.zip] * seems to run well on many cars (60-2, auditrigger and others) * almost same as 1.1.26, only change is InputTrigger/MultiToothNoneMissing N+1 trigger supported where N<=11 '''1.1.26''' [http://www.vems.hu/download/v3/firmware/v3_firmware_1.1.26.zip v3_firmware_1.1.26.zip] '''for testing ONLY''' - several cars running well on it (including 60-2 triggers) * should work with missing tooth as well (and auditrigger of course). Eliminated a rare-case condition which resulted in ignitioncut (when ecu had to be rebooted) ** more testing is under progress, also with N+1 type trigger (like 24+1 non-missingtooth "coiltype"). Note that overlapping dwell cannot be disabled: distributer is still not supported (except with manually limited dwell considering rev_limit). * launch does not retard spark if RPM drops considerably (control-range) below launch_rpm (allows high retard with otherwise weak turbo engines) * TPS-dot (and MAP-dot) acceleration enrichment kicks in sooner, and fades out gradually. Minor retune of the accelerichment enrichment added fuel amount might be necessary * knock_chan=F0 disables knock. Other values select bank individually for each ignition event (practical for V8, V6, I6 and even I5 where there are more than 1 knock sensors). ** knock_chan=00 selects knock channel 0 for all ignition entries ** knock_chan=FF selects knock channel 1 for all ignition entries ** knock_chan=AA selects knock channel 1 for every other ignition entries. Bit7 is for en Optional: Add document to category: Wiki formatting: * is Bullet list ** Bullet list subentry ... '''Bold''', ---- is horizontal ruler, <code> preformatted text... </code> See wiki editing HELP for tables and other formatting tips and tricks.