Simulation and test (first simulation) of  a APC-box

Writer: Anders Olsson, Translated by: Widde

APC-box number. 7524119 (to 900 T8-86, no cat, 155hp)

The test were executed with a complete cable trunk to the APC system.
RPM was generated by a function generator which was connected to the ignition pulse amplifier. 
The pressure sensor was simulated by a 500 Ohm potentiometer.
Knock were simulated through shaking of the knock sensor with a bolt and nut loosely fastened in the knock sensor mounting hole.
An oscilloscope was connected over the valve to read the pulse width to the valve.

Simulation 1
Knock sensor not connected, pressure sensor set at atmospheric pressure

The valve has current 100% of the time below aprox 2330 RPM.
-Maximum boost pressure ordered, due to no knock detected, APC doesn't check the knock sensors existence and the boost pressure is low.

The valve is disconnected 100% of the time over aprox 2330 RPM.
- Base boost ordered due to nonexistent knock sensor.

Simulation 2
Pressure sensor set at atmospheric pressure (12 Ohm). RPM below aprox 2330 RPM.
If knock is simulated during several seconds the valve will be disconnected.
-Base boost
It takes 5 seconds before the APC starts to rise the boost pressure.
After 30 seconds from the knock stopped, the boost pressure is back to normal.

Simulation 3
Knock sensor not connected
Pressure sensor not connected
Over aprox 2300 RPM the valve is disconnected.
- APC shut off due to nonexistent knock sensor

Simulation 4

Below 2300 RPM is the boost lowering is depended by how long the knock occurred.
Above 2300 RPM is the boost pressure lowered in only one step during knock (?)
Comment: Real strange, more tests needed

Simulation 5
Below 2330 RPM is the valve initialised when the pressure sensor gives a value above 37 Ohm.

At 500 Ohm on the pressure sensor the valve is initialised 21 % of the time.

Simulation 6
Over 2330 RPM and at atmospheric pressure (12 Ohm) will the valve be initialised 65 % of the time. The valve will always be initialised over the 2330 RPMs unless the knock sensor is disconnected.

Simulation 7
At RPM just above 2330 and at 500 Ohm (or disconnected) to the pressure sensor will the valve be initialised 21 % of the time.

Results/conclusions:
Below 2330 RPM can the APC regulate the boost pressure through initialising of the valve 21-100 % of the time.

Above 2330 RPM can the APC regulate the boost pressure through initialising of the valve 21-65 % of the time.

When knock occurs, the APC can always (?) disconnect the valve.

 

Simulation with APC 2003-May-11 (second simulation)
APC number: 7524119 (900 T8-86 155hp/240Nm)

Writer: Anders Olsson, Translated by: Widde

Simulation was performed with a complete APC systems cable trunk.
A power supply was set to 12 volts and was used as a power source to the APC system.
The pressure sensor was exchanged with a 500Ohm potentiometer.
The knock sensor was attached. To simulate knock, the knock sensor mount bolt was loosely fastened with a nut in the hole of the knock sensor. When the knock sensor was shaken, the nut and bolt could move aprox 10mm and the vibrations was enough to make the APC to register a knock. The solenoid was connected and an oscilloscope was connected in parallel with the solenoid to measure the duty cycle of the pulses.
The RPM was simulated by a pulsewave generator, which feed the ingition pulse amplifier.
The outgoing frequency could be altered and the output was a square pulse wave. A pulse counter was attached to the output of the pulsewave generator to measure the “RPM” correctly.

To activate the APC-system above 2400 RPM, “engine-noise” is needed. This noise comes from the knock sensor, and is generated by carefully “scraping” the bolt against the knock sensor. At low RPM the APC-system is really sensitive, so any “rattle” of the bolt registered as a knock. The sensitivity for “engine noise” is constant, independent of the RPM. While the knock sensitivity is reduced at higher RPM. This means that the RPM signal is very important to make the APC to not register regular engine noise as knock at high RPM. It also means that if you got a bad connection between the knock sensor and the APC, the phenomenon of APC activation at higher RPM than 2400 RPM occurs. Because the “engine-rattle” is higher at higher RPM even though knock doesn’t occur. But the APC still needs the same low “engine-noise” regarding any RPM to activate the APC.

