_____ |_ _| | | | | _| |_ |_____|
_ _ ( ) _ ( ) | | ( ) | | | | | | | | | (_/ \_) | `\___x___/'
___ / _ \ | | | | | |_| | \___/
__ __ \ \ / / \ 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 36-2-2-2 subaru trigger (like the missing-tooth triggers, or even none-missing tooth triggers) '''support sectrig-position (and third-trig position) measurement and therefore suitable for camshaft-angle-control (VVTI / AVCS)'''. 36-2-2-2 supported since early 2011, with original tigger-pattern specs. (improved syncup from firmware 1.2.27) ---- '''Summary for Subaru 4 Cyl with 36-2-2-2 crank trigger''' * Early engines '''without''' AVCS (variable intake cam) have a single VR sensor on the left intake camshaft pulley just behind the timing cover. * Later engines '''with''' AVCS have dual hall sensors that read a 4-1 pattern of slots in the camshaft mid section of the camshaft. **The AVCS equipped engines have the teeth on the cam wheel still, so it's possible to fit the VR sensor to read the cam wheel if you're not using AVCS or have the cam timing locked. This is uncommon though. EU/jdm ej207 engine 2001-2006 runs 36-2-2-2 crank vr , left cam second vr cam sensor+ A vr sensor on right and left inlet cam for feedback for avcs with a 4tooth trigger none missing ---- * See InputTrigger/SubaruEngineHsix for '''Subaru H6 engine''' that looks similar, but actually different: (19, missing, 10, missing,missing) We do '''NOT know any 36-2-2-2 subaru trigger that cannot be configured''' with VEMS. However, there are several types (and the subaru manual has been found incorrect at times, even with extremely important details like the pattern ! Despite this, all setups with sufficient information in the sufficient format have been made to work; and almost certainly will be in the future too; most likely with appropriate config; or at worst with fw mod) * publish '''details about your setup and upload triggerlogs''' (captured during cranking, with ign+inj fuses removed) '''to your project page''' if you have problem setting it up ---- '''Subaru H4 sequence ''' : (13, missing, 16, missing, missing) ('''also used in some Suzuki''', see MembersPage/SuzukiValveControl ) From firmware 1.2.12 implementation got more versatile but '''secignore must be specified (IMPORTANT)''': "Ignore sectrig pulses above" = 62 * '''secignore=3E''' ---- Please write here when you find out good values for * measure tooth for sectrig * measure tooth for third-trig values for a certain engine model ** no big deal to adjust though: usually gives 10-30 degrees trigger-position at idle, and +40..60 degrees - or so - higher reading when cam is advanced ** after VR-trigger is set up it's easy to set measure tooth to a good value (often many good values exist) after inspecting triggerlog ---- There is a contradiction about crank-pattern inside the Subaru manual (thanx for everyone contributing so we finally sorted it) '''Cam info''' '''There is a camsync pulse 45 crank degrees before Cyl 1 TDC, then the second comes 135 crank degrees after TDC''' Camsync: http://img.photobucket.com/albums/v41/Denmark/Vems%20pnp%20Impreza/subaru2001-camwheel.jpg This scopeshot shows VR, but some Subaru have cam-HALL (for sectrig and 3d trig, see AVCS below). The relative position of the cam (having the camsync pulse coming inside the middle of the long gap) was a reasonable design decision. '''Crank:''' * 36-2-2-2 setup: 36 crank teeth ** missing 2 teeth BTDC Cyl 1 (long gap from 60 to 30 BTDC cyl1) ** and 2 teeth twice before BTDC Cyl 3. (long gap from 60 to 30 and 30 to 0 BTDC cyl3) ** Seen on North American 2.2L SOHC plants (99-01 model using 2 teeth cam wheel) as well as other 2.0L/2.5L Turbo plants (most using 4-1 cam wheel) ????? http://img.photobucket.com/albums/v41/Denmark/Vems%20pnp%20Impreza/subaru2001-crankwheel.jpg http://img.photobucket.com/albums/v41/Denmark/crankwheel1modi.jpg On both pics with the wheel, the '''wheel turns anti-clockwise'''. The engine runs clockwise, but the pic is of the back of the wheel. The engine and the wheel physically turn clockwise, but '''the above photo shows the back of the wheel, so looking at the pic, the wheel turns counter clockwise'''. That confirms: '''our implementation assumed the above photographed wheel rotates COUNTERCLOCKWISE so the double long gap is followed by 13 tooth, not 16.''' ---- Crank Sensor wheel drawing, as mounted on the engine - rotates clockwise: ('''Beware: