General advice: Make sure the engine in general is in tip top condition.

There are two parameters you NEED to know: The real boost pressure and the fuel/air ratio

You *ABSOLUTELY* must get an aftermarket boost gauge before fiddling with boost levels. The factory gauge does not really measure boost, and is not accurate. This gauge can be plumbed into the line going to the computer in the passenger footwell

Controlling the amount of boost in a turbocharged engine is absolutely critical to prevent detonation

The key to not damaging your engine when you increase boost is maintaining a proper fuel/air mixture. In particular, you want the mixture to stay fairly rich to keep combustion temperatures from getting too high and causing detonation. However, what you are doing by increasing boost levels is introducing more air into the engine, therefore creating the possibility of a lean mixture. The fuel system can provide additional fuel and maintain acceptable f/a ratios to a point, but it can run out of capacity. However, it is still a good idea to check to make sure that your car is maintaining adequate f/a ratios if you increase boost at all.

Detonation:

Detonation is the spontaneous combustion of the air/fuel mixture ahead of the flame front (combustion by explosion rather than controlled burning). It occurs after the combustion process has started and is usually located in the last area to burn. As the flame front advances across the chamber, the pressure (and temperature) in the remaining unburned mixture rises. If the autoignition temperature is exceeded, this remaining mixture explodes. The audible ping (metallic clanking noise) is the explosion's shockwave.

Detonation is sometimes called preignition and is typically used to mean knocking or premature combustion. It occurs when the air/fuel mixture inside the cylinder ignites before it is supposed to or else burns faster than it's supposed to. As the piston moves upwards, it compresses the air/fuel mixture by 8, 9, or even 11 times (hence your compression ratio). Since turbocharged engines are already compressing the air mixture before it enters the cylinders, they cannot support high compression ratios--their *effective* compression is similar to or better than that of normally aspirated engines, but the compression ratio number you'll see on a chart is usually less than that of a non-turbo engine.

Here then are two causes for knocking: spontaneous combustion and quick burn.

The first: As the air/fuel mix compresses, it gets hot-- really hot. And the more you compress it, the hotter it gets. If it gets hot enough, it will spontaneously combust (that is, without the need of a spark-- this is how diesel engines work, only it's called compression ignition). The tricky balancing act with spontaneous combustion is having just the right amount of fuel in the mixture to help absorb the heat (by vaporizing the fuel and heating the mix) without putting in too little or too much. Too little can result in knocking. Too much is wasteful, pollutes, and is less efficient (meaning less power). This cause for knocking doesn't usually affect people here on the list (at least not those with non-turbo engines) since the compression ratio isn't easily changed (you'd have to mill the head, change pistons, etc). The design compression ratio is usually based on the type of fuel being used. Alcohol-burning engines see something like 14:1 ratios; diesels can see 22:1; some modern street Ferraris see 11.4:1; Turbo 944s 8.0:1, Non-turbo 944s see 9.5:1 to 10.5:1 (depending on year)

As for burning too quickly, this is where octane ratings come into play. At higher engine RPMs, the ignition control is set to actually fire the spark plug *before* the cylinder reaches top dead center (TDC). The reason is that it takes time for the combustion process to occur. If you wait until the piston is at TDC at 4,000 RPM, the burning mixture will be already on its way out the exhaust port before it's through burning (loss of power). But if the fuel burns too quickly and the ignition system is firing the plugs for higher octane fuel, then it reaches its too high a pressure (the exploding fuel is expanding) before the piston gets to TDC. That rattling noise you hear is all your internal engine parts literally shaking-- that high-velocity piston is hitting an expanding wall of hot gas. Think of it as driving your car into a brick wall. That's why knocking can destroy an engine.Most modern cars use some combination of these to prevent knocking, which can actually be happening before you audibly hear it. The key is that those vehicles have knock sensors.

Detonation is very destructive. No metals in existence, forged pistons, special head gaskets can withstand sustained detonation. If you hear detonation your best measure is to temporarily back off, before looking for its cause.

Detonation can be caused by the following:

1. Poor Fuel. A fuels octane rating is a measure of its resistance to spontaneous combustion. The greater the octane, the greater the resistance. For extreme boost levels it is recommended you use either an octane booster or some type of Aviation fuel

2. Ignition Timing. Improper ignition timing can cause detonation. A lot of the late model turbo cars have knock sensors built in that automatically retard the ignition when knock is sensed, to prevent damage.

