_____ ( _ ) | ( ) | | | | | | (_) | (_____)
___ ( _`\ | |_) ) | ,__/' | | (_)
______ ( __ \ | ( \ ) | | ) | | | | | | | ) | | (__/ ) (______/
## ## #### ## #### ## ##
_______ ( ____ \ | ( \/ | (_____ (_____ ) ) | /\____) | \_______)
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''' '''VERY IMPORTANT''': for any firmware, the '''VemsTune MUST be newer than the firmware''' (see File / Firmware info and VemsTuneVersion ). * Only use older VemsTune (at your own risk) if absolutely necessary and you really know what you are doing, and do NOT ask for support. * "Help/Manual update" only updates ini files, which is not enough in all cases. * Old vemstune (sometimes even with updated ini files) might not warn about important misadjustments, or might not show some settings at all! (No surprise it does not show values added in the future) ** Sticking with and old VemsTune (that worked perfectly for some similarly old firmware) after upgrading to a new firmware is a very bad idea. Actually it is '''best to install VT first, and do the upgrade with the new VT !''' * '''new VemsTune applies some useful config changes''' when old saved config (eg. 1.1.85) is uploaded to newer firmware (eg. 1.1.96). This saves a lot of time and headache ---- This page is 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''' (especially important if uploaded firmware is 1.1.85 or older) * '''remove injector connectors and remove fuelpump relay''' ** or otherwise make sure the fuelpump is off, eg. hidden anti-theft switch ** (especially important if uploaded firmware is 1.1.85 or older) ** 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. '''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 * 1.1.x always supported 60-2 trigger and supports dual config since 1.1.3x (read below). * very few will use 1.0.x now ** 1.0.x is unlikely to ever support TableSwitch '''IMPORTANT: use new VT, or at the very minimum, update VT ini files from web''' 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''' (in some cases it might work, never assume that). * 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. '''When upgrading to non-released firmware''' (eg. 1.2.15), the following steps should be applied: * start with a tuned engine running well on '''earlier firmware''', either latest released (eg. 1.2.11) or newer (eg. 1.2.14) * upload the '''new firmware''' (eg. 1.2.15) * optional: upload the new version ( from experimental/ or X/ directory if necessary) * enable the '''new functionality''' in config Do some testing (at least a spin around the block) with logging in each step. Report should contain the sequence (eg. GOOD, GOOD, GOOD, WEIRD IAC oscillations; with '''vemslogs''' for all steps, at very minimum for problem cases). ---- '''[http://svn.x-dsl.hu/f/ 1.2.43]''' - experimental * Subaru Subaru EJ255 related trigger change ** note: bigger (more flexible) allow/deny window ; eg. sectrig_ignore = 0xF6 (0xF0 enables sectrig gating based on tooth_cnt, not ptrig_count) ---- '''[http://svn.x-dsl.hu/f/ 1.2.42]''' - experimental, do not use yet (unless for testing, after project page + review). * receive multiple analog and wheelspeed data via 2nd serial (when enabled in config), while also sending AIM via 2nd serial (possible for multiple receivers) ** most useful when controlling 4 camshafts (quad vanos) and 2 CAN-bus (eg. BMW S65/S85) ---- '''[http://svn.x-dsl.hu/f/ 1.2.41]''' - experimental, do not use yet (unless for testing, after project page + review). * Honda K20A InputTrigger/HondaKTwentyTrigger ** more robust syncup with very high compression Honda K20A HALL-HALL trigger * autotrans, new modes * wired-in some WBO2 pid values (constants were used anyway, no benefit to change) ---- '''[http://svn.x-dsl.hu/f/ 1.2.40]''' - experimental, do not use yet (unless when explicitely told so, for testing). * no change for PWM VVTI, but for 2 solenoids per camshaft ("one advance, one retard solenoid") : ** when inverted output configured, allows '''SWAP-ping outputs''' for DUAL activation [http://vems.hu/vt/help/v3/v3_camshaft_angle_control.html VANOS] ** note: good for s259 controlled NFET or p259 controlled PFET ** not suitable for p259 controlled NFET, which needs inverting indeed ---- '''[http://svn.x-dsl.hu/f/ 1.2.39]''' - experimental, do not use yet (unless when explicitely told so, for testing). * designated improvement: significant RPM fluctuations during cranking + long dwell combination ** more igntesting ahead, to cover all cases, with these changes applied ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.38] - released ''' use with [http://vems.hu/vt/ VT] 2017-03-22 or newer + update ini from web '''compile time 2017-04-11:''' * '''SD logging''' ** bugfix: fixed an issue where VT could not read logs from SD in triggerframe mode. '''compile time 2017-03-14:''' * '''Idle control''' ** bugfix: fixed an issue where IAC 485Hz PWM frequency together with dual output configuration could lead to task overload. '''compile time 2017-02-14:''' * '''Aim ERRORS (125) channel''' ** baro problem replaced by SD logging active '''compile time 2017-02-06 16:25:''' * '''Shiftcut''' ** hotfix: when deactivation time 1020 is set-> enable on input active; removed 120msec delayed deactivation '''compile time Febr 5 or older:''' * '''NEW function -> [http://www.vems.hu/vt/help/v3/v3_tc_control.html Traction Control]:''' ** dual wheel speed slip based absolute/relative with configurable enable speed ** 4 types of independently configurable torque reduction strategies (enrich, retard or spark/fuel cut (skip cycles)) * '''Launch control''' ** allow launch switch always ON (without affecting idle control), launch will only activate when near the start rpm range ** added launch boost anytrim optionally also applied to launch boost target ** added launch