3.5 years ago (!) when I first bought the VEMS 3.3, I had some project for my Volvo 850.
Since then I had a Maserati biturbo and so I was more involved during my spare time with mechanics than with electronics or software. Hopefully the car is 200km from where I live !
Trigger
This biturbo has an odd fire engine
- with bizarre triggers
- do you have more details on the trigger ?
- VR/HALL ?
- pulse positions
so that the distributorless biturbos produced at the end were more easily equiped by Maserati with 2 ECUS, each one dealing 2with an 1/2 engine which is then 3 cylinders even fire.
Recently I aimed to use Vems to measure C0% at idle in order to fine tune the A/F potentiometer set on the Weber injection ECU.
My first issue is that I did not get the good O2 percentage for the free air test !
After you measure connection (including that you have +12V on heater+), you need to calibrate:
- GenBoard/Manual/Config/WidebandCalibration starting value (from RCal measurement)
- MembersPage/PhatBob/UserGuide - calibration steps
Wiring
Document wiring here !!!
email conversation suggests serious miswiring (like fused grounds)
You should start from 0, documenting the whole setup
- every wire you connected.
- every measurement
mcd/mct dump
Verify wbo2 config, and publish mcd / mct dumps.
email conversation suggests possible misconfiguration (although not confirmed)
Wbo2 is working
(01/09/08) What I did.
Following your advices, I have restarted from the begining using initial the Wbo2 configuration, megatune miniterminal and the config and upload .bat files.
Initial config.
wbo2_warmup_ramp=9C
wbo2_warmup_target=FF
wbo2_abs_limit=E4
wbo2_limit_maxt=A0
wbo2_fallback=60
wbo2_retry_t=06
wbo2_edgetime_corr=00
wbo2_edgetime_min=50
wbo2_ri_target=00
wbo2_nernstdc_target=9C
wbo2_pump_pw_zero=63
wbo2_calibration=AE
wbo2_heater_pid_kp=45
wbo2_heater_pid_ki=11
wbo2_heater_pid_kd=1A
wbo2_heater_pid_ilimit=81
wbo2_pump_pid_kp=15
wbo2_pump_pid_ki=20
wbo2_pump_pid_kd=00
wbo2_pump_pid_ilimit=FF
wbo2_ri_confidence_scale=80.
1) pump_zero test : I get 44.1mV -> good
2) 2x100 ohm test : Pin5-Ground : 3.98V -> good
Pin1-Pin5 : 0.241V -> bad
3) nersnt_dc_target : 9C -> A0
edgetime_corr : 00 -> BA
Pin1-Pin5 : 0.241V -> bad
4) ri_target : 00 -> 96
Pin1-Pin5 : 0.233V -> bad
5) nersnt_dc_target : A0 -> C0
Waiting 32mn or so for an 02 value to display as it is said!!
The trick is that this value may be not displaid, one as to mesure a stable dc value at Pin1-Pin5, so I got
Pin1-Pin5 : 0.530V -> bad
and
nersnt_dc_target : C0 -> B0
Pin1-Pin5 : 0.385V -> bad
but Hourrah the set point of the pump control loop can be trimmed!
6) Then it is easy,
pid_illimit = FF (get rid of hard bound),
pid_ki 20 -> FF
as it is asked, there should be a trick in the algorithm since the set point has to be reached whatever the ki value
Pin1-Pin5 : [0.500,0.530]V(instable)
7) Next I lower pid_ki to get a stable value, and I adjust nersnt_dc_target.
I finally got
Pin1-Pin5 : [0.449,0.451](good)
using:
wbo2_warmup_ramp=9C
wbo2_warmup_target=FF
wbo2_abs_limit=E4
wbo2_limit_maxt=A0
wbo2_fallback=60
wbo2_retry_t=06
wbo2_edgetime_corr=BA
wbo2_edgetime_min=50
wbo2_ri_target=96
wbo2_nernstdc_target=9C
wbo2_pump_pw_zero=63
wbo2_calibration=AE
wbo2_heater_pid_kp=45
wbo2_heater_pid_ki=11
wbo2_heater_pid_kd=1A
wbo2_heater_pid_ilimit=81
wbo2_pump_pid_kp=30
wbo2_pump_pid_ki=40
wbo2_pump_pid_kd=00
wbo2_pump_pid_ilimit=FF
wbo2_ri_confidence_scale=80
This is not a fine tuning, since one should adjust the control parameters with respect to the frequency band which is needed (AFR value for each cylinder?, noise?)and I changed the pid_kid value for no reason.
9) finally I calibrate for the sensor using : calibration = BA.
I put the sensor 10 centimeters inside the tail pipe to measure AFR. I obtain, for one pipe an AFR value of about 16 (no O2 value), and an O2 value of about 9 on the other (no AFR value).
Too much air.
