In my final despair, ( and onto Marcell proposal ) poured an older configuration the ecu, and the trigger got fixed mistake!
I copied it the dump under the other one.
greeting: Ákos
The mistake because I write the successor: The engine is running on idle speed, and the tachometer the megatune jumps onto 0. I did next measurements: I measured GND 11-17mv changes. I checked the polarity of a trigger. If the engine is running on idle speed, the trigger compared to the GND 4 mv shows tenseness, and when the turn jumps onto 0, then 10-12mv . The resistance of the trigger 550 ohms, his tenseness compared to the GND 144mv
greeting: Ákos
I made some technical modifications on the car. Turbo k26/I exchanged it for 28, 727ccm Siemens injector got involved into it, we prepared new harness, and 6 coil on plug got the car.
BMW 6cyl firmware upgrade from 1.0.46. to 1.1.18
note: It might be a good idea to avoid accented characters in page titles
- 6 cylinder
- firing sequence, 1-5-3-6-2-4
- primary trigger: 60-2, VR
- secondary trigger: not used
- ignition: 3 power-output channels, wasted spark
Ignition output channels. PLEASE SORT TO FIRING SEQUENCE, and WRITE CYLINDER NUMBER TO EACH!
- EC36 Pin10 ignch7
- EC36 Pin34 ignch2
- EC36 Pin36 ignch3
Please verify that the chosen channels are power-outputs. Measure (with ignition ON, but engine NOT running) that voltage between +12V and the ign outputs is 0 (a logiclevel output pulls down to 0V when inactive, so that would measure 12V !)
1.0.46 firmware works great with
- ignchmax=02
- h[2]=07 03 02 07 03 02 .. ..
1.1.18 firmware plan:
- ignchmax=05
- h[2]=70 30 20 70 30 20 .. ..
- h[1]=28 14 00 28 14 00 .. ..
- important: you might need to rotate either h[1] or h[2] to get the spark to the right cylinder(-pair). Document the wiring above (write cylinder-numbers) so we can give better hints
Engine does not start, shows wheel-error.
Trigger hints:
- normal tooth width: 24 (=0x18, since 6 degrees)
- missing tooth width: 72 (=0x48, since 18 degrees)
- engine phase when to reset: 160 (reset_engphase_after=A0), though this should not cause no-start
Analog input
- MCP3208 ch6 connected to EC18pin ..
- ALS or launch works well with this (if channel 6 is selected)
- TODO: verify that it measures +5V when not pulled down externally
- measures ... V when pulled down externally with 1kOhm (you can use other resistor from 100 Ohm .. 10k, just document)
- MCP3208 ch7 connected to EC18pin ..
- TODO: verify that it measures +5V when not pulled down externally
- measures ... V when pulled down externally with 1kOhm (you can use other resistor from 100 Ohm .. 10k, just document)
- neither ALS nor launch works if channel 7 is selected. What can be the problem ?
- maybe input was killed ? The mcp3208 inputs are very sensitive, connecting to +12V or electrostatic discharge can damageit.
You can upload the dump and logs via FileArea
NOS thoughts
- NOS output control
- and RPM and MAP (TPS?) ranges
BMW M20 B25 (6cyl, appr 2.5L) modified bore ( 81 mm ) now competing in 200m and 400m dragrace (earlier it was a rally car with restriction before a smaller turbo)
- high-grade forged pistons
- fire-ring-headgasketseal
- injectors 650 ccm
- factory IX drivetrain
- final drive 3.91
- whwel diameter 650 mm
- Turbo: schwitzer S2B
- external wastegate ( 35mm )
Would NOS be useful to kick in for a short time after launch ?
During launch, engine RPM drops so low that turbo doesn't work well for a short time.
When the turbo spools up and boosts, NOS could be switched off.
Would this be useful ?
Not sure why he does not use higher launch RPM (maybe he fears the clutch would explode?)
´
Be careful with nitrous oxide if the engine bogs down on launch, even a small nitrous shot will give a very large boost in torque at a low rpm. The cylinder pressures could get very high and detonation can easlily occur. -Jörgen
Choose output for N2O - see GenBoard/Manual/DigitalOut/Table
The guy who installed did not document the wiring.
We reverse engineer from the mcd/mct dump.
This is a 6 cyl engine. The 0F in h[0] looks like it came from a 4 cyl engine. (Roland has a strong 1.8L 4 cyl BMW turbo)
h[0]=0F 00 00 00 20 80 10 00
- injectors A-B-C-D are apparently fired in batch (all 4 at once)
- how are injectors connected ? Why are 4 injoutputs fired at the same time ? Why not 3 or 6 ?
- act_rpm_channel=50 mask 0x80 (128) = EC36 pin 17 Inj H
- als_tks_channel=60 mask 0x10 (16) = EC36 pin 9 Inj E
N2O enable
- mask 0x20 (32) EC36 pin 18 Inj F
- choose miscoutput channel=INJ/4 (as 20 is already in the h[2] at place 4 - remember it's numbered 0..7)
unused:
- mask 0x40 (64) EC36 pin 6 Inj G
It turned out
- ecu was used unclamped (!!)
- loaded with 6 cops, each fired once per crankrot (twice as many as cops are designed)
- ignition outputs had no protection (30A fuse is no protection. 5-7.5A per output is needed). Even the 30A fuse was blown once
Almost a miracle that the car was running after the event. (only running at NA, and missing something - likely spark - at boost.
http://www.vems.hu/files/MembersPage/poweredracing/megasquirt200804191333%206%20trafó%201.msq
http://www.vems.hu/files/MembersPage/poweredracing/datalog200804191325.xls
Working mcd mct dump
Deleted (obviously bad) confusing dates ( 2008.05.06. and Sat Apr 14 14:23:51 CEST 2007 for same mcd_mct dump )
To avoid confusion, delete the bad and leave the good (or make note of which is which).
... post good mcd mct here (so it can be identified and diff-ed to bad one) ...
check these
- als_launch_input=77 (same input mcp3208ch7 for ALS and launch ? Seems invalid)
- wbo2_pump_pid_kd=08 (using suboptimal/bad wbo2 pump PID settings)
Apply \n
wbo2_pump_pid_kp=15 wbo2_pump_pid_ki=4E wbo2_pump_pid_kd=00 wbo2_pump_pid_ilimit=FF
Suspicious parts of the diff -b -B -U2 (if diffing the good files)\n
--- mcd_mct_bad.txt +++ mcd_mct_good.txt ... -als_launch_input=76 +als_launch_input=77 -knock_retard_step=FF +knock_retard_step=00 # tach gauge now showing too low RPM ? -tach_div=10 +tach_div=17 -shiftcut_time=00 +shiftcut_time=01 # ignchmax=02 in both ignchmax=02 -h[1]=28 14 00 28 14 00 00 00 +h[1]=28 14 00 FF FF FF FF FF -h[2]=70 00 10 70 00 10 00 00 +h[2]=70 00 10 00 00 00 00 00
Cannot see obvious errors though. We should get back when it is clear which are the old and good dumps.