control range for enrich/retard increase configurable ** added launch (re)enable/disable speed configurable even when curve based mode is disabled ** downsized launch curve from 8 to 4 bins, Traction control can optionally take over when past the glide from start ** bugfix: launch boost target is only applied when below (re)enable speed (even when curve mode used) * '''Boost Control:''' ** added RPM bins independent from Lambda table for Boost Target vs RPM mode * '''N2O Control:''' ** added option to disable n2o during shiftcut * '''Anytrim input channels:''' ** added Lambda-Target: Measured Lambda substracted by Target Lambda, effectively difference from target; usefull for safety functions like lower revlimit if engine running lean ** added Lambda-Target TPSx: Difference from Lambda target scaled by TPS saturates at x TPS%, scaling 0% when TPS below idle_tps_threshold ** added Lambda-Target MAPx: Difference from Lambda target scaled by MAP saturates at x MAP kPa, scaling 0% when MAP below fuelcut_min_kpa ** added NERNST 1 (ADC1), KNOCK (ADC6) & NERNST 2 (ADC7) can be used as anytrim input * '''Input switch channels (for switch based enable functions):''' ** added THRD TRIG (SDA), WSPEED 2 (SCL), WSPEED 1 (INT6) & SEC TRIG (IC3) can be used for input switch functions if not used for trigger/wheelspeed ** added NERNST 1 (ADC1), KNOCK (ADC6) & NERNST 2 (ADC7) can be used for input switch functions * '''NEW function -> [http://www.vems.hu/vt/help/v3/v3_inputmux.html Analog input multiplexer]:''' ** added analog input calibration for muxed channels and allow selection for use everywhere * '''NEW function -> [http://www.vems.hu/vt/help/v3/v3_analog_input_filtering.html Analog input filtering]:''' ** configurable exponentially weighted moving average filtering can be applied to all mcp and demuxed_mcp configured independently per channel * '''Idle control''' ** added 485Hz IAC PWM frequency ** bugfix: Simple On/Off mode hardware settings added in VT * '''Idle AC-Control:''' ** added support for selectable pressure sensor input with configurable lockout voltages for AC over pressure and AC low pressure with fixed hysteresis of 0.196V ** added support for selectable fan speed selection when when AC On * '''AIM stream output:''' ** added generation and output of channels OILPRESSURE (9), OILTEMP (13), FUELTEMP (109), FUELPRESSURE (21), ERRORS (125) ** added selectable predefined calibrations and channel inputs for AIM pressure channels ** added selectable predefined calibrations and channel inputs for AIM temperature channel ** ERRORS (highbyte) has function activation states: ALS, Launch, Shiftcut, n2o, Traction Control, Injector stage 2, Idle, Ego ** ERRORS (lowbyte) has error states: o2 not ready, trigger error state, config slot active, baro problem, engine knocking, warning light state * '''Misc Outputs:''' ** added predefined hysteresis on all windowed configurable input variables; 200 RPM, TPS 1.95%, MAP 4 kPa ** added predefined hysteresis on all extra configurable input variables; MAT/CLT 4 degC, EGT 32 degC, SPEED 4 km/h, Launchtime 0.0s * '''WOT/RPM outputs:''' ** added predefined hysteresis on all input variables; 200 RPM, TPS 1.95% * '''NEW function -> [http://www.vems.hu/vt/help/v3/v3_fan_water_pump.html Radiator outlet temp sensor]''' * '''Fan Control:''' ** added selectable input source (also changes control mode) coolant, radiator outlet or delta coolant radiator outlet ** added support for Dual Relay based Fan control with 1 second lockout time between output adjustments ** added support for Temic and VAG J293 PWM controlled external amplifiers ** added configurable fan disable above wheel speed * '''Water Pump:''' ** added selectable input source coolant, radiator outlet * '''Fuel Pump:''' ** added PWM fuel pump control with configurable PWM output scaling by fuel requirement * '''Triggers/Wheel speed:''' ** SUBARU_TRIGGER4_ANGLE allow cam trigger edge selection rising/falling on wheel speed 2 ---- '''[http://svn.x-dsl.hu/f/ 1.2.36]''' - only for testing fw and new VT ini with expert supervision (do NOT use this unless you have VEMS ECU with analog input multiplexer AND need to use some of the inpmux channels for control AND really know what you are doing) * more use of the extra analog input channels with index 16..126 (127 is always enabled; 255 is always disabled; since +128 means inverted semantics). ** practical total nr of analog inputs: 22 with the analog input multiplexer channels (6 new above 8 mcp3208 channels and 8 internal basic channels) or total 40+ with the external input-extender (2nd rs232 input traditionally used for GPS; request "full 2nd serial RX/TX" port in order comment to also have output for CAN-dash) ** not just for logging any more, but for activations (ALS, launch, shiftcut, ... even config-switch) ** [http://svn.x-dsl.hu/f/VT_1236_config.zip VT ini.zip] must be unzipped under VT installdir (where most recent 2016-mm-dd VT was installed; overwriting older ini-s) to support 1.2.36 firmware (do NOT click "update ini files from web"). ** Do NOT use 1.2.36 for now if it's problematic to save the v3_firmware...zip under webcache/fw/ + using Tools/fw webtool/(Re)install , or unzipping VT_1236_config.zip to the right place ** Installer can manually verify with $7E$B1$7F$34$7E Manmcdmct and adjust if necessary. Create project page, test thorougly and wait for confirmation. Linked to the project page: Manmcdmct result and List of files under VT installdir (preferrably with md5sum or sha1sum or sha256sum values) are needed in case of any uncertainty ---- '''[http://svn.x-dsl.hu/f/ 1.2.35]''' - obsoleted by 1.2.36 '''has known problem of mcp3208 sampling freeze, revoked''' * '''BroadcastDatastreamAim via 2nd UART''' (assuming 1-wire /I-button not used in same setup) ** useful for several AIM