The exhaust pipe is new, so the exhauts manifold is cracked (which happens on the Biturbo) or the sensor still catch some air from the end pipe?
Q. How deep inside the pipe should be set the wbo2 sensor to give a true value ?
SJ7
Q.If I input Vems with the signal from the VR cam sensor, can I output the
calibrated pulses from SJ7 which seems to be connected with the
Ref Pulse Output of the LM1815 ?
A. yes, the output of the LM1815 is a good place to "tap" the signal. Just make sure it does not get 12V or other damaging signal.
Problem with the miniterminal serial link
. The serial link from the Megatune miniterminal to Vems used to work without problem on my older notebook (Windows NT).
I tried to make it work using my new notebook Toshiba wich is equiped with a serial port. After some attempts and plays (!?) I got in response to man the following garbage:
man $Pw ´´´´´´´$(Px ±²±±±±±(,Pw™ª«©©©©©,0Pv•¡¢¡¡¡¡¡04Pu�™›ššššš48PoƒŽ������8<Pgz�ƒ„„„„„<?P_oxxxxxxx?<P_oxxxxxxx:<P_oxxxxxxx:<P_oxxxxxxx:<P_oxxxxxxx
(2@ABCD (2<KZijklm
. I immediately tested back the Vems communications with the older notebook: I get now the same garbage in response to man, however the megatune display shows the gauges correctly and active.
The audit log seems allright :
+++
Comm port = 1
Comm rate = 9600
Timer interval = 50 (I have tested some other values)
+++
. Next I tried to communicate through the keyboard and the LCD. I get the Hello message on the LCD but I was only able to make work the left and right arrows of the KBD. The KBD is an AZERTY and is a COMPACT model, but even after AZERTY/QWERTY translation, mcd or other commands do not work. However I remember that the COMPACT KBD are special.
Q. For what reasons do I now get that garbage in response to man?
Thank you for your help
A: Type Man not man!
26/10/08 - Thats'it ! (it was an easy one ;) )
Same week end, I made a special tubing for the wb02 to insert into the tail type. I measure an AFR > 15. My exhaust line is new but the exhaust manifold is possibly cracked (very common on the biturbo). Same day, I did not have more success with offEngineTesting of the input trigger...
Output trigger / coil type input trigger configuration
I connected the stim-output signal to the InputTrigger with a 1K resistor as said on the Output Trigger page.
I entered msp0a and I obtained a nice rectangular signal period 20ms, then I configured
primary_trigger 0x01
tooth_wheel 0x04
tooth_wheel_twidth1 0x0A (I did not try 0x2D or 45 deg)
trigger_tooth 0x01 (why not)
secondary_trigger 0x02 ( coil type no cam)
Result : no RPM value on the megatune gauge !
I also tried the 36-1/no cam configuration and some others ... but that requires tooth_wheel=23 (decimal 35) !
5cyl no secondarytirgger "camsync" example. Only 1 ignout channel, distributer ignition, with 5 window campulses connected to primary trigger:
<PRE>
primary_trigger=03
secondary_trigger=02
tooth_wheel=05
trigger_tooth=00
another_trigger_tooth=01
crank_minper=60
tooth_wheel_twidth1=40
tooth_wheel_twidth2=02
cam_sync_r_edge_phase=01
cam_sync_f_edge_phase=02
reset_engphase_after=F0
</PRE>
Q. What is the configuration which works in this case? Which case ? Make a chapter for it and make it clear what trigger-type exactly you want to test. There are zillion 36-1 and 60-2 configs.
Instead of stim-out, a PC soundcard is often easier: ElectronicDesign/TriggerSignalGenerator
Trigger 4+0
With the help of the uk forum I realized that the trigger configuration "4+0" (why not to have the distributor still used ?) I was looking or more easily "4+1" should be handled only by the recent firmware version.
So I tried to upload version 1.1.43Alpha2
- Check everything running as usual
- Save of config and tables
- mci -> "Hardware serial number: V3.2-n00???"(I badly noted the end number) This surpises me since the order says V.3.3..
- mdV -> 1.0.23
- 1st run of upload-firmare-12x12.bat -> "LoAddr= 0 HiAddr=e73b" then "No hello from AVR"
- 2nd run -> ... "AVR said hello" "Flash erased" "-...-" then garbage/binary sent to the screen (1sec for each 80 bytes line), why not echo of the writing I thought, after waiting 4 hours I computed that this should be enough and I stopped the process.
- Power on/off and 3rd run -> "LoAddr= HiAddr=e73b" then "No hello from AVR"
- check megatune and miniterminal : dead
- run upload-firmare-boot-12x12.bat -> "didn't get reply from AVR, out of sync"
Now I am stuck...
