| Home | Gallery | Fire | Specs | Vac diag | Hardware | Seals | Faults |
Background Notes Oz951
Fuel management system – background information
OPEN LOOP – this is the condition where the DME is not using the oxygen sensor output to adjust the fuel mixture. This condition is triggered when the throttle is opened >60 in a short period of time. Open loop means that the fuel management system is using a predetermined amount of fuel derived from a table (map) within a chip or similar.
CLOSED LOOP – this is the condition where the DME adjusts the injector duty cycle based on the air-fuel ratio as measured by the oxygen sensor. Note that closed loop mode can be used all the way to redline if the throttle is opened gradually. Rapid gear changes and hard acceleration will of course send the DME into open loop.
Air Flow Meter (AFM) Output Signal
The way the AFM works is to output a voltage from 0-5V representing the amount of air flow. The 944 computer or DME expects to see a voltage somewhere between 0-5V from the AFM.
Adding in modifications to get more boost means that the AFM would hit that 5v max output much sooner than redline. To cope with the extra mods it is necessary to massage the air flow signal being sent to the DME, this is what the ARC2 and Link MAP systems do.
When the DME sees a particular voltage and it is operating in closed loop mode (after rapid throttle opening) it goes to its chip based fuel map and finds an injector duty cycle number. This will deliver a predetermined amount of fuel (for a given fuel pressure) that should match the airflow being measured if the fuel map is good.
Limitations of original AFM – the original AFM was calibrated for a stock motor such that it would be providing 5v output at an airflow that corresponds to the maximum airflow that a stock motor could achieve (call this airflow X). When the engine is modified with bigger turbo’s and exhausts etc its air flow capacity is increased. On this modified engine the stock AFM will hit a point where it is outputting 5v at airflow X and as the engine continues to higher RPMs and greater airflow the AFM will still continue to be capped at 5v because that is the maximum output signal. The DME will see the fixed 5v signal and continue to use the injector duty cycle that corresponds to that 5v. But since the airflow is actually increasing whilst the fuel delivery remains capped the fuel mixture will be getting leaner and leaner as RPM increases.
MAF systems measure the air flow and generate a similar range of voltage to what the AFM would do (0-5v), the chips that go along with the MAF will generally have their own fuel maps that will use higher duty cycles to deliver more fuel that is needed by cars with mods such as MAF & bigger turbo etc. The ARC 2 allows fine tuning of the system by further massaging the air flow signal up or down (within the 0-5v range). The massaged air flow voltage will then just correspond to a slightly different number from the fuel map in the chip.
MAF Adjustments – in a poorly matched MAF system the output voltage my not correspond well to the air flow capability of the engine so that you end up in a situation like the AFM where the voltage becomes capped at 5v for a given airflow. Adjustment of the ARC 2 will be of no benefit from this point on because the output will already be at 5v and any positive increments to the airflow signal will have no effect (because its already at 5v) Whilst tuning on a dyno this could give the impression of being in closed loop because +ve adjustments of the fuel computer will not increase the richness of the mixture.
The link MAP system generates an air flow voltage between 0-5V that can be input to the DME. Some tuners provide link systems that come with chips (for DME & KLR) that are burned with fuel maps to suit certain cars. The benefit of the system is that the fuel maps (injector duty cycles) are based on particular mod configurations and injector sizes so that the duty cycles in the fuel maps are likely to be producing good mixtures as long as your mod configuration is matched to the chips that you have purchased. The Link tuning module acts in a similar way to the ARC2 in that it allows the user to add or subtract a certain voltage from the standard output at particular RPMs and load conditions. This adjusted airflow voltage will then correspond to a certain injector duty cycle from the chip fuel map.
Fuel pressure regulator - Increasing the fuel pressure will allow the standard injectors to deliver more fuel for a given duty cycle and will allow an across the board increase in richness. This can be useful for modified engines which have a higher airflow and which need more fuel.
Fuel injectors – At some point it becomes necessary to increase the capacity of the fuel injectors because the engine capacity may demand more fuel. Initially the fuel delivery can be increased by increasing fuel pressure and then by increasing duty cycle but when the injector is being run at max fuel pressure and max duty cycle it will be necessary to install larger capacity injectors if more power is needed from the engine. Once bigger injectors are used it will be necessary to match the fuel pressure and duty cycles (fuel map in the DME) to suit the delivery rate for the new injectors.