compatible receivers (TX => RX of multiple VEMS EAW52 rounds, dash, dataloggers simultaneously) on 2nd UART ** while android is connected to main RS232 port (in triggerframe mode, which has some advantages; but TF is request/response protocol so need software relay if data to be sent to multiple receivers, as multiple RS232 "TX" wires cannot simply be connected - except with RS232/RS485 adapter but that half-duplex mode is special: not enabled in standard firmware and VT and android software...) * note: using newer compiler 4.8.1 (which we must always handle with care: seems good on bench and engine, but more bench and engine tests are in progress; Currently only experts use these new features in very special apps) ---- '''[http://svn.x-dsl.hu/f/ 1.2.34]''' - experimental, only for testing and capturing vemslogs with expert supervision * Note: if '''validate''' shows '''warnings'''/errors after eg. '''1.2.31=>1.2.38 upgrade''', might need to '''temporarily change to 4 controlled camshaft to disable some values''' to suppress warning (change back to 2 camshaft mode if appropriate) * reverted to the Apr 18th compile which was great success with [https://vems.hu/vt/help/v3/v3_electronic_throttle_control.html#BMW_S65_S85_ETC BMW S65 CAN bus] * '''BMW S65 v3 controlled quad Vanos and v3/CAN-bus controlled IAC and ETC (DBW electronic throttle)''' ** uses SPI-bus CAN addon module ** (unlike the dummy throttle) bridgeoutput driver not needed (controlled via CAN-twisted-pair, terminated bus). Note: unlike S85, S65 has 240ohm in all actuators (3x), ecu requires 240 ohm termination to get to the 60ohm total ** BMW s65 is already in racing competition in the UK, owner reported car running much better than on the stock management. Final 430hp (e85) with factory NA engine (also on same dyno measured better than stock, which is rare for NA engines: BMW tunes them quite well) : at least it confirms good behaviour (on car) of the 1.2.34 firmware with 4 cam control + CAN. Needless to say, this setup is for experienced installers. * improved sync for WBO2 sampling => EGO control when 2 WBO2 channels are configured and used for control. ---- '''[http://svn.x-dsl.hu/f/ 1.2.33]''' - internal version * analog input multiplexer (with VT that supports it - scheduled release 2016-03) ---- '''[http://svn.x-dsl.hu/f/ 1.2.32]''' - experimental, only for testing * adjustable [http://vems.hu/vt/help/v3/v3_electronic_throttle_control.html ETC] VBATT compensation code * anytrim extensions; more input sources, modes and channels (5 instead of 3) available, more info [http://vems.hu/vt/help/v3/v3_anytrim_control.html anytrim help] * (experimental) Ign+Lambda tables, possible to tune using Y axis is calculated load, see [http://vems.hu/vt/help/v3/v3_alphan_settings.html alpha-n help]. ** smooths out the transitions when using ITB's VE(Alpha-N) and Turbo, normal users use KPA (or TPS) * Honda F20C later type / K20A, sometimes sees 1 sometimes 2 crank pulses between two following cam pulses (supported again) * table Kpa unit = 4 (>510 values in kPa based config and table axis values) fully supported '''be ware this also needs [http://vems.hu/download/v3gui/dev/VemsTune-Install-2015-03-16.exe VT Dev >= 2015-03-16] version''' * boostcontrol ** speed based closed loop control available (previously only refdc based) ** boost sensor calibration (when using external boost sensor) no longer shared with baro calibration * * AC idle_up control (refdc/steps add) with (optional) FAN link and (configurable time) delayed compressor relay activation [http://vems.hu/vt/help/v3/v3_idle_ac.html more info] ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.31]''' - only for testing (no known issue, should be at least as suitable as any earlier version) * '''Electronic Throttle "integral scaling curve"''' ** See [http://vems.hu/vt/help/v3/v3_electronic_throttle_control.html integral scaling curve] ** ETC PID I (integral term)'s max contribution to actuator PWM duty% can be nonlinear with abs_error (abs_error=abs(TPS-target)). Allows even better follow of very sudden target-change, while still reaching target precisely. Dummy configuration: constant 100% (and 60..100% for integral max positive/negative effect, the other 2 new variables) would be compatible with 1.2.30 and earlier semantics. * supporting a very special (uncommon) setup: speed density for fuel and TPS for other tables ** normally if MAP is used for fuel, then MAP is also used for every other tables. Normally TPS is used just for fuel (with MAP compensation or without) or TPS for everything. All these have been supported for a long time and worked well. ** TPS for everything except fuel is very uncommon and required minor changes: now used on a Porsche 911 (for some reason, experimentally) with this 1.2.31. ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.30]''' released (no known problems) * MembersPage/JorgenKarlsson/MiataNb implemented (Miata trigger, with camsync) ** Triggerlog was not provided. We made test wav and tested with that (tested OK). ** in VemsTune update ini-s from web; configlets/trigger/Mazda_Miata_NB_4_3_Camsync.txt . ** Note for the curious: configured similar to subaru 6+7 (with camsync of course), but no advanced missingtooth filter (config.primary_trigger bit3=0 ) : * '''LSU 4.9 with standard v3 hardware officially supported''' See [http://vems.hu/download/sensors/LSU4.9/ printable pdf] ** Power up ECU and measure nernst voltage (with nothing connected. Config does not matter at this point). If less than 4V (above GND), then '''install a 27k resistor between nernst and +5V''' for correct operation (pullup resistor for nernst current according to Bosch requirement: nernst reference current replaces the LSU4.2 reference oxygen gas connection-tube, which was sensitive to clogging. LSU4.9 sensor has faster response, and longer life expected under similar conditions). ** If the 27k nernstpullup resistor is installed (inside the box or outside), nernst measures >4V (typically appr 4.9V) when open circuit. With a 1k pulldown test resistor towards GND) measures typically appr 175-180 mV => don't install another 27k externally if already installed internally. ** '''LSU4.9 must be selected in config''' (independently possible for 1st and 2nd WBO2 channel). Other than that, same PID values work (eg Ri target is 165). ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.29]''' testing * Subaru 6+7 (also Suzuki and Fiat Stilo) allow operation without camsync * minor PID change: sum(error * I_term) instead of sum(error) * I_term ** normally same result, except when I_term is changed suddenly + significantly during tuning (makes tuning a tiny bit easier) ** if upgrading an engine with ETC (eg. from 1.2.21 fw), upgrade to 1.2.29 (2014-11-24) or newer. ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.28]''' testing only * special support was needed for some 60-2 BMW V8 with weird position of the sectrig missing tooth (BMW S62 only ? Thanks to Rija for triggerlog) ** compiled with -D MIN_MAX_PRIM_PER_SEC_CNT ** TODO: point to an explanation of a related special max_prim_per_sec_count=.. secignore=.. example with triggerlog image * enter TF mode immediately on startup, also after password unlock (fw after 2014-07-02 14:19). (= v3 config corruption practically impossible) ** for several years before 2014 VemsTune used the safe TF mode for communications, but the old unsafe "mt" commands were still supported, which made fw corruption possible in case of communication noise. With startup mode=TF config corruption is practically impossible (due to eg. connecting modem or GPS to the wrong connector, or using a cheap BT-RS232 adapter or other source of communication noise ) ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.27]''' testing only * faster syncup for most trigger types * including honda extrapulse trigger (with or without camsync) ** also always correct syncup for Honda extrapulse trigger under all known conditions ** that's great but where is the file? neither http://www.vems.hu/download/v3/firmware/ or http://svn.x-dsl.hu/f/ has it *** Under the experimental directory. Direct link: http://www.vems.hu/download/v3/firmware/experimental/v3_firmware_1.2.27.zip ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.26]''' testing only * faster syncup for coiltype + camsync ** also for InputTrigger/LanciaCosworth (which was actually very hard to start with recent fw). ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.25]''' testing only * For testing (or if living on the edge) use 1.2.25 (not 1.2.24). * no config change (but check earlier important things, eg. 1.2.23 CLT_IGN_ADJUST) ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.24]''' internal testing only (currently under investigation, not recommended at this time) * '''added 5 cyl non-standard 30-2 and 30-1 crankwheel support''' ** unlike 4/6/8 even cyl engines, '''5 cyl''' needs special "avoid tooth" for missing-tooth setup, so please, if possible: with 5 cyl, stick to 15-1 (homemade wheels) or the factory 60-2 ** instead of inventing new missing-tooth trigger setups (the 30-2 and 30-1 was now tested, and also confirmed working well in the field). ** It's not relevant to coil type triggers, but why not mention the 5 cyl options while we're at it: c270 (factory 135 tooth on crank); or the reasonable c010 trigger is also option (5 pulses on crank); c020 (10 pulses on crank) would work (with campulse not exactly at crankpulse) but not reasonable. * internal: TPS overflow avoid (which was not impossible earlier with certain tps_low/tps_high values if TPS > tps_high happened. Was not possible with most tps_low/tps_high values, or if calibrated to max TPS reading less than 99.8%) * 2014-04-02: '''improved SD card support for sdv2 cards''' (very common nowadays), and SDHC also ** still only the cards shipped with v3 SD-card socket are supported, any other cards are "YMMV" (your milage may vary), but with the new code the set of cards that are likely to work well grows greatly (and hopefully not many types of cards produce insidious results while apparently seem to work). ** almost all SD-cards we tested worked; notable exception: a Chinese clone of "Elite Pro" 2GB card that, when used from SPI writes partial blocks (yes, unbelieavable), and shows other anomalies (files don't show up in VT, so not confusing). * For '''1.2.23 and older fw, disable SD card logging if SD card is not inserted''' (and verified operational), or if there is any chance SD card will be removed. (disable SD-card in config when removing card, or insert new working card) ** verify operation with SD-card overview ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.23]''' recommended for testing * '''CLT_IGN_ADJUST (coolant/load based ignadv adjust table)''' ** useful to decrease spark advance for high CLT, high load (high MAP), eg. some engines (many Subarus) need lower ignadv when very warm to prevent knock ** also possible to increase spark advance slightly for warmup ** take a look at the coolant/load based ignadv adjust table, make sure it's not random (should be set to some reasonable value during upgrade, but do verify it). Set it to all-0 if no better idea. '''Don't install 1.2.23 or newer firmware without making sure reasonable values are set in this table''' * as always, VemsTune '''must''' be newer than firmware, use the most recent, and '''don't forget to run iniupdate''' ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.22]''' - for testing * same as 1.2.23, but compiled without CLT_IGN_ADJUST (no coolant/load based ignadv adjust table) ** preferrably use 