• T3/T4 50trim Turbocharger (nocash kit)
• Supra Twin Turbo 540cc/min fuel injectors
• Apexi S-AFC
• Aeromotive adjustable fuel pressure regulator
• Two stage water injection
• Greddy Intercooler
• K&N FIPK Intake
• Apexi AVC-R boost controller
• Ross forged pistons

Session 1

Fuel Pressure – 43psi

AFC Values:

Boost: 1.10kg/cm^2 (15.6psi)

Peak HP:185.6 @ 6460

Peak Torque: 197.3 @ 4807

Wideband datalog

Notes

Running very rich as expected

Session 2 – repeat of session 1.

Peak HP: 179.1@6476

Peak Torque: 186.8@4680

Session 3 – Same run except increased boost to 1.20kg/cm^2 (17.0psi)

Peak HP:229.7@6116

Peak Torque: 206,1@5501

Notes

Still running rich – AFR went from 9.35 to 9.79 near peak power.

Comparing Session 1 and Session 3 – 17.0psi in red.

Peak power increased 43HP from 1.4psi of additional boost. Strange kink at 5000RPM on the higher boost run.

Session 4 – Same as session 3 except fuel pressure lowered one turn.

Peak HP: 245.1@6097

Peak Torque: 211.8@5620

Notes – AFR increased from 9.75 to 10.25 around peak power. Peak HP increased 15.6HP.

Comparing session 3 and 4 (session 4 in red):

Session 5,6 – Same as above except increased boost to 1.30kg/cm^2 (18.5psi)

Around peak power, AFR leaned from 10.25 to 10.45. Judging from the AFR, one would think that you would want to go more lean. This would have been my intention, but out on the road, I looked at the wrong part of the graph! So, I increased fuel pressure a half turn for the next run, which (naturally) made less horsepower.

Session 7 - Increased fuel pressure 1/2 turn.

Peak HP: 251.4

Peak Torque: 224.9

Notes – wideband certainly shows slightly more rich. Power dropped 6HP from 257, to 251.

Session 8 - increased boost to 1.35kg/cm^2 (19.2psi).

Peak HP: 266.0

Peak Torque: 234.1

AFR was about the same.

Notes -Increasing boost from18.5psi to 19.2psi increased peak HP by 15. Comparing runs 7 to 8 (19.2psi in red):

Session 9 – AFC Adjustments

Comaring runs 8 to 9 (run 9 in red):

Notes – Power dropped about 8HP. Not sure why – AFR was about the same.

Session 10 – More AFC adjustments going to put in more fuel (put some timing back in).

During this run the ECU detected detonation. This is based on the fact that the TVIS LED changed state whenever the throttle was in any other position besides closed. That is as soon as you touch the throttle the LED changed. After about 1 minute of driving the ECU returned to normal operation and also permitted full boost (TVSV LED). I do not know why there would be any detonation from a less agressive AFC setting. As you can see, the AFR was also rich.

Since the ECU returned to normal operation so quickly, I continuted tuning.

Session 11 – Same as above except changed 5500RPM to -4.

Notes – Power has returned. This run is very similar to run 8.

Session 12 – After a couple tries at other AFC settings, I ended up with this. This is after several other tries.

Comparing first graph (15.6psi) to last (19.2psi) – 3.6more psi of boost:

I drove the car with the above settings, and all seemed well – no detonation, and the engine seemed to be running very well. My only guess as to the detected detonation during session 10, is because of too much timing. I believe that the stock ECU is somewhat agressive with timing, and that is why such rich AFRs are required. Thsi is where an aftermarket engine management system would really help.

• Apex-i S-AFC (Super Air Fuel Controller).
• LoveHorsepower Fuel Pressure Regulator Kit (Aeromotive Fuel Pressure Regulator).
• Innovate Motorsports LC-1 Wideband Lambda Cable (Wideband Oxygen Sensor Kit).
• LoveHorsepower Stage Two Water Injection.
• 550cc/min Fuel Injectors from the twin turbo Toyota Supra.
• GTech meter – Pro Competition Version.
• TVIS (Toyota Variable Induction System) and TVSV (Toyota Vacuum Switching Valve) LEDs.