Check for a defective knock sensor. Many late model engines have a "knock sensor" on the engine that responds to the frequency vibrations characteristically produced by detonation (typically 6-8kHz). The knock sensor produces a voltage signal that signals the computer to momentarily retard ignition timing until the detonation stops.
A knock sensor can usually be tested by rapping a spanner on the manifold near the sensor (never hit the sensor itself!) and watching for the timing change while the engine is idling. If the timing fails to retard, the sensor may be defective -- or the problem may be within the electronic spark timing control circuitry of the computer itself. To determine the cause, you'll have to refer to the appropriate diagnostic chart in a service manual and follow the step-by-step test procedures to isolate the cause.
Sometimes a knock sensor will react to sounds other than those produced by detonation. A noisy mechanical fuel pump, a bad water pump or alternator bearing, or a loose rod bearing can all produce vibrations that can trick a knock sensor into retarding timing.

3. Lean air/fuel ratio. Lean running will promote detonation. Lean mixtures increases heat which increases the chances of detonation.

4. Exhaust Gas back pressure. A build up of back pressure can cause heat to build up, which can aggravate detonation.

5. Intercooler. Anything that lowers the intercoolers efficiency will increase the chances of detonation. eg paper covering intercooler.

6. Ambient Heat. High performance turbo systems often run near the detonation threshold. Hot days can push the car over this threshold.

7. "Read" your spark plugs. The wrong heat range plug can cause detonation as well as preignition. If the insulators around the electrodes on your plugs appear yellowish or blistered, they may be too hot for the application. Try the next heat range colder spark plug. Copper core spark plugs generally have a broader heat range than ordinary plugs, which lessens the danger of detonation.

The ideal AFR is 12.6:1, this condition is RICH NOT LEAN. Maximum power is at lambda 0.86 with an AFM of 12.6:1. This is taken from the "bible" on Bosch injection systems by Charles O.Probst, SAE in his book "Bosch Fuel Injection & Engine Management". Chapter 7 is on Tuning for Performance and Economy. On page 7-7 he specifically states, "Peak power is achieved with a slightly richer air fuel mixture (when Lambda is approximately 0.9). Minimum specific fuel consumption is achieved with a slightly leaner mixture (when Lambda is approximately 1.05). So, max power is on the rich side, not lean and the "ideal" value is 12.6. Note that the curve is relatively flat in this area so a small variation has little effect on the curve value.

There are two ways to monitor fuel/air (f/a) ratios: First, the function of the oxygen sensor is to monitor f/a ratios for the ECU. It produces voltage in proportion to the a/f ratio. High readings (~.9v) indicate a mixture which is conducive to long engine life. The reading should go to ~.9 and stay there as rpms increase. When you reach the limits of the fuel system, the voltage will start decreasing as rpms increase. Lower voltage = leaner a/f ratio = possible detonation, so back off the boost if this occurs. The factory thoughtfully provided a wire which can be used to monitor O2 sensor voltage in the passenger's footwell, in the DME wiring harness (orange pin 24), so you can connect a voltmeter to it or an A/F meter e.g. Lumenition, and get a display.

The stoichiometric (STOICH) air/fuel ratio is the chemically correct ratio, theoretically all of the oxygen and all of the fuel are consumed. The mixture is neither rich nor lean. However, due to the fact that combustion is never perfect in the real world, there will always be a small amount of oxygen left in the exhaust. This small amount that is left is what the oxygen sensor measures. The smaller the amount of oxygen that is left in the exhaust, the richer the A/F ratio is, and the higher the oxygen sensor voltage is. The on-board computer monitors the voltage from the oxygen sensor. If the computer sees an oxygen sensor voltage greater than .450V, it immediately starts to reduce the amount of fuel that is metered into the engine by reducing the on time to the fuel injectors. When this happens, the A/F ratio starts to go in the lean direction, and the oxygen sensor voltage starts to go down. When the voltage drops below .450V, the computer immediately starts to increase the fuel metered to the engine by increasing the on time to the fuel injectors to produce a richer A/F ratio. This occurs until the oxygen sensor voltage goes above .450V. This repeating cycle happens very fast (many times per second). The computer is said to be in closed loop. It is constantly monitoring the oxygen sensor voltage and adjusting the on time of the fuel injectors to maintain a stoichiometric A/F ratio. This A/F ratio produces the lowest harmful exhaust emissions, and allows the catalytic converter to operate at peak efficiency, therefore reducing the exhaust emissions further.