1.2.23, and 1.2.22 only for comparison (1.2.22 has no advantage except that ini-files are similar to 1.2.20-1.2.21) ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.20]''' - obsoleted by 1.2.22 and 1.2.23 * acceleration enrichment within 25 msec (was 100 msec) ** with 1.2.20, one might need to set 1/4 (quarter) of previous accel enrichment amount ** retune accel enrichment if possible: some adjustment of the bins could help, especially for lower (dV / dt) values * several other speed improvements (eg 60-2 V8 engine >10000 RPM, with double high-frequency wheelspeed signal, injangle, knock, camshaft-angle-control, ETC, VT-comm all active at the same time). * narrowband lambda sensor: selectable input channels (2 inputs, similarly selectably individually for each cyl, like with dual WBO2) ** (NBO2 might have not worked in some previous fw versions) normally not used for real installs anyway (only used for factory harness plug-n-play before swapping to LSU4 and bigger injectors) * requires uhex bootloader (all boards and ECU purchased since 2011-03 has uhex bootloader) Verify in VemsTune File / "Firmware info": Marked by the 'u' as in "v3.3_u009583" ** older devices upgradable after locking certain internal "fuse" bit with an [https://shop.vems.hu/catalog/cable-p-165.html AVR-ISP] device (or a cheap 5-wire DSUB25 parallel printer port stk200 or similar ISP cable with avrdude). * EVO trigger (with camsync; would be c004 without camsync): supported with wider range for threshold angle (slightly lower than 90 deg allowed: if starts hard because of extreme angular acceleration at startup, try lower than 90) * HEMI / Elise (Rover-K) triggers supported: without camsync, or with camsync (new). * Injector staging: scales actual pulsewitdth, not pulsewidth + deadtime (verify tuning after upgrade, especially if staging was enabled) * known issue: with some configuration (only when stepper iac configured) ADC reading might stop. Fixed in 1.2.22 Any install configured with stepper iac should not use 1.2.20, upgrade (see 1.2.23) ---- '''[http://svn.x-dsl.hu/f/ 1.2.19]''' - for testing only * ETC related (more choices - to get right): ** selectable TPS/PPS1 threshold for IAC activation ** ETC selectable PWM frequency. * planned (under construction): faster and more precise TPS/accel enrichment ---- '''[http://svn.x-dsl.hu/f/ 1.2.18]''' - for testing only * support "secignore" = 1..30 again ** 1.2.16 and 1.2.17 was restricted to secignore = 255 ("ignore sectrig pulses above...") after cranking / initial syncup was made faster (for coiltype or missingtooth + sectrig triggers) ** secignore is important for some BMW with multitooth type (eg. 6+1 or 8-1) sectrig (most often exhaust cam), like the S54 and similar, where practically secignore = 10..16 (decimal) work ** reminder: if the interesting sectrig pulse in triggerlog shows up at tooth10 (every 2nd crankrot), than use +3 (secignore=0D, that is 13 decimal) * '''triggerlog optical improvements''': distinguish different ingout entries ** and visualize primtrig relative gaps. Eg 2 (or 3, for N-2) times taller pulse shown at missing tooth), and TDC also seen better (due to compression slowing the piston) ** makes it easier to set ignition outputs (strobe is still needed for proper "TDC after the trigger" setting but at least easier to get close this way) ---- '''[http://svn.x-dsl.hu/f/ 1.2.17]''' - for testing only * higher (1 crankdeg) resolution for knock window start position ** better knock-sampling support for high individual cylinder retard (>10 deg) like used with odd-fire (although odd-fire knock is still untested on real engine). * with ETC configured, inj-H (and/or INJ-G) acts as PWM-command (for external power-module) if INJ-H (and/or INJ-G) is not enabled in ANY injgroup ** this-way ETC power-module can be connected to older controller without disassembly and internal mods (note: for any InTake/DriveByWireThrottle application, it is still obligatory to setup safety relay and 2nd sensors and configure the safety functions with proper acceptance thresholds, and requires high quality installation in every respect) * the "Pump on After powerup" function supports non-p259 fuelpump (also the time restarts from correct button touchon; otherwise measured from powerup of course). ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.16]''' * "unpower IAC when not moving" configurable (iac_conf bit0, same bit as for stepper) when IAC PWM dual-solenoid is configured * of course new VT is needed (even from [http://vems.hu/download/v3gui/dev/ Dev] * virtually same functionality as 1.2.15... the reason for this version: ** experimental 1.2.15 (from 2013-08-20 or so) had stumbles (due to misfire as it turned out, that took some time to reproduce). Now fixed, and named 1.2.16 to prevent any uncertainty. ---- '''[http://svn.x-dsl.hu/f/ 1.2.15]''' for benchtesting (obsolete) * compiled with KNOCK_ALTERNATIVE (with tuner-defined knock thresholds in a 8x4 table, 8 RPM and 4 kPa bins; not 8x8 as earlier, and not the more complicated 2-knock-sample per event that was default) ** made lotsof tests (with new VT and Preferences / update of course) with good results. Notable exception (yet untested): odd-fire maserati (or any other weird setup with >= 10 degrees cylinder-specific spark delay for any cylinder) should use earlier firmware (will be supported again of course). * somewhat more tolerance if wheelspeed signal amplitude diminishes at high frequency * made als_retard configurable absolute/relative * made coiltype (eg. c004 "4+1" or c024 "24+1") trigger with sectrig sync reasonably fast (was annoyingly slow since 1.2.7, if attempted after the initial fuelpump "priming") * '''Release of 1.2.15 is delayed''' because finding misfire ('''missing a few consecutive ignition pulses''') symptom early