1979-1993 & 94 Conv [Subscribe to Daily Digest] |
This is zulu-india-golf, do you read me?
Well, regarding APC stuff... Nick, first of all I'm to inform you that APC ECU's installed on C900 are analogue. Yep, they're analogue, not digital. A-na-lo-gue. Like big mainframes from '40-'50s. Napoleon, Enigma, Moniac and Saab APC (the 1st revision) are of the same nature. What does this mean? I won't tell you about triggers and "AND/OR/XOR/NOT" operations. For you this should mean that there's no memory. Not a single byte. You can't store a value and restore it the next cycle. Or the next minute or hour. So there're no maps (boost vs rpm, boost vs temperature, etc). There's a simple (but working, uh?) algorithm: boost until it blows out :) Kidding (50/50). Let's take a tour thru the APC era's. You aren't on a hurry, are you?
------------------------
Intro that is a must. PWM and duty cycle.
You can try Wiki but long story short - imagine yourself triggering a switch. Once a second, no rush. Tick-tick-tick-tick. In this case the switching frequency is constant (2 secs per period = .5 Hz) and duty cycle is 50%. But let's have the switch turned on for 1.5 sec and off for .5 sec. The frequency will remain the same (it'll be turned back on every 2 secs) but dty cycle will be 75% (1.5/2*100%=75%). 100% Duty cycle means that the switch is always on. 0% means that it's always off.
PWM means that you only can turn the switch fully on or off and you regulate the power flowing thru the switch by changing duty ratio only.
Clear?
Ari, don't heart me much, here're lot's of folks that don't follow what I was just talking about.
------------------------
APC on C900 and very early 9K.
Yep, 100% analogue. System consists of ECU, RPM sensor, pressure sensor, knock sensor and the valve. The stuff works the following way: with ignition on and pressure below the preset value (resistors, capacitors - no eeprom chips and other stuff) - ECU is triggering the valve with 12 Hz frequency and ~80% duty cycle. It'll reduce the duty cycle (thus there'll be less pressure applied to the wastegate)
-----------------------
Intro - the valve connects the wastegate either to intake manifold or to atmosphere. There's no 3rd option.
-----------------------
so it'll reduce the duty cycle either when the pressure rises to specified (yep, that's about bunch of R's and C's soldered to the board) level or the signal from knock sensor exceeds the specified level (K pot, potentiometer = var resistor). The boost level at high RPM is decreased bc of the design of the ECU (it's always listening to the knock sensor and can't distinguish between a noise and the engine knock; engine is very noisy at high rpms).
Bc of analogue nature the ECU can't do the retard step-by-step. And it can't calculate the retard. The signal is of such a value? then we're going down (up). Or we stay at the basic 80% duty cycle.
The speed of changing the duty cycle is programmed by the resistors and capacitors. You can adjust it with F pot (or P? Not really sure, need to pull up the archive and drill it; but whatever, we're on a overview). With low setting you'll get slow boost restore but with high setting you'll get overshoot. But believe me that it's not the best way to go. First of all bc the pot is very sensitive and the second... You've got Bosch EFi, right? Go to DI or Trionic mod.
DI with ext knock sensor (DI/APC system) - M1988-93 9K
The system consists of just the same staff as the 1st version although MAP sensor is of another design as well as RPM sensor signal is changed a bit. And valve driving freq is now 90Hz, not 12.
MAP sensor was canceled in '91 and M1991-93 ECU's use load signal from Fuel Injection ECU to calc the boost (load/RPM=boost)
It's almost XXI century and Saab finally decides to enter digital computing :)
There's quite a few stuff on the web and no algorithm description in the factory manual but I've got the ECU and surely I've open it. Intel CPU, Eeprom chip and other stuff - it's definitely digital computer. I bet there're ignition maps and boost maps. And there're constants determining the retard of ignition or boost that ECU decides to apply. And all other stuff. Just change the some bytes in the eepprom chip and you can go to 50 psi (only carry testate estate first).
