Building a BMW M10 racecar engine [Homepage]
2006-05-07 Car started, and everything seems to work great so far. Will start tuning soon, but cant find any information on how to set lambda-target table? Please advise! Running 12x12 standard and engine is normally aspirated with (soon to come) 100 hp/lit.
As this is a race car engine this is very simple. You can set the entire table to lambda 0.82-0.88, the same value in the entire table. If you want to sacrifice some response and improve emissions you can change all load sites below around 70kPa to lambda=1. Let all load sites above 3500rpm or so stay at the richer value (0.82-0.88). This will also improve your consumtion figures if you drive to the track. The idle quality will be better if you don't set that range to lambda=1. Going to 0.9-0.95 usually make the idle much smoother. Idle is of course even smoother in the 0.82-0.88 range but it can be a bit smelly.
- If you get an answer on this, please let me know. I just set mine to where I guess it would close ie: a bit leaner at idle and cruise, then high map at higher rpm, I set mine around 12.5:1. I turned on the ve learning and have been driving around and changing the lambda table to make it run right and let the auto tune do its thing. This is probably not even close to right but it was the only thing I could think of and the truck is running awsome now.
2006-02-16 Genboard is in car! Haven┤t cranked yet since the battery was low, but so far so good. Most sensors showed reasonable values and no fires started yet. A small problem with TPS-signal but i have got some leads to try out before I say its a problem. Configs and Tables is now available further down on this page. Please let me know if you see some wierd stuff in there..
No spark! Set up a distributor ignition (single coil) but cant get spark on the plugs. Put a diode on the coil to se if there was some sort of signal from the EC36 pin 24 wich I use as output, and yes, the diod "blinks" WEAK so there is some sort of signal. So I took the cable out from the EC36 pin 24 and hit it directly agaist ground (the body) without changing anything else and then I get i spark!
Seems like I have a bad or weak signal/ground?!
I have tried between 3 and 16 ms of dwell, doesnt matter.
I did set up the distributor ignition since my wasted-sparkcoil was broke in some way, couldnt get a spark from it in any way. Is there any possibility that this messed up the Genboard?
Put the same diod on a Injector output to se if there was signal, and then I get a GOOD "blink" on it!
Some components was put in the wrong place on the Genboard.. Now it works! :-)
Engine phase; According to the manual (link? maybe it's the old non-updated sourceforge instance?) it is 180 for 36-1, but using the formula "Engine phase = 2 * (dec * real teeth + dec * missing teeth)" gives (3x35+3x1)=2(105+3)=2x108=216.
216 is right for 36-1 trigger (must be a multiple of 72!)
Edited table http://www.vems.hu/manual/html/ch09.html#table.enginephase to show the correct values for 36-1
- 4-cylinder 2 valves/cylinder engine.
- 2058 cc normally aspirated.
- Rebuilt head with optimized intake and exhaust ports.
- Special cast intake manifold.
- TWM 45 mm throttlebodies.
- 304 Dr. Schrick camshaft.
- BMW S14 rods and custom made JE pistons.
- Genboard 3.2.
- VR cranktrigger from BMW M50-engine and triggerwheel made by Samuel Lindqvist.
- Coilpack from BMW M43-engine.
- BMW S14-engine injectors.
- TPS? -Not solved yet.
- Will run wasted spark since I dont have a camtrigger.
Config and Tables
Total newbie on VEMS, but gets lots and lots of help from Johan Rius (Thanks a lot!!)
Questions to Rikard
As we spoke earlier you were thinking of going with a distributorless setup, is it using:
- the missing tooth trigger wheel and wasted spark coils
- using a cam sync sensor and separate coils for direct fire ?
In either case you can still use the missing tooth wheel as the crank sensor (primary trigger) for a more accurate engine speed.
If you're going for a direct-fire setup, the easier trigger arrangement would be cutting of all but two teeth from the 36-1 wheel you have and use the Honeywell GT1 Hall sensor. These two teeth should pass the sensor about 60 degrees before TDC, i.e. place the sensor and then determine what teeth to keep.
A: I┬┤m think about running Wasted spark since it seems easier (dont need a camtrigger etc) and I also dont see a gain in hp by running direct fire. By easier I mean less machining of parts since I cant do this by myself.
Q: Thinking about using the Honeywell GT1 Hall-sensor, but my toothweel has a smaller "tooth spacing" and "target thickness" than the GT1-sensor needs according to its specifications. Is it possible to use it anyway since I┬┤m not even close to the 100kHz its supposed to handle? 8000 rpm is less then 5kHz with 36 teeth triggerwheel..
A: I guess you would have to try. With a multi tooth setup it's easy to watch the count of the wheel errors to determine if you have trigger sensor (or signal) problems. I know it's not very polite to rev the engine up to 8000 rpm in neutral as there is no load on the engine but it's the fastest way to find out if there is trouble. Any errors and it should appear in the wheel error count on page 2 of the LCD display as a kind of rev limiter as ignition (and fuel? don't remember) is cut when the Genboard looses track of the wheel tooth count. Here is good reading for finding any trouble: InputTrigger/TriggerLog
Thoughts about VR-trigger
I also have a VR-sensor that I first intended to use. It was installed in another car that was to be a VEMS-testcar, but it became a daily driver instead.. Im a bit scared of using a VR-sensor since they seem so much more sensitive to other things around them, but maybe Im a bit too scared..? The Hall-sensor seemed lika a god idea since it seems more foolproof..
The VR sensor works very well, you should not be afraid. Just connect all GND and GND5 from the same (strong - thus electrically close to the battery) point outside the ECM case. Also, don't connect the GND and GND5 inside the case. GND and GND5 used to be temporarily connected on non-preassembled genboards with a small arch at the middle on [this pic] right from the FETdrivers. This onboard GND and GND5 connection arch was for safety in case user forgets to connect GND5-s: some people forgot to cut the arch. Now the necessity of strong GND and GND5 connection is stressed instead (which was necessary for good operation anyway).
Actually, if you connect the VR sensor in reverse, it's revealed in 15mins by applying InputTrigger/TriggerLog at cranking. A badly connected HALL sensor can easily be damaged.
Also, there is now the '''advanced multitooth filter, see toothrel_normal and related variables (see GenBoard/UnderDevelopment/FirmwareChanges) that reveals any multitooth problem early (makes it easy to decide if problem is with trigger or other subsystem)
The sensor is taken from a BMW E36 with a six-cylinder engine (M50 I think. Anybody┬?┤s using this sensor?) and the triggerwheel is the same as Samuel Lindqvist is using on his M3E30 and it works fine. Looks like this:
How to display/log engine parameters
Q: Since Im building at race engine, it would be nice to get Oil temperature and pressure in the LCD-display.
- get the signals to mcp3208 (ADC for extra analog inputs) chip, pin7 (channel6) and pin8 (channel7) are usually used. See the onboard MAP, and secondary_trigger=HALL options for [assembled controller] standard EC18 pins.
- MegaTune logs the collected analog data so it can be analyzed (see LogAnalysis)
- decide on which LCD page, and which part the data should be printed, in what format
- displaying on the LCD is relatively simple, small changes in lcd_display.c
How about fuel pressure ? You don't need that ?
Fuel pressure would be great too! Actually, there are even more things that would be good to have such as oil temperature for differential etc.
Q: How many different input temperatures etc can the genboard handle and where do I connect them (on mpc3208)? I know the first two now.. Thanks a lot!!