.I am still running the stock computer which I tweaked in two stages which are described below. Note that stage 2 was more effective and flexible and is what I am currently running. Note that the B6T only has ECU controlled fuel delivery but ignition is centrifugal with vacuum advance and boost pressure retard. There is also electronic knock control and retard. This actually amounts to a coarse 3D type of map with load vs. RPM and while obviously not as accurate as full ECU control, is still quite effective at allowing high boost.
The standard injectors on the B6 turbo engine are approx. 300cc/min which means that they are capable of flowing significantly more than 100kW of flywheel power. According to 21st Century Performance a good rule of thumb is:
kW x 7 = cc/min x No. Injectors
=> kW = (300 x 4) / 7 = 171kW
Consequently with an increase in fuel pressure and some other tweaks maybe 200kW at the flywheel is possible without resorting to larger injectors.
The percentage fuel flow increase with regard to fuel pressure is not linear but rather the square root of the new pressure divided by the old pressure. For a new pressure of 55 P.S.I and the stock pressure as measured on my car of 34 this calculates to:
SQR(55/34) * 100 = 127% of the original flow
Originally I set the fuel pressure very high (55-60 P.S.I) and adjusted the spring tension of the air flow meter to compensate at lower loads. Doing this requires the removal of the black plastic cap on the top of the air flow meter. You need to cut the silastic seal and pry the cap off. You can then clean off the silicon with a razor blade and re-seal it with some foam rubber weather strip or something similar. Tightening the spring has the effect of the computer thinking there is less airflow than it has been calibrated for and thus the lighter loads are not too rich even with increased fuel pressure. At higher boost we reach the extreme of what the ECU will respond to and thus the extra fuel pressure is a benefit and combats high boost lean out.
This was reasonably effective but still meant that mixtures were possibly on the lean side of safe at high boost and was also putting a fair bit of strain on the new Goss fuel pump (which even though under warranty, it would be nice if it lasted more than two years!)
I later discovered while scouring the net for info. that there is a hidden air bypass screw which is buried under a steel cap on the top right of the air flow meter. I removed the cap by drilling it, putting in a large self tapper and pulling it out with pliers and then replaced it with a plastic cap. Opening the air bypass (which is a simple screw) has the same effect as tightening the spring tension of the vane in the air flow meter and has the benefit of not increasing the flow restriction. It is also a lot easier to adjust and allows a finer grain of adjustment too.
Stage 2 enrichment modifications involved making a small circuit consisting of a pot, boost switch and relay that puts a variable resistance in series with the water thermo sensor when boost hits about 11PSI which has the effect of enriching mixtures from 11 PSI onwards in a way that is fine tunable with the pot. Note that the water thermo sensor is a variable resistor which increases in resistance when it is cold and decreases when hot.
It is handy that the B6 turbo has an external boost cut solenoid ... disconnected of course :-) This switch is normally open but closes at about 11PSI and serves as a perfectly good boost switch, sitting unused in the engine bay, that activates just at the point where the mixtures started to lean out.
A simple circuit diagram will be coming soon but basically, a relay is placed in series with the water thermo switch and a pot is placed in parallel with the relay (but in series with the thermo sensor). When the relay is closed (i.e. less than 11PSI boost) the pot is effectively switched out of the circuit since a resistor in parallel with a closed circuit gives almost no additional resistance. When the relay is opened (> 11PSI boost) it means that the pot is now in series with water thermo switch and has the effect of increasing the resistance by whatever it is set at (about 5k worked in my case) which makes the ECU enrichen the mixtures. Fortunately the ECU enriches mixtures quite significantly, even at high boost ... I suspect that enrichment is a simple percentage adjustment factor which is constant across the entire load/rev range (any chip experts out there know for sure?) This enrichment allowed me to lower the fuel pressure to 45PSI (a sensible figure that many EFI cars use in standard form).
Tuning the mixtures was done with a multimeter reading the oxygen sensor and I also have a Jaycar mixture meter in the car so I can keep an eye on things all the time. I think this is more accurate than people think because the response curve for a zirconium oxygen sensor is said to be very similar across cars so I can get a rough idea of the A/F ratio ... at least within .1 lambda I would think. Below is a diagram showing typical EGO sensor response:
.