Lean Spool

[Alex B]

Engine leans out on turbo spooling up when running at 90-100% throttle.

Mapped lower loads with more or less steady constant load ok, however once started to map full throttle loads of 200 kpa and above keep getting lean lambda during turbo spool. The engine is Subaru STI 2.5 with Cosworth heads, 260-11.3 custom cams,Cosworth intake manifold, Greddy TD06 20G turbo, 4-1 exhaust manifold. Usually THxRPM fuel comp map used to solve the issue with few % added in the turbo spool rpm range (2000-4000 rpm), but not this time.

Screen shot below already has increased accell comp and +15% at 100% TH, but doesn't solve the problem, just some parts got richer.

Which comp strategy would be best suited for this case?

Best regards
Alex

[Martin]

I normally see quite a big difference in fueling requirements for NA engines thats got longer duration/more overlap setups when the RPM ROC changes. Especially where the engine gets more effiecient. Where you get the dips in the Fuel table that points to the lower VE areas itll ask more as you ramp into the areas where the VE rises. Altitude also makes a difference in where the VE will start rising.


If you look at your fuelmap, is that where the VE starts rising? Have you tried RPM ROC as a trim in the problem area?

What do you mean by an accel comp? accel enrichment will only be added if the accel channel exceeds the accel threshold. i.e. if its TP itll only be added it dTP/dt exceeds the threshold.

[Scott@FP]

Does this engine have AVCS and/or VVT?

[Martin]

What injectors are you using?

[stevieturbo]

Maybe I'm missing it.


Are you trying to tune using various compensations, or using the main map vs rpm load sites ?

[Alex B]

Martin,

The VE is raising steady over the boost levels. The problem is that under steady boost fuel is ok, but as soon as turbo spools up engine runs lean during this condition.

I did not consider the RPM ROC, however I tend to think that if used may interfere at some other operating points.

Yes I meant accel enrichment under accel comp. I scratched that idea since timeframe is to long for it.

We are using Bosch 1000cc injectors from T1 Race (aka Injector Dynamics).

Scott@FP
Yes, its a Subaru 2.5 STI with intake AVCS, Tomei custom spec 260-11.3 cams, Cosworth CNC heads with big valves, Cosworth intake manifold.

Current situation: We have solved the problem using Secondary load table THxRPM with 10% comp at WOT up to 3600rpm, although its a compromise solution with some overfueling and some underfueling operation, but nonetheless no La 1.0+ operation on spool up.

The final note is that after 1 week of usage by happy customer, engine blew We think that valve seat became loose and started the chain reaction. This is second engine failure with Cosworth heads. First one was very similar. Have a look

[Steve M]

Answer from Cosworth on the big valve heads as posted on Naisoc

A friend of mine, running his 10 sec sti on a cosworth large valve head/cam kit have lost one of his cylinder number 2 intake valve after 8K miles of street use and some runs at the drag strip, here it is the answer from cosworth, wich makes sense, so you guys running larger valves, next time make sure you use the subaru procedure for installment, and turn the intake and ex LH head cam sprockets in the right direction, and turn them with the less angle possible and do not make them move, wich is difficult, but can be done..

so here it is

"The main issue is that it's difficult to keep the cam gears in the proper position when installing the belt and that you don't open intake and exhaust valves at the same time causing contact. As long as you follow the factory service manual instructions exactly, you will be fine. I've attached a PDF copy of the timing belt procedure. On page 5, they explain the importance of the steps and rotation directions. With +1mm valves and higher lift cams, the danger of valve to valve contact during timing belt installation is even greater.
once the valve is slightly bent, the higher spring rate and constant contact of the valve to the valve seat tries to straighten the valve. Plastic deformation sets in, like bending a paperclip back and forth, and then the valve head falls off causing failure. We inspected a number of these valves and the material showed evidence to support our diagnosis.
It's always number #2 that fails, so we're confident this is the problem. Some people argue about piston to valve contact with the AVCS full advanced or retarded, but we've bolted on a cutaway head and manually clocked the avcs cam to each extreme and never came close to piston to valve contact. If the cam timing is phased properly you don't have to worry about the AVCS.

[Scott@FP]

If you are in full intake cam advance in the spoolup areas the high overlap can cause 'short circuiting' where fresh unburned fuel/air mixture is drawn out the exhaust during overlap, this will give a false lean reading but the engine runs well.

[Holmz]

If you are in full intake cam advance in the spoolup areas the high overlap can cause 'short circuiting' where fresh unburned fuel/air mixture is drawn out the exhaust during overlap, this will give a false lean reading but the engine runs well.

Now many questions...

So what needs to be used?
EGT?

How does this affect things for a n/a engine with a lot of overlap?
Will I be back to looking at the plugs and have to not pay too much attention to Lambda?
Last question is on cams with a lot of overlap, what keeps the intake from igniting during the overlap when there is not a lot of advance when one uses a wasted spark>

[MarkMc]

Holmz,
An answer to the "last question"....basically nothing. I have had a customer who saved some money by going wasted spark with a monster overlap cam (Cleveland V8) and at starting sometimes it tried to blow the intake manifold off.