the plotted function is wrong''': it shows a group of 13 teeth, instead of 16! This caused a lot of confusion) http://img.photobucket.com/albums/v41/Denmark/Vems%20pnp%20Impreza/subaru2001-crankwheelpic.jpg ---- '''just thinking''' The choice to have the long gap detected at 30 BTDC and 0 TDC is weird. * for fastest sync during cranking, 40 BTDC and 10 TDC would make most sense. * for most precise timing during the whole range (and also for programmer's convenience) 90 BTDC and 60 BTDC would make most sense We can look at it as 15+3,3+12+3 * TDC is 30 crankdegrees after the single missing gap ** in other words, TDC is the 4th tooth after the single long gap ---- Subaru 36-2-2-2 '''implemented with camsync''': derived from c024 (12 cranktooth + 1 campulse) '''Original implementation notes''' - might be useful, but the configuration with the configlet is much easier now * |http://www.vems.hu/download/v3/firmware/experimental/ 1.1.76 experimental] works ** 1.1.78 will be modified and tested later ** start from a sane config, than at the bottom of primary trigger dialog click the subaru configlet (so primtrig subaru mode), than change: ** nr of tooth on wheel = 24 (yes, 2*12, not 30 !) ** tooth width = 30 degree (=3 * normal tooth) ** another_trigger_tooth=6 if 4 cyl *** (3 if 8 cyl, but assuming 4 cyl here) ** '''reference tooth from top to bottom: 0 18 12 6 ...''' ** secignore=62 (62 for H4 engine; or 94 for H6 engine) ** optional: disable camsync. *** But beware: it is ment to work with camsync. Might or might not be possible to set up to operate without camsync. With COPs, use camsync as ign-dualout overloads COP ignition coils easily. Than "save" + "burn", and might need to reboot. Please measure+figure how to set trigger tooth and TDCdelay. (tooth 0 is after the long gap after the group of 16 normal tooth). So triggertooth=4 + 90 deg TDC-delay or triggertooth=5 + 60 deg TDC-delay SEEMS appropriate. (The strobe will tell) ---- '''Test pattern''' - codenamed "s362" in wav filenames and signalgenerators Here is a [http://www.vems.hu/files/MembersPage/JasonRussell/Subaru/subaru_36-2-2-2_trigger-invert.zip wav file] (with CamSync. Inverted to play properly. 735 RPM) * this wav does not match the above spec ** the campulse comes in the first long gap, not the second. But it changes even within the file. The first crank rotation was good. The second crank rotation had the wrong number of teeth after the double gap. A cut and paste error. *** Note the wheels in the above photo rotate '''anti-clockwise''' since looking at the back. The drawn wheel turns clockwise. The wrong time-line in the subaru manual next to the wheel drawing was the cause of the confusion. ** So 16 pulses, '''gap, tooth 0, another gap,''' 13 pulses, gap, and it restarts ** verify any pattern (not just the first cycle), and reverse in time if necessary, also invert in audacity if needed. ** Than sectrig-triggerlog showed pulsetrain is reasonable for some testing. ** 2010-06-16 and newer VT also plays 36-2-2-2 in '''"Tools/Play Trigger"''' if '''s362''' type is typed in the combo ---- '''Testing''' - some [http://www.vems.hu/files/Firmware/config/trigger/Subaru6Plus7/vemscfg_test_2010.06.18-s060-s362_1.1.76_1.1.78good.zip configs] that were used for testing Since the '''2010-06-15 experimental 1.1.76 seems to act properly (36-2-2-2 with camsync or with subaru 6+7 with camsync). Can be tried on bench or engine.''' * 1.1.78 (since 2010-06-18 experimental) seems to act well on s362 (36-2-2-2 with camsync also with subaru 6+7 with camsync). Subaru code is basically same as in 1.1.76. Test on bench first, if you intend to use this. (and remember to review all the new settings like per cylinder ign-delay and fuel-pw-adjustments) ---- '''Jason's notes''' * Jason can get it to trigger with camsync disabled, and can enable camsync and the rpm becomes closer, but trigg errors. If you stop stim signal it will not register rpm again with camsync enabled. Here is a [http://www.vems.hu/files/MembersPage/JasonRussell/Subaru/v3.3_n002843-2010.06.13-00.20.52.vemscfg vemscfg file] of the setup described above (trigger settings) In Jason's engine there is the possibility of using a 4-1 cam signal. It's for the 2004 WRX STi variable cam system (disabled on this engine for now) and those sensors are still there but unused. We should figure out the 2 pulse camsync, but if the 4-1 is easier to handle quickly, I'm ok with