The APC got a ”default-cycle” for the pulse length to the solenoid. The goal of this simulation was to investigate this. It’s already known that the solenoid starts to “chatter” (close = lower the boost pressure) at high RPM even though max boost pressure isn’t yet reached. When modifying the APC it’s difficult to keep the boost pressure at high levels at high RPM. The APC is closing the solenoid during some part of the duty cycle when the RPM is rising though it “should” be fully open. In the stock usage it’s no problem, it’s designed that way. All stock APC-ECUs are configured to lower the boost at high RPM and 

All stock APC ECUs is configured to lower the boost at high RPM and the opening of the solenoid can be an advantage in the stuck function, just to make the control circuit to ”land” in the correct ”working field”.
When jumpering of the R42 is done the control circuit senses the real boost pressure and tries to achive constant boost during the whole RPM range, but the ”default-cycle” of the solenoid will only marginally be effected by the jumpering. In reality the ”default-cycle” of the solenoid will have the same shape but be moved 200 rpm upwards in the RPM range. The solenoid will ”chatter” at 3400 RPM instead of the 3200 (if the pressure sensor senses ambient pressure).
The setting of the F-pot doesn’t effect the ”default-cycle”.
In other hand the P-pot has a rather strong influence on the ”default-cycle”.

With jumpered R42 and P-pot =15kOhm (stock), the solenoids starts to open @ 3400 RPM and is opened 38 % above 3800 RPM.
With jumpered R42 and P-pot =30kOhm, the solenoids starts to open @ 3600 RPM and is opened 27 % above 4200 RMP
With jumpered R42 and P-pot =50kOhm (max), the solenoids starts to open @ 3800 RPM and is opened 19 % above 4000 RPM

So, the P-pot does really effect the ”default-cycle” much. Unfortunately the P-pot does also control the boost rise rate, so controlling the ”default-cycle” with the P-pot isn't an option.
The influence of the P-pot over the ”default-cycle” is the cause that cars with a low base-boost runs better at high RPM if both P-pot and F-pot is at max. The APC tries to rise the boost pressure but the low base-boost setting makes the wastegate valve to be pushed open. When the RPM rises the ”default-cycle” will increase the problem, but with the P-pot at max the problem with the ”default-cycle” is mildered. This is one of the reasons that many thinks that the P-pot sometimes controls the maximal boost pressure.

The other endpoint of the control circuit is reached when the pressure sensor is disconnected. The control circuit can't open the the solenoid more than 76 % of the time, not regarding the RPM . That's why the typical ”chattering” of the solenoid when the pressure sensor is broken. The knock circuit can close the solenoid completely, not regarding the RPM.

The control circuit can't switch between the endpoints really fast. The control circuit can change the open time of the solenoid with 40% really fast, but if larger change is needed then it takes significant more time. If a larger change is needed the whole control area is moved up/down and it takes the same amount of time to move the area back. If you let the APC try to rise the boost for a while and then suddenly increases the pressure, can the solenoid ”chatter” in many seconds before the control area has been moved down to the desired boost cut.
I suspect that the rapid control circuit can be described as the ”P-part” of the control circuit, i.e. That part that only control the boost proportionally. That part that moves the control area can be described as the ”I-part”. This can be shown if the APC is slightly adjusted wrong. At first the boost swaggers a bit and then stabilizes, often on a level lower than the ”F-pot” level. When the RPM rises the boost pressure will slowly increase. If you then change gear fast, you often gets a different boost increase. This is almost surely a cause of the control area has been moved up, i.e. The solenoid is open during some time while the next gear is selected, and due to the ”bottoming out” the control circuit on the previous gear and the control area was moved.
The optimal adjustment and mod of the APC should be that the control area didn't need to be moved at all. This would need a base-boost that is almost the half of the maximum boost (which leads to the need of a stronger wastegate actuator).

The knock circuit works with different levels of boost-cut. 
At very short knock pulses the boost is insignificantly lowered and the solenoid is almost directly returned to the previous frequency before the knock occurred.
At a short knock pulse the solenoid is opened in steps of aprox. 14%. This action returns in aprox. 2 seconds.
At a more powerful knock the solenoids opened in steps of aprox. 16%. The APC is the waiting for aprox. 160 ms before it listens for more knock. If knock still occurs the solenoid is opened with even more 16%.
This is the same as the APC is listening for knock with a frequency of 6.3 Hz. This is why the ”knock-LED” ( a LED connected between +12v and pin 19) is blinking with a fixed frequency when the APC is sensing knock.

When the knock has stopped, the solenoids open time is increased with aprox. 5% per second. This is why the boost drops fast and slowly rises if the APC senses knock. To go from an fully open solenoid to a completely closed takes around 1 second if heavy knock occurs. That's why it's not recommended to run too high settings on P or F-pot in accordance to the boost pressure when knock occurs.