September, reproducible with certain config at least with 2013-08-28 and 2013-08-29 builds ** not reproducible (not seen at all) on earlier builds like 2013-08-12 or so ---- '''[http://svn.x-dsl.hu/f/ 1.2.14]''' for benchtesting * '''Electronic DC-motor based throttle''' with 2 pedal-sensors and 2 throttle-sensors ** 2 dimensional RPM, PPS1 => throttle-target curve ** curve for verification of PPS2 (2nd pedal position sensor) ** curve for verification of TPS2 (2nd throttle position sensor) ** different PID values for above target and under target: the throttle response could be tuned to yield perfect response with virtually no overshoot or undershoot (several throttle-bodies tested) ** ~500 Hz PWM, via precise HW-pwming (OC0). Injector-PWM-ing cannot be enabled simultaneously with ETC. (either high-Z inj or low-Z with series power-resistor is possible of course). ** safety relay shutdown (and configurable overrun fuelcut) if throttle way above target, or 2nd sensor verification (PPS2 or TPS2) failed - with configurable thresholds. * iac2etctarget: how much IAC.position can effect electronic throttle target 0 .. 24.9% ** effectively work can be split between IAC and Electronic Throttle in ANY ratio * TPS overrun fuelcut delay configurable (0 .. 3.1 sec) ** time measured after TPS released (even if RPM > overrun_fuelcut, fuelcut won't kick in for the configured delay). ---- '''[http://svn.x-dsl.hu/f/ 1.2.13]''' for benchtesting * select EGO1 and EGO2 for each injector (EGO correction based on WBO2 channel1 or channel2) * For odd-fire maserati, injection is 30 crankdeg delayed every 2nd injevent ---- '''[http://svn.x-dsl.hu/f/ 1.2.12]''' for testing * for '''missing tooth trigger''' (60-2, 36-1, etc...) with camsync, '''secignore=246''' (which is 256-10) means that sectrig pulse coming at tooth 0-10 is ignored ** this is '''useful if sectrig normally comes right before tooth0''' (which is the pulse after the missing tooth), as in some BMW (sectrig pulse at the middle of the missing gap 6..12 crankdeg before tooth0 pulse; with HALL the chosen edge can help but not with VR sensor), but sometimes sectrig is delayed to slightly after tooth0 (either because of VVTI action, or cambelt sloppyness, or agressive clutch action). With this config it tolerates this and does not resync (no stumble) ** another example: secignore=240 (which is 256-16) means that sectrig pulse coming at tooth 0-16 is ignored. Example values: ** secignore=246..250 for wheel with 22 or more teeth, ** secignore=250..253 for wheel with 10..11 teeth ** secignore=253 for wheel with 7 teeth ** secignore=254 might work for wheel with 5 teeth (or even less) but it's not recommended: sectrig pulse should NOT come near the primtrig missing tooth for any low toothcount wheel (or any homemade wheel, since those should be HALL anyway, and it's easy to satisfy this condition by choosing the right primtrig edge and sectrig edge) ** secignore=255 means sectrig pulse is not ignored, no matter at what position it comes. * '''Added support for SUBARU H6 pattern'''; 36-2-2-2 but a different pattern for '''engine H6''': 19, missing, 10, missing,missing - see See InputTrigger/SubaruEngineHsix ** Normally used with camsync (and camshaft-angle-control ~VVTI/AVCS). ** "Ignore sectrig pulses above" = 94 (secignore=5E, or 5F) for H6 engine ** For Subaru H4 engine, set "Ignore sectrig pulses above" = 62 (secignore=3E, previously didn't matter), see InputTrigger/SubaruThirtySixMinusTwoMinusTwoMinusTwo * Convenient calibration of WBO2 2nd sensor (without plugging sensor2 into socket of sensor1) ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.11] - Released''' * LCD page9 can display 4 EGT and 2 lambda ** it seems popular now (especially in USA) to log multiple channels that v3 can do easily (some v3 ordered with 8 EGT-inputs and 2 lambda) * During launch, launch_maxkpa takes over boost_target (if lower) ** nice on tires * 2nd ibutton switches to eeprom page-B ** example usage: page-A can be "wallet-mode" and page-B can be high power (or "fallback to safe-defaults" and "cutting-edge under-tuning", whatever one likes) 2nd I-Button is easy to configure in new VT (remember the base rule: VT newer than firmware). Options in short (internally coded in a backward compatible way in config prohibit_config bit 7:6): * 00 : Ibutton disabled * 10 : Ibutton enabled, antitheft, no keepalive * 11 : Ibutton enabled, antitheft, with keepalive * 01 : Ibutton enabled, but only for switching to page-B with ibutton_code2 (prohibit h/l : firmware version dependent): I-button touchon not required to start the engine. ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.10 ] - offered to wide audience for testing, extremely promising''' * Tranny drumsensor support: when enabled, gear detection is calculated from shiftcut analog input (and 6 specified voltages, 1 for each gear) ** in this mode, shiftcut activates when outside the ''voltage[gear] - shiftcut_potlow .. voltage[gear] + shiftcut_pothi'' window ** configured shiftcut_potlow and shiftcut_pothi should be small values (around 0.15 .. 