Main difference - the ecu drives the ignition and boost at the same time. But it retards ignition first and only if it doesn't help then it lowers the boost. Thus the boost is much more smooth. The 2nd difference - there're ignition and boost maps. This means that the ECU tries to maintain the level of boost that is mentioned for current load/speed/temperature and other conditions.
Combined ignition+boost+injection ECU (with ion-current knock sensing - no ext knock sensor) aka Trionic. 9K M94+ and 900NG's.
MAF->MAP, no knock sensor, rpm sensor of completely new design, APC valve is driven at different frequencies at different rpm to avoid resonance effects. 32bit MC, 32K (or 64?) eeprom - nice thing. Lot's of maps, lot's of constants. Either dealler equipment req'd or special skills to change the ECU settings.
But how does this b/s drive the boost? Man, imagine there's a huuuuge reservoir full of air at say 2 bar. It's so huge at it can hold this pressure for an hour even if we make a hole (a small one) on it's top. And there's a football ball, empty. And there's an Y-styled valve that connects the ball either to the reservoir or to the air outside. And you can trigger the switch as fast as you wish.
- if the switch is always on (the ball is connected to the huuuuuhe reservoir) then the pressure in the ball is definitely 30 psi specified above. Always on - means 100% duty cycle, refer to the 1st intro.
- if the switch is always off (0% duty cycle of the switch) then the pressure inside the ball is 1 bar (normal atmo pressure)
- if you trigger the switch with 50% duty ratio and fast enough then I do believe that the pressure inside the ball will be 1.5 bar. Agree?
Let's get back to the wastegate. On LPT Saabs (and most turbo's from other car manufactures) there's no valve that drives w/gate. Just google for w/gate actuator pic and you'll see that on such cars YOU CAN DISCONNECT THE HOSE TO THE WASTEGATE! Yep, you'll need to cut the actuator arm to have the w/gate throttle "pre-closed" at the level specified, but the idea!.. On the max boost the throttle stops closing and this means that you can just have it set to this position and forget about feedback from the intake manifold. AFAIK all turbos before Saab 99 turbo were adjusted in this way.
1st conclusion: having w/gate throttle set to specified position you set the boost the specified level. You can set it mechanically (just get it fixed at the angle you wish) or pneumatically (rod+membrane+spring+hose from the intake manifold)
What does APC do? It applies the pressure it calculates to the membrane thus setting the w/gate throttle to the position it needs to get the boost it gonna have. Let's imagine two boundary conditions:
- if the APC valve keeps the w\gate connected to the intake manifold then we have the base boost only
- if the APC valve keeps the w/gate connected to the atmosphere then... big baga-boom. But very high boost first.
This leads us to the equation I posted twice: at full throttle and high rom Pboost = Pbase/(1-Qapc), where
Pboost is the current boost level
Pbase is the base boost setting
Qapc is the current duty ration of the APC valve (1=100%, fully open=w/gate connected to atmo; 0=0%, fully closed, w/gate connected to the intake manifold)
What all this is for? Remember recommendations about Pbase setting? Yep, APC can set any boost above the Pbase level but in theory only. In practice there're limitations: 6% and 5.5% are about the same for you but can cause quite different boost levels. Remember the graph of Y=1/X function? At very small X its VERY sensitive to X volatility. Although when X ~ 100 it's about the same for X=100 and X=100.1 But check it for X=0.2 and X=0.1
*Ring-ring - rings the bell (C) - Bart Simpson*
The next time we'll friendly and no obligation talk about carpentry and home painting. I tried to catch you several times but you either don't accept calls this time or just don't pick up.
Zig
posted by 91.77.18...
, Wed, 4 Jun 2008 11:41:59
, Wed, 4 Jun 2008 10:41:51
No Site Registration is Required to Post - Site Membership is optional (Member Features List), but helps to keep the site online
for all Saabers. If the site helps you, please consider helping the site by becoming a member.
 |
 |
 |
 |
 |