Air Fuel Measurement. – Lets assume that you have purchased a MAP kit with chips to suit a bigger turbo and 65lb/hr injectors. This kit will have the MAP kit programmed so that it will deliver a 5v signal at the expected airflow capacity of your engine and it will have a duty cycle based on the 65lb/hr fuel injectors and it will be expecting that you operate the FPR at a certain fuel pressure (say 3 bar). The MAP kit will deliver reasonable fuel ratios and power because the chips and their fuel maps are reasonably matched to your cars mods. However every car is a little different ie exhaust size, head flow characteristics, specific turbo model etc. These differences will result in slightly different flow characteristics from the vehicle used to tune/write the fuel map in the chips you obtained with the MAP. Those differences in flow characteristics may require you to slightly adjust the air flow signal output by the MAP unit at certain RPMs That will cause the DME to use a slightly different injector duty cycle at those RPMs so that you achieve the correct mixtures at those RPMs. In order to do this tuning it is necessary to know the AF ratio at each RPM so that you can spot the points that need adjusting. You can then use the tuning module or software to adjust the MAP kit air flow signal at that particular RPM. Wideband O2 sensors with RPM logging are needed to take these kinds of measurements. The typical dithering AFR gauges are not sufficient for this purpose as they are really only good for telling you that your mixture is at stoic or not at stoic. When these cheaper gauges show stoic they are accurate but when the gauge reads anything other than stoic the actual amount of richness or leanness indicated is very inaccurate (the regular sensor display is only designed to provide an accurate reading at stoic)
LINK Logging – use the link system to log the air flow signal all the way to redline in each gear and verify that the signal only gets to 5v at about redline. That will indicate that the output of the LINK air flow signal is a good match to the flow capacity of your engine. If its hitting 5v before redline your engine will be heading towards lean as rpms increase beyond the cap point (rpm)
944
fuel pump relay – Pressurising the fuel system
To
help diagnose in a safer fashion, don't turn the car on. Pull the DME relay, and
run the fuel pump by jumping a wire as follows. If you look at the holes
standing on the drivers side of the engine compartment you will see three rows
of holes. The first row has three holes, the second has one hole, and the third
has two holes. Jump a wire from the middle hole of the first row to the single
hole of the second row. This will run the pump and if its a fuel rail leak or
more likely a fuel hose leak at the crimp connector you will be able to see
this.
Full fuel system problem diagnosis & result
The text below is from a post and some follow up responses I put on Rennlist
Here
is how I spent my weekend: After doing all this I realized that this list of
steps is a bit of a guide on steps involved in diagnosing a no start condition
so I thought it might be a handy reference for the future.
After putting my car back together after the fire I got ready for the first
start by doing all the basic checks for fluids, connections, free rotation etc.
I also powered up the fuel pump from the DME relay socket and checked for fuel
pressure and absence of leaks. All was OK. But when I went to start it it
wouldn’t run. The car was configured with 65lb injectors (with ballast
resistors), AFPR, and MAP & guru chips, new ref sensors new coil, wires cap,
rotor and O2 sensor. The fuel rail showed pressure so I suspected ignition. I
swapped the DME relay for a known good one, still no start. I did notice that
the plastic around the DME relay pins on the fuse box was slightly melted –
the DME relay that was in the car during the fire must have had a high current
run through it when the fire shorted out some of the wiring. None of the wiring
inside the car looks like it was ever hot or melted and I also had a good look
at all of the wiring underneath the fuse box and it looks normal.
I checked for spark at the coil – it was good and then at the plugs – it was
good. (so the reference sensor, coil, wires, plugs, DME and cap/rotor were
good). The Apexi AVCR lets you monitor the TPS output and I noticed that the TPS
output wasn’t changing when I pressed the accelerator, so I replaced the TPS,
still no indications of throttle opening. I then noticed that the boost gauge on
the dash was actually showing a slight positive pressure with the engine
off….. that’s odd. Since the TPS and boost signals come from the KLR I
swapped the KLR out for another one. Now the TPS and boost indicated correctly (KLR
bad). Still the car would not start.
I checked the fuel pressure again and realized it was actually a bit too high.
So I pulled off the AFPR and inspected it, sure enough it had a bad diaphragm.