0.2V) not 2V and 5V as for traditional shiftcut activation ** of course, shiftcut/launch sharedinput makes no sense together with drumsensor=enabled * bugfix: fixed the undefined behavior, after we found and reproduced by an otherwise invalid setup (floating analog input): when LCD is enabled in config (regardless of wether LCD is connected or not) AND lcd_pagestep enabled (or inverted0..inverted7 input selected) * bugfix: for wasted-spark ignition when ignout0 was among the channels, eg 0,2,0,2 (or 2,3,0,2,3,0) limited the dwell below the expected 360 crankdeg (minus a few hundred microsecond). ** 0,2 (or 2,3,0) worked but the multiignout config is normally preferred (for better indication in individual-power and knock). Either ignout setup works fine now. ---- '''[http://www.vems.hu/download/v3/firmware/ 1.2.9]''' available for testing * PID boost with "target based reference PWM duty" option is easier to tune for very quick reaction and small overshoot (if the hardware actuator is good) ** reference DC can be configured in function of RPM, boosttarget ** PID can be configured to switch to PD when boost is way too low, or to PI when boost is too high ** actually, configurable thresholds for operation ranges: ** boost off ** max-DC (to help spoolup) ** ref-DC ** ref-DC + PD ** ref-DC + PID ** ref-DC + PI * communication improvement: send overview of all pages command was improved internally. Better behavior when connecting during driving (eg. android) ---- '''[http://cell.dyndns.ws/f/ 1.2.8]''' for testing * wheelspeed dependent PIDboost uses wheelspeed2 when config.als_max_cut bit7=1 (wheelspeed1 when bit7=0) * experimental tachout change ---- '''[http://cell.dyndns.ws/f/ 1.2.6]''' for testing. - ALS/launch combined * wbo2 "quick-on-line" improvement (might be more generic than 1.2.4 - 1.2.5, eg. graceful with more sensors/cases, one more reason that 1.2.4 - 1.2.5 is unlikely to ever get out of experimental status). More testing follows with 1.2.6 (and testing only). * when '''both ALS and launch functions are activated''', if ALS_OVER_LAUNCH ( config11 bit3 ) is set, depending on TPS: ** if TPS < TPSx ALS is authoritive ** if TPS > TPSx launch is authoritive ** where TPSx is als_retard_maxtps or als_cut_maxtps (whichever is higher) ** in other words: in the configured ALS TPS-window => ALS acts, at higher TPS launch will act. ** note that because of this, depending on the setup, at some points higher TPS can result in lower RPM ! Provide feedback (Fero, especially you !) * Mazerati was reimplemented in 1.2.6 (under testing) for very quick sync at startup and ** using 10BTDC teeth during cranking (like a kickstart motorbike). ** If RPM > cranking_thres than old behaviour applies, eg. time from short-gap (and per-cyl sparkdelay). ** '''config also changed, use latest ''': ** start from a 6 cyl config ** apply '''configlet from vemstune''' (bottom of primary trigger dialog), which overwrites config13 variable so after that enable alpha-N, matretard is necessary. Plus, just in case, check these: ** ODDFIRE set ! ** next trigger tooth=2 ** reftooth: 0 10 8 6 4 2 .. .. ---- '''[http://cell.dyndns.ws/f/ 1.2.5] was experimental, now obsolete, WILL NOT BE RELEASED''' - * '''new "trigonLONGgap" option''' implemented in InputTrigger/ShortGapTrigger (that is coiltype + fiatstilo/subaru + advanced missingtooth filter => this configuration traditionally triggs on short gap) set trigger_multitooth_Nminus1 (bit4 = 1) effectively '''primary_trigger=BB instead of AB''' ** trigonLONGgap means: '''throw away short gaps, and trig on longer gaps''' (somewhat like the missing-tooth detection, but unlike a proper N-1 or N-2 wheel or the best MembersPage/YasecElise , don't assume that there is a sufficient number of sufficiently placed short-pulses) ** could be useful for MembersPage/FPhil/OddFireWastedSparkEvilStartUp apparently timing measured from the pulse after the 90 degree gap is better (than timing measured from the pulse after the 30 degree gap) ** would also be useful for Nissan 350Z V6 engine where 10,10,10,10,10,10,10,10,10,10,20 crankdeg pattern repeats (yes, that is 3 missing teeth 120 crankdeg apart so not suitable without camsync even for wasted spark or igndualout limphome) *** hard to tell out why Nissan didn't use a proper 36-1 or MembersPage/YasecElise or InputTrigger/SubaruThirtySixMinusTwoMinusTwoMinusTwo trigger, but at least an improvement from the terribly unprecise (even at normal running and high-RPM) Nissan-practice of timing from cambelt-driven camshaft where cam oscillates with 110 msec period. * fuel consumption on LCD page11 ** l/h ** and for the wheelspeed fw also l/100km (which is == cl/km that is centiliter/km, or centiliter/mile if wheelspeed is calibrated to miles per hour). '''[http://cell.dyndns.ws/f/ 1.2.4] - was experimental, now obsolete, WILL NOT BE RELEASED'''. SubaruEj (or maserati) should use newest firmware, other installs should use a released firmware. '''Workaround for the weird Subaru''' --design-- evolution where the measurement of the '''4th trigger''' is required because the camsync does not move with cam actuation * [http://vems.hu/vt/help/v3/v3_camshaft_angle_control.html page bottom shows: third trigger = SDA = Atmega128/pin26] ** the 4th trigger is next to it (SCL = Atmega128/pin25) same as wheelspeed2. * at configured "sectrig measure tooth" firmware measures wheelspeed2 pulse position instead of sectrig when speed2_sensor=01 ** so configure in VemsTune '''Base Setup/Speed Sensor/2nd wheelspeed sensor calibration = 1''' ** also configure in VemsTune'''Base Setup/Speed Sensor/2nd wheelspeed divider = 1 (for future compatibility)''' ** just in case it's not obvious: '''DO NOT configure 2nd wheelspeed sensor calibration = 1 if not having 4th trigger sensor'''. Value=0 to disable, or actual calibration if using 2nd wheelspeed input as wheelspeed input (frequency measurement) ** when ordering assembled controller for such Subaru, write in the order comment: "suitable for subaru shortpulse triggertype wheelspeed2 input and VR=>HALL divby1". The default aggressive hardware filter or the VR=>HALL divby4, which is perfect for wheelspeed input, would prevent this operation so this is very important. ** otherwise everything is same as for [http://vems.hu/vt/help/v3/v3_camshaft_angle_control.html sectrig based intake cam control] proven in BMWs, Ford, Honda, Suzuki, etc... (target table, measure tooth, output configuration). The only difference is that with this "2nd wheelspeed sensor calibration = 