Im on the way for fix a problem. Today i make pressure test and found leakage around TB shaft, common problem(holds only 2-3psi). And also found, that cleaned black build up around butterfly was some factory sealant.
So some of you was right about "vacuum" leak, but i think main problem will by high ecu sensitivity on butterfly sealing VS idle speed regulation.

Ordered shaft seals, then i will dissasembly TB and try to rebuild it completely.

 

Throttle shaft seals installed. Throttle plate resealed by some hard wax and teflon lubricant.
Idle screw calibrated again. Now ISCV steps is little lower than before. TPS sensor set to 12,5%.

Unstability is still here.

I recalibred first three-four collumns in MAF calibration and first two in MAF smoothing, to have O2 trim more constant(was bit rich around 1500rpm on neutral). No influence on RPM drop when i put higher or lower values here. Only 02 trim changing.

Tried add some values in ISCV demand RPM adder. Have no influence on ISCV steps value on datalog, when RPM drop happend. So no influence on my problem.

Tried to add some timing on High octane map under 1000rpm. Tried lean last 500mm row on fuel map from 13,3 to 14.7.

From datalog, you can notiece, there is large O2 trim change(5-15% negative), immediately after revving... take some time to get back around zero, where is tuned for.

Any idea?

 

Going deep in this f.king problem, to many threads on this forum, no solution found.
Tried another two tricks.

I put cca 5cm of rubber hose - diameter same as AFM, inside large diameter of hardpipe intake. To reduce diameter jump - lower potentional turbulences.
In same time, i smooted pipe inside around BOV outlet. Because i read some problems in this area with KN typhoon intakes.

Again, no change.

Made a log, until engine stall. I notieced, there is decel fuel cut(0ms injection) and around 1500rpm injection going back, but going much higher than is needed. Just before stall(cca 1000rpm) there is spike to 7,5ms and this probably flood engine. On steady idle is 2.3ms.
When this spike is up to 4ms, than engine stay running, bud rpm dip to cca500rpm.

Problem is, that airflow Hz is still the same in this situation, no spikes or other violence(30-40hz). So why ECU randomly pump so much fuel?

Datalog enclosed, stall is on the end.

 

I found its definately fuel problem. Ics steps doesnt go lower than the same value as for idle.
Rised columns at 500rpm/30/40/50/60 load to max lean 19.4 and its little bit better, but little bit. AFR still jumps to 10.

I cannot lean it more, as i wrote airflow HZ are same or little higher at this rpm dip, so if i change MAF scalling or smoothing, its affect AFR at idle.

Slowly thinking about Speed density instead MAF...

 

This problem only started after cleaning the MAF and throttle body?

 

Im not sure. As i bought car with death clutch, run just few miles and then going to repair it and clean everything.

I will try to play with RPM idle adders, to try slow rpm drop before idle(1500rpm) as precaution, not as i tried rise values under idle(under 1000). But it just my thought now.

 

Im not alone boys, one of foreign videos from youtube of this IDLE DROP problem...

 

Adding RPM adder at 1000-1500 doesnt have influence.

I found another potential source of problem. As i wrote, im running without blow off valve. There is massive surge of air going back throught MAF, which is probably my fault. Installed some old open piston blow off valve, and now almost no idle drop stationary, when drive there is drop, due slow bow closing(air leak). So im going to source OEM blow off valve.