that. There is a 4-1 on each cam, but I expect it's sufficient to just read one of the two (just like this 2 pulse cam trigger). http://www.vems.hu/files/MembersPage/JasonRussell/Subaru/36-2-2-2withcam4-1.gif 2013-09-24 : '''triggerlog captured with primary crank vr , right cam sensor as 2nd trigger Hall, and left cam 3rd trigger Hall (exhaust cam)''' Assuming the triggerlog was captured with fully retarded position, it seems a good idea to swap secondary and third trigger (use the red pattern for sectrig) * primary 36-2-2-2 * rightcam 4-1 ** secondary (white) on the triggerlog. Proposed: third trig * leftcam 4-1 ** 3rd trigger (red) on the triggerlog. Proposed: secondary trigger * secondary and 3rd do both move when avcs is activated! ** that's good ! ** It is very likely to get it start that way, but unfortunately sectrig is a bit too close to tooth0: that can result in trigger error and loss of sync due to cambelt sloppiness at sudden angular accelerations ** if that is the case, please report it. Secignore=F6 or similar (240-250) originally for missingtooth wheels might help (but need testing for this, and unlikely to work with 1.2.16-1.2.17) http://img.photobucket.com/albums/v41/Denmark/null_zps564e2f12.png The settings in the dropdown menu does not work here!! This is a 3 trigger setup were left and right cam Will love when avcs is activated , and not like the 4trigger setup were the 2nd trigger does not move There are 2 types of cam sensors: * '''Type 1 : crank 36-2-2-2 and cams 4-1 Hall with avcs''' ** Primtrig is known to work, and should be possible to configure sectrig also. Will bench any setup: Write details and triggerlog => project page * '''Type 2 : crank 36-2-2-2 and cams 4 tooth even spaced with avcs (known to work) ---- * from 1.1.92 VVT has special firmware support, the OLD Subaru implemetation changed, ** trigger event is the first pulse after the double missing if the cam trigger received in the tooth detect zone http://www.vems.hu/files/MembersPage/JasonRussell/Subaru/subaru_vvt_a.PNG http://www.vems.hu/files/MembersPage/JasonRussell/Subaru/subaru_settings.PNG * For valid measure Right sensor must connect to secondary trigger input ** Connecting the Left channel to secondary trigger input can measure as [http://www.vems.hu/wiki/index.php?page=MembersPage%2FSuzukiValveControl Suzuki] trigger, needs to re-configure tdc and tooth-wheel too. * VVT measures from the trigger cam tooth to the configured measure tooth ( default = 3 ) * Bench tested [http://www.vems.hu/files/MembersPage/JasonRussell/Subaru/v3.3_s001212-005-SDCard-2011.05.25-10.00.25.vemslog vemslog] and [http://www.vems.hu/files/MembersPage/JasonRussell/Subaru/v3.3_s001212-2011-05-25-09.57.20.triggerlog triggerlog] * Assuming that is the most retarded position. (correct ?) link to a site, were there is crank 36-2-2-2, plus cam 4-1, http://www.msextra.com/forums/viewtopic.php?f=131&t=36430 this is what there is on the subaru´s when the run inlet avcs, so NOT 36-2-2-2 and the 2pulse cam wheel A triggerlog, where * apparently B is connected (not A). B would require different config (derived from the suzuki setup), so connect A to sectrig if possible. * also, there is an offending extra pulse ! ** maybe because VR input is used with a HALL input ? ** we need to find the cause of this pulse to cure the no-sync problem (otherwise the primtrig is detected and RPM seems correct) http://vems.hu/files/MembersPage/BaldurGislason/triggerlog_subaruimprezasti_2004usdm_ej25.png ---- '''Camshaft-angle control''' Firmware originally supported camshaft-angle control with missing tooth primary trigger. But someone on irc requested 36-2-2-2 (without writing details here or there). We added firmware support for 36-2-2-2 (with the factory 2-pulse camsync) camshaft angle control (1.1.82), but there are still open questions... * upto 2008 only intake cam is with avcs(vvt) * exhaust cam is not controlled ** good, because now the special 2-pulse campattern is supported for intake cam only (only the normal 1-pulse campattern is supported for the exhaust cam). * the max possible crankadvance is 46degrees(23degrees cam) Engine has '''2 VR + 2 cam-HALL sensors in total''': * '''36-2-2-2 on crank''' * '''2 cam-HALL sensors (only on US model)''' ** Jason from USA ordered VR,HALL,HALL for "4 IGBT, manual throttle/no ETC, Subaru 36-2-2-2, and needs 3rd trigger input for second intake camshaft" ** others run VR as previus