Answer : If it successfully uploaded the new firmware, com speed is now 19200 and it outputs AIM binarydata. Check UK forum thread http://vems.hu/www.vems.co.uk/forum/index.php?topic=653.0 Also this trigger type isn't successfully tested so you are on your own with this.
- That 4+0 idea of configuration was for bench test only.
- Indeed the changelog of version 1.1.43alpha2 says that 19200 is now default. I tested the (nice) uploader script but I forced 9600 com1. I also tried to launch the Megatune but have the message "Megasquirt controller code version - received ???? from controller"
What about the 4 hours of binaries sent to the screen. It probably shows that the upload had failed. Is the boot firmware still allright set for 19200? What to do if not? Do I have to upload the boot firmware?
What about the mci telling my harware is version 3.2? Is Vems 3.2 compatible with the updated firmware?
Anyway some tests to do the another weekend ...
Answers: Did you change megatune speed to 19200 also? Binary data after firmware update is the AIM protocol stream that Vems outputs if it's not disabled from config or manually via terminal (send A, D, Q, S or Man at right speed). Bootloader still in 19200 and firmware is compatible also with 3.2.
- No, I did not change the megatune speed (9600 default). I stopped the stream by switching off/on (what is that stream for ?(. Hope it it was harmless?
1.0.46 configuration for 36-1 no cam
I switched for baud 19200 as one friend has kindly suggested and I got the Megatune Miniterminal link to run.However
- The numerical characters of the AZERY keyboard of my Toshiba Tecra A9 were not recognized by the miniterminal of the Megatune version from the 1.1.43alpha2 package.
- The Bray terminal was non usable because it was receiving that AIM binary stream as soon as it was connected without any chance to stop it
- Megatune 2.25 crashed when launched from my Notepad Dell Qwerty under Windows NT4.
Then I uploaded firmware 1.1.18.
- The numerical characters were not recognized by the miniterminal of the AZERTY Toshiba
- Megatune from the Dell qwerty was good, but the output trigger gives no signal for the stepper motor port (pin 4 E18) !
Then I uploaded firmware 1.0.46
- Megatune from the Dell qwerty good, output trigger for 36-1 good (check by a scope +5V -5V). I fed that signal to the input trigger pin through a 1kOhm resistance and I configured as said by the Section 4.7.d, that is
primary_trigger 0x01
tooth_wheel 0x23
tooth_wheel_twidth1 0x0A
trigger_tooth 0x01
secondary_trigger 0x02
another_trigger_tooth 0x12
tooth_wheel_twidth2 0x06
-> No rpm on the gauge.
Finally I thought myself that indeed I am feeling stupid running the Vems system but on another hand I was good enough to check my abilities on a test bench ;)!
Do you know where can I pick the configuration and table parameters + firmware version which allow me to bench test for sure using output trigger and for 36-1 no cam?
---
Is the output trigger test supposed to work?
Following Marcell advice I put a 4700pf capacitor on the line of the 0utput trigger signal which feeds the input trigger point.
After entering mst62msq21msp01msp00 (I have configured for 36-1), the rpm indicator went up to 25500 tr/mn and back erratickly to zero, the "trigger error" alarm was risen.
- Changing the value for msq21 or msp did not stabilize the indicator.
- Configuring for 36-0 makes the indicator constant and no trigger alarm?
- Doubling the "rpm constant" doubles the indicator value (nice,but how to choose that constant? measurement ?)
- Putting no capacitor, that is having simply a square input signal -5V,+5V, keeps the indicator to zero in any case.
My conclusion is that the Output Trigger signal is of no use for me to make basic testing of the firmware 1.0.46.
Note. In the case where nobody would be able to utilize the Output trigger signal for trigger testing, why to keep the page in the Wiki?
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++
02/04/2011
Indeed it is a slow project. My Maserati biturbo my 1987 is now working fine and I am actually looking inside the eprom of ignition ECU (Marelli Microplex 602F).
From time to time I play with the Vems Ecu on a bench which is easy now thanks to Vemstune.
The engine car is 6 cylinders odd fire, sequence for ignition is 150°-90°-150°-90°-150-90°.
Triggers are 6 pins (30°-90°-30°-90°-30°-90°) on crank pulley, 2 pins on cam pulley 25° apart.
I made a .wav file for testing
Does anyone know if Vems firmware is able to track this profile?
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
April 12 2011
This is great, Marcell has opened a new odd fire 6 cylinders page !! ;)
http://www.vems.hu/wiki/index.php?page=MembersPage%2FOddFireSixCyl"
There is a view for the ignition system of the biturbo there
http://www.eurospares.co.uk/partTable.asp?M=3&Mo=478&A=1&B=27127&S=
where you can see 2 of the 6 trigger poles which are set on the crank pulley.
The cam pulley can be seen here,
http://www.eurospares.co.uk/partTable.asp?M=3&Mo=478&A=1&B=27110&S=
however the 2 trigger poles are hidden.