1" configuration the position of the 4th trigger pulse (connected to 2nd wheelspeed input) is measured instead of the sectrig pulse * ButtonImmobilizer '''keep-alive=enabled mode supported again''' (was not supported about 1.2.0 -1.2.3, accidentally) ** after enabling, test that after >30 second powerdown ECU is in "safe mode" after boot (and only allows running after touching on I-Button with configured ID) * when 1-wire IButton antitheft is enabled '''warning_light flashes''' after powerup before 1-wire Ibutton touched on (when injection and ignition is prohibited) to remind the driver (antitheft mode) ** Note: warning light '''must be p259 output''' for this to work, because other outputs are disabled in this state * wideband change: comes up faster if ECU powered down only for a second and sensor is still very warm. Not all sensors behave same under these conditions and this might not always work as expected (remember, if wbo2 does not measure sensor is likely to get clogged so revert ASAP if experiencing that), so we're collecting experience and will apply improvements ** also improved under extremely rich conditions (like 0.65) ---- '''[http://cell.dyndns.ws/f/ 1.2.3]''' - highly experimental, Viper only (for anything else use 1.2.2 or older, really) * Subaru type (or short-gap, similar to maserati, can start with the "maserati" configlet at the bottom of "primary trigger" dialog) * advanced filtering=enabled ** with disabled it won't work * nr of wheel on tooth=10 * tooth width=72 degree * set ign channel count to 8 to be able to set variables (finally change back to 5) * reference tooth=0,9,8,7,6,5,4,3 ** note: the missing 2 entries: 2,1 are wired in for this trigger * ignout=0 1 2 3 4 0 1 2 (just example) ** note: the last 2 (idx 8,9) entries are alias idx3,4 so 3,4 in this case (5 wasted spark channels are used but all 8 configured +2 aliased internally in firmware) * '''only change ign channel count to 5''' after setting all 8 channels of reference tooth and ign outputs. * inspect triggerlog carefully, and preferrably start with benchtest ** benchtest with Viper "Z020" trigger: seems ok ** configlet (bottom of primary trigger dialog) will be available for Viper in new VemsTune to adjust all the above several settings with 1 click ---- '''[http://cell.dyndns.ws/f/ 1.2.2]''' - TESTING - very nice for audi 135 "c270" trigger Beware: triggers other than the 135 tooth auditrigger are not well tested at this point (and 1.2.2 offers little improvement for them anyway), so benchtest before playing on real engine. 60-2 seems good at least. * note1: TDC after the trigger (ign_tdcdelay) might have changed ** to the safe side: without adjustment, 1.2.2 ignites 1 tooth (=2.67 degrees) later than earlier firmwares ** If unsure, leave earlier config, only compensate after strobing ! * note2: '''divby9 is recommended for 135 tooth''' - which means tooth width=24, tooth count=30, next trigger tooth=6 and related reference tooth array * note3: '''only divbyN is supported''' (so even for divby3, which is not recommended from now, and not well tested) divbyN and divider=3 must be excplicitely set * for non-135 tooth, the recommended divbyN range is 5-120 (max 128? but practically 5-65 or upto 71 anyway). DivbyN N=2 needs benchtest before use ** uploading [http://vems.hu/files/MembersPage/MarcellGal/Audi/etc/5cylAudi_c270_divby9.txt divby9 configlet] or [http://vems.hu/files/MembersPage/MarcellGal/Audi/etc/5cylAudi_c270_divby27.txt divby27 configlet] allows changing only the divider related parameters without ruining VE and other tunings. (in new VT, will be on bottom of primary trigger dialog) * '''[http://vems.hu/files/MembersPage/MarcellGal/Audi/etc/v3.3_u003637-2012.06.28-10.44_14800RPM.vemslog vemslog 135 tooth c270 running at 14800 RPM]''' with [http://vems.hu/files/MembersPage/MarcellGal/Audi/etc/v3.3_u003637-A-2012.06.28-10.44.36divby9.vemscfg this divby9 vemscfg] ** This 2 million pulses per minute (14800 RPM with auditrigger) is on table benchtest with a clean generated signal and both wheelspeed calibration=0 (off). ** With real "hazy" signal and wheelspeed calibration non-zero (enabled), the maximum RPM is obviously lower, but the mechanical stresses blow up the engine before the ignition delay deviation due to high-RPM would be >0.5 crankdeg. ---- '''[http://cell.dyndns.ws/f/ 1.2.1]''' - intermediate non-release only for audi 135 tooth "c270 divbyN" (N=9, 27 or possibly 3) * For audi c270 divby3, the trigger mods and motorsport features like camshaft-angle and multiple wheelspeed inputs lowered peak-RPM. In 1.2.1 we made +n*100 RPM possible (still not 11k RPM like we had on bench earlier, but >9100 RPM on bench). Config checklist: ** make sure HALL-dirac is OFF (both triggers) ** sectrig_maxrpm=40 (enable nissan temporarily to set to 6400 RPM, disable nissan) ** reboot ** Note that sectrig camshaft-angle measurement (would be useful for adjustable cams) is off above sectrig_maxrpm (eg. 6400), and triggerlog sectrig might be lower precision even at lower RPM. '''1.2.2 pushes auditrigger RPM-limit further (and should make 1.2.1 obsolete)''' ---- '''1.2.0''' - for testing (test results welcome, of course) * fixed 1.1.99 issue about "comm sometimes bailing out and sending CRC error (after about 30 mins in lab, some installers reported 5 minutes in field) ** engine running was not effected by this, but annoying when tuning, so 1.2.0 is better than * ButtonImmobilizer '''ADC keep-alive=enabled mode NOT supported''' (=> from 1.2.0 to 1.2.3 ADC keep-alive must be disabled for the immobilizer function to work) ** this was by accident, sorry (and thanks for reporting, support immediately added back at 1.2.4). ---- ((GenBoard/UnderDevelopment/FirmwareChanges 1.1.x)) OLD versions 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.