Questions
Insufficient injector opening configuration misinterpreted as "number of bins" issue"
GenBoard/Manual/Config/InjectorOpening
It's not the power of the engine (100+ hp/L at 100kPa, but with forced induction) or the lack of vacuum (80kPa) at idle is that make things change rapidly.
The reasons for the bad curves (that are hard to map properly and take many bins):
- big injectors: req_fuel=11
- and wrong injector opening configuration
Injector opening is important here, because of the big injectors (at low kPa the pulsewidth is very low).
Here is a mct/mcd dump from the box (maps and config): http://www.vems.hu/files/MembersPage/MattiasSandgren/volvo_16v_race_vems.cap
injopen=00 and injocfuel=00 is a bad idea, especially injocfuel=00. Unquestionably it's a wrong direction to try known wrong injector-opening and compensate with VE. The tables show this clearly: the 20kPa VE is lower than 10kPa, VE starts to increase again above 40kPa.
Drop the 10kPa bin (so the 20kPa bin takes effect) and add injocfuel to get good lambda at very low MAP.
The worse injector-opening values => the harder the tuning for big injectors:
with proper injector-opening values 3..4 bins are perfect under 100kPa (it is proven on >100Hp/L NA racecars). With bad injector-opening values, more pressure bins would be needed and still result in suboptimal behaviour. In additionally, without interpolation maybe 100 bins would be too few. Other misconfigurations would make it impossible to tune with 1 million kPa bins.
The owner of the PV is used to Haltech and has mapped 9-10 cars previously, he is really disappointed and feels that even 16x16 would be on the thin side for his race engine unless he understands the proper way for this problem: importance of injector-opening effects. On his Haltech he had 32 fixed pressure bins, and the engine ran silk smooth. Maybe Haltech had no proper way for injector-opening configuration so a workaround was necessary.
It is absolutely sure that it's possible to tune this powerful engine perfectly with 8x8 configurable bins if interpolation is used.
8x32 (rpmxmap) was much better on his F3, but he has worked with the newer models too and feels the same way. After he understands injector opening he'll see that it is a much better solution than the workaround he applied before (especially if he ever gets to change the injectors :-).
problem candidates:
- wrong injector opening values - this is the primary problem. injuocfuel=20 or above is needed, and tune primarily injector opening values below 40kPa instead of VE.
- kpaofs=00 maybe this was needed because of a "(" typo in MegaTune config (see MegaTune). This might make lower than actual values at low MAP.
Ruled out:
- setup (nonlinear, wavy transfer-function MAP sensor output?) Map is stable at all load points, it has been a prime suspect but ruled out
- lack of multiplication with MAP. If config13 bit2 would be set to 1 (which is not sane for speed-density, but often used for AlphaN) than multiplication with MAP is not used, so the VE table will be very steep and harder to tune. config13=02 so MAP multiplication is used, and this is OK (VE table confirms this).
- hybrid_rpm_a=00 and hybrid_rpm_m=00 so speed-density is used, but even for speed-density, the config13 bit2=1 would disable the MAP multiplication
Interpolation
It is absolutely impossible that interpolation would be the problem, unless at the variable intake or variable valve actuation point (which is not appropriate here).
Is there no interpolation in often referred Haltech so that many kpa bins are needed?
Interpolation between table lines/columns can only be disabled in vems when it makes sense: for variable intake or variable valve actuation point: with 2 equal neighboring bins (making a solid switchover point).
- The first derivative of the function to be mapped can be any high: it just does not require more bins with interpolation
- High 2nd derivative value requires bins, but 2nd derivative is absolutely limited by nature
In the RPM direction, for some engines there can be 2..3 small waves in the VE where 9..10 bins can come handy (though 8 configurable bins with interpolation and incredible EGO correction is significantly better than any number of bins without incredible EGO).
In the MAP direction those waves (high 2nd derivative) are simply not possible. T
- lambda unbelievably smooth (even 4x4 bins would be enough)
- VE unbelievably smooth below 60kPa if injector opening config is not neglected: in this case VE is the problem, because of bad injector opening config
- VE unbelievably smooth smooth above MAP=60 kPa (upto 512kPa or whatever), and
- ideal ignadv just cannot ever have more than 7 significant breaks in the MAP direction. ( confirmed: Ignition is fine)
More kPa bins for VE table can compensate for bad injector-opening config
Q: How long until larger tables are available in both firmware and MegaTune?
12x12 or 16x16.
16x16 will surely be supported in tuningsoftware, mainly for marketing (avoid blaming table-size when the problem is elsewhere, like injector-opening or not understanding other things; also to make the tuning process somewhat easier by absolutely eliminating the need for moving around bins, ever).Great!
There are things we should sort out before we give up and wait for larger tables, we would like to know what those would be.
- Configure MegaTune.
- firmware just needs to be recompiled with VE_SIZE_RPM=16 and VE_SIZE_KPA=16 and some testing. I will try this in a week when I'm back from vacation, on 12/5-2005, if someone else hasn't tried it before me!
On a happier note, ignition and wide-band lambda is working out great on this motor!
Status log
- Mar 5th, 2005
Another BWM owner needs a Genboard install. Visit MembersPage/MattiasSandgren/PeppesBmw to check it out. We have some trigger issues to work out.
- Feb 12th, 2005
Added a page about a BMW 728 that I'm helping to convert to VEMS/Genboard with help from all of you and MembersPage/EmilLarsson in particular. Visit MembersPage/MattiasSandgren/BmwSevenTwentyEight to check out the wiring issues.
About me
I am the main site administrator of [Sävar Turbo Site] and I have in recent years become more and more interested in combustion engines and cars in general. I have a computer geek background but mechanics always intrigued me - not only electronic gadgets.
I have several cars but my two main projects are :
- BMW E30 325iX (1989) MembersPage/MattiasSandgren/BmwEthirtyIx
- Volvo 142DL (1972) MembersPage/MattiasSandgren/VolvoHundredFortyTwo
The Volvo is long term, my focus is currently on the BMW. I plan to replace the Motronic ECU of the 325iX completely.