So I replaced it with a new 3bar unit (to suit my MAP system). Fuel pressure was
lower but still no start. I decided to run the stock AFM because at that stage I
was not sure of the functionality of the MAP electronics unit (had it been
shorted in the fire ?) I took the MAP out of the loop and put the AFM in. I
decided I was also going to put a 2.5bar FPR in their as well to suit the stock
AFM. So to get rid of excess fuel pressure I decided I was going to pull the
fuse for the fuel pump and crank the engine, well I’ll be damned …the car
started! I ran it for about 1 minute by babying the throttle and then killed it
to check for leaks and problems. I put the fuse back in and went to start it
again…. No start. I pulled the fuse again….it starts. Weird. I didn’t know
the car could run without the fuel pump on. It actually idles nice and stable at
950RPM and whilst doing so the fuel rail pressure reads zero. I have not let it
run for more than a minute or two like this. I put the Map back in and by this
time I had a hand controller for it (thanks Dwayne) and was able to verify that
it still had all the correct inputs and outputs and had retained all of the
tweaks to the maps. Just like with the AFM I was able to start the car only with
the fuel pump fuse removed. At some point I had already tried another DME and I
switched DME’s again, both would allow the car to run only with the fuse
removed. I tried the different FQS settings for different injector sizes and
that made no difference either.
So at this point I am thinking the engine does not want to run with high fuel
pressure, so I go ahead and swap the 3 bar FPR for a 2.5 bar FPR and it makes no
difference. So the I think OK is the high pressure causing high injector
electrical load on the DME. (high injector load from high fuel pressure can
overload the DME so it shuts down the injectors) I pull an injector clip and it
makes no difference. I have one odd injector clip on the harness so someone has
been messing with the injector harness at some point. I check the resistance of
the harness to all injector plugs and I think I get a high resistance at one
time but when I check it again it seems OK. Maybe I have intermittent high
impedance in my injector harness? OK I already have 4 new injector clips so I
replace all four injector wires and verify the resistance of each wire back to
the DME Plug, they are all good no matter how much contorting of the harness I
do, this step also verifies that the ballast resistors are good. Still the car
does not start with the fuse in.
I am starting to scratch around for ideas now. So far as I know a bad fuel
dampener shouldn’t cause a high fuel pressure problem but I have another one
sitting around so I swap that out, no difference. OK I decide to swap all the
injectors back to the stock items and take the ballast resistors out… still no
go. Could I have a blocked fuel return line causing pressure in the rail to be
too high ? To check that out I disconnected the fuel line from the fuel pressure
regulator and also where the hard-line joins the rubber line at the fuel tank,
then I blew that line through with compressed air. I also blew lower pressure
air through the short rubber hose at the fuel tank and could hear the bubbles in
the tank. I collected the fuel from the return line and there were not any
debris in there. So that does not appear to have been the problem either. Whilst
I was back their I removed and cleaned out the fuel pump check valve but that
did not fix the problem.
Hmmmm WTF, actually this was not the first time I had thought WTF, but now I
meant it even more. At this point I had swapped out or tested all of the
following:
DME, DME relay, KLR, various FPRs, dampener, Injectors, Injector harness, check
valve, ref sensors, rotor, cap, coil, wires, spark plugs, TPS, MAP Kit, AFM,
ballast resistors, fuel return line, 02 sensor.
Next steps:
I will plug a Noid Tester in to verify injector pulses so that I can see if the
DME really is shutting down the injectors. There may be a wiring issue somewhere
that is only present with the fuel pump fuse in place. I checked it out more
last night and noticed that the o2 sensor and the turbo coolant pump relay also
operate of the fuse so perhaps there is an issue there. I will hotwire the fuel
pump to test that scenario.
--------
Well
here is the latest, I plugged in a noid sensor last night and verified that the
injectors do get pulsed when the fuel pump fuse is in. With that in mind I tried
starting the car with various amounts of throttle and eventually got it to start
with the fuel pump fuse installed. It was way way rich and pumped out plenty of
black smoke from the exhaust and stalled as soon as if I let off the gas, prior
to that though I revved it higher around 3-4000 rpm and the smoke was a lot less
at those revs. The smoke was also less after a minute or so and that was
probably a result of the O2 sensor warming up. I watched the AF gauge go from
its starting position of 0.47v which would normally be slightly leaner than
stoich and steadily move around to very rich over a period of a minute as the 02
sensor warmed up. The gauge never dithered and the only time it did not read
rich was when I would modulate the throttle quickly.
Since the rail pressure is normal and the injectors are functioning at that
pressure I have to look at whether the injectors are being commanded to provide
extra fuel (cold enrichment or bad airflow signal), or whether the injectors
have a wiring problem that is inadvertently causing them to add more fuel. I
doubt there is an injector wiring problem since I have replaced all that and
cross checked the resistances. The airflow signal is not a factor in the MAP
system so now I have to look at cold enrichment. I am thinking maybe a bad
engine temperature sensor or associated connector/wiring.