After the above components were installed, the tuning process could begin! First the S-AFC high throttle map was zeroed. The low throttle map was set to -7% at most RPM points. The Low throttle point was set to 30%, and the High throttle point was set to 50%. Our turning setup consisted of the GTech meter to datalog RPM, horsepower, and torque. A laptop was connected to the LC-1 kit to datalog air fuel ratios. We set out with a goal of 11.5 for the air fuel ratio when under full load/full boost. The fuel pressure regulator (FPR) was set to a base fuel pressure of 30psi with the engine off (no vacuum). All of our tests were done with water injection turned on using a two stage setup. The first stage (3 gallon per hour nozzle) coming on at ~9psi of boost and the second (5 gallon per hour nozzle) at ~14psi. Distilled water was injected.

All of our tuning runs were done in 2nd gear by zeroing the GTech meter, and starting off in 1st gear from a stop. We would accelerate normally up to about 3000RPM, then shift to 2nd gear and apply full throttle until ~7200RPM. The first run was done at 15psi of boost, and resulted in a very rich mixture, and poor performance. In fact, the engine was running so rich that below about 5500RPM the engine felt like it was bogging down. It was not happy at all!

Once we have the datalog from the LC-1 and the GTech meter, we could graph both of them on a computer. Then using a graphics program, the resulting graphs were resized to have a similar scale. Here is the result of the first run:

The top graph is the air fuel ratio (AFR) from the LC-1. The bottom graph is the HP and torque vs RPM from the GTech meter. As can be seen, power down low is very poor, and the fuel mixture is well below 10.0 during the engine run. Way too rich! You can also see how the power really nose-dives around 3000RPM where too much fuel starts to come in. The power graph is also very jagged, and indeed, it felt that way.

We then lowered the fuel pressure down to about 25psi, and performed another run. The engine was still running very rich, and was still bogging at lower RPM points, but not as badly. The decision was made to increase boost pressure to about 17psi. This resulted in still poor performance down low in the RPM range, but more pull up top, and with an AFR above 10.0 up top. While doing the tuning in 2nd gear, it’s a good idea to do some runs in 3rd to achieve a higher load on the engine at lower RPM points. This allows the turbo to spool up to full boost while the RPMs are still low (~3500RPM – 4000RPM). Detonation is more likely to occur under high loads and low RPMs rather than high loads and high RPMs. This is one of the reasons why F1 engines spin so fast – there simply isn’t enough time for detonation to occur at those engine speeds. When doing one of these tests, the ECU detected detonation. This was noticed by the behavior of the TVIS LED. When the ECU detects detonation, it will activate the TVIS under any throttle position other than closed. It is possible that the ECU detected what I’ll call false detonation. Sometimes by running so rich the ECU will think it has detected detonation. I’m not sure why this occurs, but it apparently does indeed happen. I also cannot say for certain if this is what happened at this point in time.

I then started new, and increased the fuel pressure to about 32psi. I then began to do more tuning runs again, this time only concentrating on the upper RPM ranges and using boost to get the AFR to come up. At about 18psi of boost (guessing from poor notes!), this is the resulting AFR:

As can be seen above time 14.0 seconds, (about 5000RPM) the AFR starts to come up. This is where the engine really starts to pull very well, even though it is still rich and not even reaching 11.0 AFR. More 3rd gear runs were done, and again detonation was detected by the Toyota ECU as indicated by the LEDs. Dang!

Since the engine experienced detonation, and were still running very rich, we restarted our tuning effort with a different strategy. I decided to increase the base fuel pressure to 38psi and work from there. Now I know what you’re thinking – you were already rich, and now you want to add more fuel? Are you crazy? Well – bare with me.

