Low end fueling with Vanos and turbo

[ryancyates]

I’m in the process of tuning my built S50B30 turbo.  The fuel map is pretty dialed in except for one area where I’m having an issue.  This is my first time using a LinkECU so I was hoping someone with more experience could give me some insight as to how I should tackle the problem.

Basically, the area of 0 rpm to 3,500 rpm and -10kpa to +10 kpa has two widely different fueling requirements depending on whether I’m lightly leaving from a stop and vanos cam angle is at rest or I’m crossing that range at WOT with 40° of cam angle advancement.

I can set it up so leaving from a stop is fine but then I get a harsh lean condition on spool up at WOT or I can set that area up for 12.5:1 AFR at WOT and I’m seeing 10:1 AFR while gently entering that area of the map and no Vanos engagement.

As far as thoughts I’ve had, I could reduce the vanos angle til after crossing that area of the map but it does slightly affect spool characteristics.

I haven’t used LinkECU’s closed loop control so maybe that’s a solution?

Then I was wondering if this is where I could do something with 4D fueling and cam angle?

[Hodgdon Extreme]

Are you on traditional fueling or modeled?

Closed loop works great and would probably deal with the issue, but it's sort of band-aiding improper characterization of your fuel/VE table... It's really better to simply have the VE of your engine properly characterized.

Obviously, different valve event timing changes VE; hence your varying lambda.

A 4D/overlay target table would be a good solution to introduce cam angle as a factor in VE.

[ryancyates]

52 minutes ago, Hodgdon Extreme said:

Are you on traditional fueling or modeled?

Closed loop works great and would probably deal with the issue, but it's sort of band-aiding improper characterization of your fuel/VE table... It's really better to simply have the VE of your engine properly characterized.

Obviously, different valve event timing changes VE; hence your varying lambda.

A 4D/overlay target table would be a good solution to introduce cam angle as a factor in VE.

I’m currently using traditional fueling.  It’s made it a lot easier to familiarize myself with the software and Link method of operation.  I had originally tried modeled from the start and I was fighting too many variables and user end errors to get anywhere.

So in this instance my first thought would to disable vanos and tune that area then enable vanos and tune that portion via 4D map?

With cam angle as load reference x rpm?

[Adamw]

Are we talking about a "genuine"  S50B30 with ITB's, or the S50B30US with a single throttle?  I assume the ITB one since I think the US only had the M50TU style on/off vanos.  

If ITB, can you attach a copy of your tune so I can see how the fuel is set up?  

[ryancyates]

Yes its a car that I've imported from Japan in 2021.

I've attached the tune file.  While the fuel map is still "blocked out" as I call it, I like to work with map regions and smooth afterwards, it gives me great results at 1bar of boost.

You can see from the Vanos map that I have it set up closely to the original file from the stock chip with some changes to help with spooling.

The Vanos map is referenced TPS x RPM with 42 degrees of advance coming in around 3,000 RPM and quickly tapering at 5,000.

So in 4th gear at WOT and 3,000 RPM the Vanos is at 42 degrees advance car is already making close to 3 PSI and the AFR is a spot on 12.5:1.

If I'm leaving from a stop light and cross into that 3,000 RPM - 0KPA area under light throttle, there is maybe 8 degrees of cam advance and my AFR will dip into the low 10's.  You can see this in the hard shelf between 0 KPA onward.

If I set the fuel table up to accommodate for the light throttle portion in that area I see AFR in the 19's under WOT and full advance.

S50 MGP.pclx

[Adamw]

I have never seen an ITB engine give consistent lambda when using MAP or MGP as the fuel table load reference, - and that is without the complication of VVT added.  MAP just doesnt seem to be a good indicator of airflow when it is not coming from a large single plenum.  On GTR's etc where I have experimented with MAP in the past, under some driving conditions like the start of a hill you might open the throttle from say 20% to 40% to maintain speed and you will barely see a change in MAP but it will go dead lean as the throttle slowly opens so obviously air flow has changed but MAP just doesnt always reflect that.   

So if you haven't already tried, I would start again using the suggested alpha-N + MAP with open loop target correction.  Use fine TP breakpoint increments at small throttle openings and you can spread them out at larger openings.  I usually start with something like 0,2,5,8,11,15,20,30,50,70,100 and only add more if there is some area between 2 rows that doesnt track well.  Tune the fuel table on the lowest boost setting, then try it on the highest to see how well lambda follows target.  In some cases if the turbine is restrictive at high boost or there is some other factor causing a significant reduction of efficiency at high boost then it will drift rich, in those cases you have to enable a 4D table with MAP or MGP on its axis and just pull a little fuel out in the areas it drifts rich. 

I haven't done many ITB engines with VVT though - an RB26 with the VCam and an S54, but I dont remember needing to do anything special with 4D tables etc to keep mixture in an acceptable ballpark with vvt movement. 

[ryancyates]

That's interesting and it makes sense.  This is also the first turbo ITB car I've tuned.  I have a built N/A Euro M3 that I've tuned with the stock DME using an extended MAF fault map as an alpha-n table.  I assumed the reason why I was having difficulty here is because the Motronic ECUs use different fuel maps for idle, cruise and WOT, where WOT is just a single column referenced by RPM and injector constant, where I'm trying to do all of this from a single table while adding boost.

I figured I would be ok since I'm pulling the map reference off a 1/8" nipple on the brake booster rail, but I guess the whole point of ITB is quicker response and cylinder filling so I will go out this evening and see what I can do.  From the looks of it the partial throttle section of the map closely lines up with the current MGP map as is, so that's great.

Just to make sure I'm understanding this correctly, the AFR map is like a background correction adding injector time based on MGP?  I guess this would be to create a more linear map?  Like how my current fueling map boost range fuel vales are only 71-84 for 0-15 psi?

Honestly, I didn't even realize I had it turned on in my current map til I just checked.

[ryancyates]

For future reference to anyone in need of this information, the suggested tuning method has worked very well.

I still get minor AFR variances at 3,000-3,500 depending on cam angle but they are in the area of 13.5-14.5 AFR and no longer 10 AFR or lower and not worth nitpicking a quick transient with no detriment to the engine.

[Adamw]

18 hours ago, ryancyates said:

Just to make sure I'm understanding this correctly, the AFR map is like a background correction adding injector time based on MGP?  I guess this would be to create a more linear map?  Like how my current fueling map boost range fuel vales are only 71-84 for 0-15 psi?

Yes, this factors target lambda into the fuel equation.  When your equation load source is set to MAP, and the open loop target table is turned off, if you doubled the MAP, this would indicate the inhaled air mass has doubled so the ecu would give double the injector pulse width to keep the fuel mixture the same.  But in general when you double your MAP you dont want just the same fuel mixture as before - you would usually want a richer mixture due to the combustion pressure and temp increase.  If open loop target correction is enabled and say we were targeting 1.00lambda at 50kpa and 0.9lambda at 100Kpa, then you would get not only the doubling of the inj PW due to the doubling of MAP, but you would get an extra 10% due to the 10% richer target as well.

So, 3 main advantages with open loop correction:

1. The fuel table is flatter as you have noticed, it usually also means you can get away with less fuel table cells.  The fuel table is more like a VE table reflecting air flow as it doesnt have mixture variation baked into it.
2. After tuning you can quickly and freely try different mixtures at any operating condition without having to retune or adjust the fuel table.
3. It takes care of the required richer mixture as MAP/boost increases when using alpha-N as described above.  You can vary boost etc after tuning without messing anything up.