Since the oxygen sensor output is non-liner and very sensitive at the stoichiometric A/F ratio it will cause the A/F meter LED's to bounce back and forth rapidly. A very small change in A/F ratio causes a large change in oxygen sensor voltage as can be seen on the graph. This causes the A/F ratio meter LED's to rapidly cycle back and forth, and is normal operation when the computer is in closed loop and trying to maintain a stoichiometric A/F ratio.

The oxygen sensor is very accurate at indicating a stoichiometric A/F ratio. It is also very accurate at indicating an A/F ratio that is richer or leaner than stoichiometric. However it can not indicate what exactly the A/F ratio is in the rich and lean areas due to the fact that the oxygen sensor output changes with the oxygen sensor temperature and wear. As the sensor temperature increases, the voltage output will decrease for a given A/F ratio in the rich area, and increase in the lean area as shown on the graph.

During wide open throttle (throttle opening greater than 80% as indicated by the throttle position sensor) the A/F ratio will be forced rich by the computer for maximum power. During this time the oxygen sensor outputs a voltage that corresponds to a rich A/F ratio. But the computer ignores the oxygen sensor signal because it is not accurate for indicating exactly what the A/F ratio is in this range. The computer is now in open loop, and relies on factory programmed maps to calculate what the on time of the fuel injectors should be to provide a rich A/F ratio for maximum power. The A/F ratio meter should indicate rich during this time.

During hard deceleration the computer will command an extremely lean mixture for lowest exhaust emissions. This may cause the A/F ratio meter not to indicate anything. The A/F ratio is so lean that it is outside the range that the meter will indicate.

Injectors

1. size, of course

2. the pulse duration, obviously

3. but also, the delta pressure across the injector.

The standard Fuel Pressure Regulator is rated at 2.5 bar (36Psi)

Injectors are 370 cc

Flow rates

The industry standard for measuring flow is 3 bar (43.5) psi and 80% duty cycle

Standard injectors flow between 260cc - 265 cc at 36 psi at 80% duty cycle. If you were to hold the stock injectors wide open, 100% duty cycle, they would still flow nowhere near 370cc at 36 psi.

lb = cc / 10.5

A 370cc injector would equal a 35 lb. injector (rounding off).

Diagnosis

I do not know what is 944 S2 turbo, but if it is anything like US. 951, there a few test for you to try if you are suspecting boost controlling circuit (cycling valve, wastegate, etc...). Get a good manometer to be able to get accurate boost pressure readings.

Check Compressor Bypass Valve (CBV). It is a little black can with three connections (two larger, and one smaller). It sits on top of a Rubber Sleeve(RS) that goes from air flow meter to the turbo fresh air intake. The other larger fitting of CBV connects with the metal pipe from intercooler to throttlebody.

1 Take out CBV, hold it so that smallest (vacuum) fitting is pointing up, give a horizontal fitting a good blow job, it should not leak.

2 Try the above step, but connect a source of vacuum (18 in Hg) to the smallest tube, you should be able to freely blow into it.

3 Try sucking on the small tube, it should hold vacuum

Conclussions If any of the three cases fail, replace CBV. The failure in step 1 could be the cause of low boost.

1 Pinch the rubber part of the hose that leads from the banjo bolt in the Intercooler Pipe(IP). WARNING Go for a ride (very carefully do not overboost). WARNING Do not forget to remove this clamp after the test.

Conclusions If a car feels like it just got a steroid injection (you should get full boost and then some) your problem is probably the Cycling Valve(CV).

Unless you have superhuman upper extremities, you will need to remove intake manifold to get to it. It is a three connection electrically operated air valve. At any given time the air from the IP connected side is bled to the WasteGate(WG) control or into the RS to the intake of the turbo. As a safety feature this CV is normally (unpowered) allowing the wastegate to see the full pressure, thus forcing it open. Your KLR controls boost pressure by electrically activating the CV. Once the CV is out it can be tested.

1 Get under the cold car in the morning, start it, or vice versa, and touch the small exhaust pipe that leaves WG.

Conclussion if hot, it is probably your WG.

The above tests should narrow your problem down, if not fix it. You could also have an air leak under pressure somewhere. I do not know how to find those, may be compressed air and soap water can help.

In a few recent threads, there was a talk that not all the cars see the full boost unless under WOT. With that in mind if your throttle cable is not adjusted properly, you may never see WOT, and full power.