Certainly after upping the fuel pressure to 38psi, things were rich everywhere. I kept increasing boost slowly until I reached 22.6psi. At this point the upper RPM range (above 5500RPM) started to get close to 11.5AFR. It also really pulled hard up top here. Since the lower RPM ranges were still too rich, and taking fuel out with the FPR would have resulted in too lean a mixture up top, my only option left was to remove fuel with the AFC. One other important note, is that I noticed that the injector duty cycle was going to 100% above 5000RPM, pretty much regardless how much boost I ran (above 15psi). Since the ECU behaves this way, really, the only way to tune the top end is by adjusting fuel pressure. If you use the AFC to tune the top end, I would be very worried about too much timing being added in. So – off to tuning the lower RPM ranges with the AFC by taking out fuel/load. I was very cautious in doing this, and only took out 1-3% change per run, and checked to make sure there was no detonation, and that the AFRs were still good (and rich). I still wanted those ranges to be rich since by telling the ECU there is less load with the AFC, it will advance the ignition timing. After numerous runs, I came up with this:

As you can see, in the upper RPM ranges where the AFR comes up to about 11.4 it really starts to pull great. In the lower RPM ranges, it’s still very rich (probably could be made more lean with the AFC to help low end torque out). That said, at least it’s not in the low 9s anymore! Here are the AFC settings for the low and high map at this point:

This is a big improvement over our initial 15psi run with the AFC zeroed out – a gain of 81.6HP and 55.8ft-lb of torque with 7.6psi more boost pressure. How’s the T04E-50 trim turbocharger doing? At 22psi the turbocharger is operating is a good range on the compressor map. With a pressure ratio of 2.49, the turbo is pushing approximately 40.0lb/min at 7200RPM. At 4500RPM, it is pushing 25.0lb/min. Plotting these on the T04E-50 trim compressor map:

So, at 7200RPM the turbocharger is in a nice 75% island and cruising along somewhere between 105,620RPM and 118,700RPM. Now that’s some serious revs, and it sounds sweet too! Looks like we could run more boost – more tuning to come.

• O2 sensor signal for A/F ratio – higher values indicate more rich.

• Boost pressure.

• Air flow from AFM – lower values indicate more airflow.

• Throttle position.

• RPM.

In addition, temperature sensors were added inside the intake pipes before and after
the intercooler to measure the intercooler’s efficiency.

The data can then be logged to a serial port at the rate of approximately 20 samples per second.
RPM is currently only measured once every second.  The BasicX chip was soldered onto a board,
and along with peripheral components, was mounted in a box.  Snap connectors were used for the
prototype version.  Future versions (if any) will use a 15pin connector.  It is also possible to connect this
board up to a display.  The display used was the ‘very smart display module’ from simmetry.com.
This module has since stopped working, and I’m looking into a different display to use.

The following data was taken during a short 25 minute drive.  The data was
then imported into Excel and stripped down to the just the ‘interesting’ points.
Here are some preliminary results.  As can be seen the intercooler efficiency rises
(yellow line) when full throttle is first applied, then slowly drops the longer the
turbocharger is producing boost.  Sorry for such large images, but it is easier to
see the data this way.  Also it is difficult to see the efficiency in this graph since
it is ‘off scale’ with the temperatures.  The outside temperature was 82 degrees F.
The formula for intercooler efficiency is (BI-AI)/(BI-Out) where BI is before
intercooler, and AI is after intercooler temperature.  The following three graphs
are all shown at the same time interval.  Notice also how choppy the RPM graph
is since it is currently only measured once per second.
Plots were taken running 17psi of boost with a T04E-50 compressor trim,
PT upgraded CT-26.  The Greddy IC has replaced the stock unit.

Although not shown, O2, airflow, and boost were also measured.

The temperature sensors were mounted into the intake pipes (click to enlarge) using JB weld.

Pictures of the prototype board:

In the last picture above the snap plugs are for sensor connections and power, the 9 pin serial is
for data logging, and the 4 pin PC power connector is for the display module.  If another is made,
it will use a 15 pin connector to neaten up the design and a PCB board will be used instead
of wire wrapping.  The 2nd picture above is a prior board which doesn’t contain some necessary
additional (8) components for filtering the input signals.  The first picture above (the bottom
of the board) contains those components.

Update (10/23/2003) – speed sensing added, but not tested yet.  The LM1815N adaptive variable
reluctance sensor amplifier from national semiconductor (samples available!) was added to the board
along with necessary caps and resistors to add the capability for measuring speed.  I have not tested it
yet, but have used these chips with much success for reading position sensors for camshaft position.  The
plan is to determine speed by reading the stock speed sensors used for the ABS.

Another plot of the intercooler efficiency over a 20 minute period.  You can get a good idea of
how long it takes the intercooler to recover.

Excel spreadsheet including all data from this run can be downloaded here (zip format).