stated!!: 4-1 pattern (see pic above) on '''right and left inletcam''', at the back of the cam (for cam position feedback) * 2 pulse camsync (secondary vr on left camwheel) ** this is constant position, does not change when the cam PWM solenoids are actuated, right ? ***yes this was used for camsync and is constant, I believe the second trigger is a leftover from engines not running avcs,as this have also been removed from the later engines,that all run avcs,the new engines only run the cranktrigger, and the 2feedback sensores on the right and left cam ** please confirm that this is not needed when setup for camshaft-angle control (probably used as backup in the factory setup, maybe to better tolerate loss of cranksensor signal ?) ***Yep, not needed '''Trigger inputs - 3 should be enough''' * 2 onboard (primary+secondary trigger) * and one external [http://shop.vems.hu/catalog/hall-p-162.html DSUB9 VR to HALL] is '''not needed because cam sensors are HALL!''' ** this 3d can be connected to the 2nd cam-input, traditionally called "exhaust cam input" - even if it's not exhaust cam in this application. (and at least 1 wheelspeed is still available) ** but the special pulsetrain must be implemented for the "exhaust cam". '''Output solenoids''' ? * 2 solenoids (both are PWM) one on the left side and one on the right side ---- 1/5-2012 Need help to to solve low Vr signal. Question 1 As you can see below images of the vvt cam vr sensor is not a strong signal positive 200 mV negative 300 mV. * actually hard to tell the DC bias. But it seems very likely that because of extremely high sensor resistance=1.9k '''the primtrig DC bias'''= 5000 * 1.9 / (1.9+18)='''480 mV''' is too high. ** Solution is easy: use weaker pullup R30 >= 47k or higher pullup resistor instead of 18k. Even removed R30 (which is infinite Ohm) can be tried but that adversaly effects noise immunity so max 100k is preferred. (hint for a new install: order with eg. "R30=100k" in the order comment, no extra charge) What can we try to do to make the Lm 1815 read the signal? * adjust the DC bias (pullup) to make sure the positive peak is > +150 mV (+250 mV or higher even better), and the signal goes subzero (< -20mV at the negative peak, also important). Question 2 This version of Subaru is not like the US made ons as shown on this side. This Subaru is sold in Europa and Japan. It has 4 spike´s on cam distributed 90 degrees and not 3 like the US models. Also all sensors are Vr. Type (US models uses hall on cams) * '''4 pulse on cam''' (evenly spaced 180 crankdeg) '''is suitable for third trig, but NOT suitable for sectrig''' ** it's not a firmware question: the pattern MUST be different in the next crankrot. Absolutely required to be useful for sectrig (engine-sync) ---- 36-2-2-2 crank, plus a 4tooth camfeedback on each cam, This trigger setup is used on JDM and Euro 2.0 AVCS STI engines from 2001-2005 were it changes to hall cam feedback sensors with 4-1 feedback on each cam.(like the above US version) scope shot 1 : vr cranksensor 36-2-2-2 + right VR cam feedback sensor connected to secondary trigger measured with ecu pluged in http://img.photobucket.com/albums/v41/Denmark/f4480a56.jpg scope shot 2: VR cranksensor 36-2-2-2 + left VR cam feedback sensor connected to 3rd trigger connected by VR-Hall converter this pic is direct from sensores without ecu connected http://img.photobucket.com/albums/v41/Denmark/bcad1d6e.jpg Scope shot 3: VR cranksensor 36-2-2-2 + left VR cam feedback sensor connected to 3rd trigger connected by VR-Hall converter this pic is with ecu connected. http://img.photobucket.com/albums/v41/Denmark/558e660f.jpg '''Not getting any output through both cam LM1815 pin12''' Suggestions to get it working ? * 3 vr sensors are all 1.9kohm * very '''important: measure DC input voltage also''' ** primary 490mv secondary 0.1mv third 0.1mv * how is the LM1815 pin5 set up ? ( connected to GND, connected to +5V or left open ?) **tried with +5v and open(measurement is open) * how do you measure output from LM1815 pin 12 ? Pulses (depend on resistor*capacitor value, but ) can be extremely short ** especially the primary trigger is set up to output very short ~44usec pulses to allow 10000 RPM * 135 tooth = 1.35 million pulses / minute for auditrigger. ***we tried with peakdetect on scope only triggerlog also only shows primarytrigger firmware 1.1.96 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.