-------
Yes it turned out to be the temp sensor, the one with a connector like a fuel injector. The sensor itself turned out fine but the connection to it was bad. That means the DME was figuring on a very cold engine and commanding lots of extra fuel. I repaired the connector and now she runs
Wastegate and boost controllers – background information
Single vs Dual Port Wastegates
Before I get into how the boost controllers work it is necessary to have a few clues about how a wastegate works. The stock 944 wastegate contains a valve like the exhaust valves in the cylinder head. The valve is connected to a diaphragm within a small chamber. On one side of the diaphragm there is a spring to resist movement of the diaphragm until a certain level of boost, say 13 PSI. Boost pressure from the intake system is fed to the other side of that diaphragm which causes it to move when the boost pressure is high enough and that opens the valve. When the valve is open, exhaust gasses take the path of least resistance and go through the wastegate instead of having to push their way through the turbo. Because the exhaust gasses are now bypassing the turbo the turbo starts to slow down and so the boost pressure will drop as well. When the boost pressure drops sufficiently the boost signal to the wastegate diaphragm will not be enough to keep the valve open, so it will close. Exhaust gasses now have to go through the turbo again and boost will start to build again if there is enough exhaust gas flowing.
MBC with stock wastegate. A manual boost controller is an adjustable restrictor that goes into the line between the wastegate and the banjo bolt connection to the intake. If it is fully open, all of the boost signal will get from the banjo bolt to the wastegate and the wastegate will open as if no MBC were fitted. If the MBC is adjusted so that it restricts some of the flow going through it then the boost signal that goes to the wastegate diaphragm will also be reduced. That means that the wastegate will not open as soon so the turbo will continue to generate boost ie you will get a higher boost level when the MBC is made more restrictive. (turned counter clockwise towards negative).
MBC with dual port wastegate. A dual port wastegate has two sealed chambers one, on each side of the diaphragm. As usual the boost signal is applied to the underside of the diaphragm in order to lift it and open the valve. If zero boost pressure is applied to the top side there will be a pressure differential and the valve will open. If some boost pressure is applied to the top side then it will offset some of the boost pressure on the bottom side and the valve will open later when boost pressure is higher. An electronic or mechanical boost controller is used to control he boost signal going to the top side of the wastegate. A MBC with a dual port wastegate must be turned clockwise towards the + to increase boost, this is the opposite of the case with a single port wastegate. The reason is that as the MBC is turned towards + it is sending more boost signal to the wastegate top port and off-setting more of the boost on the lower port. That means that a higher level of boost will develop before the wastegate can open.
MBC faults
I installed an MBC initially and it worked well after some teething problems. Just after I had been through a trial and error method of setting the boost level, one of the plugs in the MBC came unthreaded and blew out. That meant that all of the boost signal was escaping and the wastegate was not opening when it was supposed to, this could be seen as higher than expected boost on the gauge. It took a little while to find this fault because the offending plug was the one that was out of sight the way I had the MBC mounted. Do not assume the plugs are all in tight from the factory….tighten them yourself prior to installation.
Pretty simple to install electrically:
Speedo - dark blue w/red wire coming out the ride side of speedo (the vertical
wiring connector (connector B) at the instrument cluster)
Throttle position - KLR terminal #22
RPM sensor - DME terminal #1
OPTIONAL (I wouldnt bother) --- Injector duty cycle -DME terminal #14, Install a SPDT switch to toggle between RPM and injector Duty.
14.2psi == 1 kg/cm2
14.5psi == 1 bar
Rennlist posts by Rage 2, Danno et al -----> The feedback speed might be off for your wastegate. Did you learn the
gears? If so, go to the F/B speed. If it's underboosting, turn down the F/B speed. If it's off, start off with a middle number (5) then slowly
turn down the speed to get the proper max boost (or turn up the speed if you're overboosting). If I remember correctly, I used a setting of
3 for all 5 gears on my stock wastegate (It's been a while). You can set this for every gear to ensure you get the same amount of boost at
all times.
One other item, check the "learned" boost curve, at over 5000rpm (approx) the duty should be 100%. That's about when the turbo fully runs
out of steam.
Once you get the boost dialed in for all 5 gears, you can edit the boost curve, so it's not flat. Add more boost on the bottom end, you can set
rpm specific boost levels, and it'll automatically update the duty cycle to your settings. It's really a lot of fun =).
The Apex-i isn't hard to set up, it's the manual that's written really poorly. Totally useless manual IMO. A quick lesson in getting the Apex-i set up :
1. Set gear judge, drive in 5 gears, any speed, AVR-R will know what gear you're in
based on speed vs rpm.
2. Choose Program A or B. Set target boost level (say, 1.00kg/cm3)
3. Start at a low duty, say, 50%.
4. Set to monitor the boost, then do a run in 3rd gear to redline (you'd want a highway)
5. If boost too high, lower duty. If boost too low, raise duty. Repeat step 4.
Once the target boost is close enough, the AVC-R will then set a duty curve
automatically to get to your target boost level. It'll keep learning as you drive.
Final step, if you're not getting enough boost in 1st and 2nd, you can use the start duty
for the individual gears and raise the initial start duty cycle to get boost to shoot up quicker,
and hit your desired boost. It's a dangerous feature IMO, cause you'll overboost in those gears
going uphill.
Of course, this is if you want to get the most out of the boost controller (get to max boost quicker without overboosting). You can use it in what I call "dummy mode"
where you set a flat duty cycle and just get the same boost level at all rpm's (if your turbo can provide) with a small sacrafice in turbo response.
That takes about 10 seconds to setup.
PSI kg/cm2
5 0.35
6 0.42
7 0.49
8 0.56
9 0.63
10 0.70
11 0.77
12 0.85
13 0.92
14 0.99
15 1.06
16 1.13
17 1.20
18 1.27
19 1.34
20 1.41
21 1.48
22 1.55
23 1.62
24 1.69
25 1.76
26 1.83
27 1.90
28 1.97
Pertinent background info
1. First up I have replaced all my vacuum tubing and intake hoses and there are no leaks.
2. I blocked off the Idle stabilizer valve initially whilst I sorted out the low rpm zones.
3. I put the system into open loop with a 1.8K resistor across the altitude sensor connector.
4. My 951 is not stock it has various mods that are listed on my website
5. The whole point of this tuning was a) to get the car driving smoothly and b) to get the car running slightly (safely) rich whilst I do my test drives prior to going to the dyno for fine tuning under load with a wideband o2.
6. I used a regular narrow band AF gauge to monitor the mixtures (in open loop it does not dither) and I used either the hand controller or a laptop to monitor zones etc.
Observations – Revving at idle, I found that the rpm initially fluctuated even if I didn’t move my foot. The cause was the map switching from rich to lean zones and back. The different mixtures affected the rpm. Ie at idle it was in zone 305 which was fine if I revved it slightly it would go to zone 205 which was lean and it would then cycle between those zones so it seemed like the idle was hunting. I fixed that by increasing the fuel in zone 205 until it was just slightly rich. I then revved to 1400 rpm (zone 310) and the idle would hunt again because the system would cycle between zone 310-rich and zone 210-lean. So I increased fuel in 210. I kept repeating this process in 500 rpm increments until I could rev it as high as I liked and the rpm was stable. This worked very well. The actual increases in fuel in each zone can be seen on the worksheet . Then I took the car out for drives in 1-3 gear and had my girlfriend call out the zones whilst I a) looked where I was going and b) watched the AF gauge. I made a note of which zones were lean and which were overly rich and then we stopped and made changes in those zones. As a starting point I made each row have about the same level of fuel ie zones 400-475 I set to +6%. Once the low speed driving stuff was sorted out we went for a short drive on the highway and I tested low boost initially (to ensure it was not lean) and then higher and higher boost. If I observed lean conditions whilst on the highway we made a not of roughly what row the map was in and then we used master AFM to increase fuel a percent or two to see how much extra was needed in that area to bring it out of lean.
Based on those observations I made more changes to various rows of the map. Then I plugged the laptop in and did some data logging. I wanted to see where the map was spending most of its time and which zones it was moving through to get there. I used Excel to analyse the data and the spreadsheet is here . The spreadsheet allowed me to see which zones the map spent most time in (that’s how I colour coded the zone worksheet) and when I sorted the results on zone and o2 reading I could see which zones were rich and which were lean. 82-86 is ideal for the AF reading in the logs but I was initially aiming to make sure it generally stays above 80. Whilst driving, when you back off the accelerator the fuel will be cut and the map will drop to a 200 series zone and the mixture will show extreme lean-that is ok. Note also that the logs show the zones as a number between 16 (zone 100) and 159 (zone975). The first sheet of the spreadsheet shows the overall results of a a5 minute drive with a few short boost runs but mainly freeway driving. The AF mix is generally stable around the high 80’s.
Copyright
© 2003 , — All Rights Reserved
Porsche® and the Porsche Crest® are Registered Trademarks of Dr. Ing. h.c.
F. Porsche AG
All other trademarks used within this site are property of their respective
owners