JasonAndreas

The 964 does not have a long term O2 trim, just short. There is also no O2 sensor correction during idle for fuel just ignition (similar to WOT).

Geoffrey

Wow, no long term trim for O2 correction? And none for idle? How does it maintain proper AFR readings at idle at different altitudes without barometric pressure compensation? What is the relationship between O2 reading and ignition timing at idle? This seems more simple than I had been led to believe.

Red rooster

Jason,
Whoops , there is/must be idle fuel correction by O2 sensor volts for correct cat converter operation.. Just try disconnecting the O2 with the motor idling and measure pre cat emmissions.
Agreed that by remapping, the part load fuel, transient enrichment can be increased so giving a sharper feel. I think that Lorenfb has assumed a fast O2 sensor / system response time to make the usual " no effect " statement.

This is getting very technical . Feels like being at work !!

All the best

Geoff

Geoffrey

Ignition timing at idle will affect the emissions as well as the variability of the idle speed, although there is an idle stabilizer valve (ISV) to help maintain proper RPM at different engine operating temperatures. Ignition timing around the idle RPM will help stabilize the idle and prevent stalling (LWF example). I am unsure why the O2 reading would be correcting the ignition timing though. I don't understand that feedback mechanism.

JasonAndreas

I think I need to go through the code a little more before I can say yes or no with any certainty to the O2 @ idle or the O2 modifying the ignition timing other than the timing changing while the O2 sensor is not upto temperature. There is definitely no correction of the fuel trim or injector opening values like there is during partial throttle.

Red rooster

Jason,
So if you are right ,modifying the idle fuel map would lead to a corresponding idle fuelling level ? This just doesnt fit with what I have een /done/understood !!
Very interesting though.

Geoff

Lorenfb

"The DME contain 3 seperate fuel maps" - Jason -

Right! But NOT separate maps for EACH O2 value. As I said before there're offsets
(added/subtracted values) to the basic/neutral maps & as Geoffrey mentioned
too.

"Ignition timing at idle will affect the emissions as well as the variability of the idle speed," - Goeffrey -

Right! And there's really NO need to be changing idle timing other than for different country
requirements (NOx) as was the case for the 3.2 DME.

"Its just that there is no direct fuel trim/lambda correction for the O2 at idle (like there is for part throttle)" - Jason -

What?????? There's O2 control of Lambda at idle.

"Wow, no long term trim for O2 correction?" How does it maintain proper AFR readings
at idle at different altitudes without barometric pressure compensation?"

- Geoffrey -

That's OBII 993 and later. The 964 has a pressure sensor input.

"Idle fuel is controlled by the AFM which is controlled by the ISV dwell factor"
- Jason -

Of course! Just like on any engine where the amount of intake air affects the fuel intake,
e.g. carbs, MAP system, EFI, etc.

"The O2 feedback is ignored during partial throttle if the AFM load factor is increasing (i.e. acceleration) by a programmable amount."

- Jason -

Very short term at open loop, and that's why DUMPING more fuel really does
little in a performance chip versus "pushing" the timing. Just check the fuel maps
of various chips.

"AFM ADC increases by 3 steps the O2 is ignored temporarily." - Jason -

That's rather small: (0 < AFM < 5.0 volts / 256 (8 bit word)) X 3 ~ = 60 mv.

"The DME contain 3 seperate fuel maps (ROW, USA, M150=NO O2) and 6 ignition maps for partial throttle (ROW, ROW+TIP, USA, USA+TIP, M150, M150+TIP)!"

Right! But not all the ignition maps are selected at once, i.e. the DME coding
by the "jumpers" (coding plug etc) determine the maps selected.

Lorenfb

"Here is the fuel map @ idle, so there has to be an error in my disassembly of the routine that processes the O2 input and sets the flag stored at (temporary) ram location 0x2F.0."

I don't know what you're using to disassemble AND whether you're interpreting the assembly
language properly. The O2 process is very simple for interfacing to a narrow band O2 sensor
as used on all Porsche DME until the wide band O2 sensors, e.g. Cayennes:

1. A reference O2 voltage is set @ about 500 - 600mv via comparator/op amp
2. If the actual O2 voltage is above the reference (rich), then the injection pulse width is reduced
by some delta time
3. The u-p then waits and then samples the O2 comparator, if its output is still greater
than the reference, the injection time is again reduced by the delta time.
4. The system will then reach a point where the sample will be less than,
and the u-p will increase the injection time.
5 The opposite process occurs for a lean condition.
This process can be observed with the 964 as "idle hunting" (25 - 50 RPMs).

This basic operation occurs whether the comparision of the O2 voltage
is done before of after an analog to digital conversion, i.e. a byte comparison
in the 805X.

"There are separate base fuel & ignition maps available if you disconnect the O2 sensor but the ECU won't use them unless jumpered."

That's right, but you MUST "tell" the DME that no O2 sensor is being used. The DME doesn't know
the you've disconnected O2 by anything other than:

1. The input O2 voltage is out of range. or
2. The input O2 voltage is not changing.

And these just cause a DME fault. Basically, the DME couldn't care less whether the O2 is
working/connected since it still has the reference voltage for a nominal mixture setting.

"Do you know anyone that actually tried to use the altitude sensor/switch? "

It's pin 46 on the DME. I'm sure it's used just like on the 3.2 but as an analog input versus a switch
input on the 3.2.

Geoffrey

I thought that there was an idle swtich on the TPS that selects the Idle map and a second switch that selects WOT map. Anything in between is using the 3d AFM/RPM map.

Where is the barometric pressure sensor in the system?

Jason, The graph you posted is the idle correction map or the idle target map? What AFR does the ECU try to get the engine to idle at?

"Very short term at open loop, and that's why DUMPING more fuel really does
little in a performance chip versus "pushing" the timing. Just check the fuel maps of various chips."

The real reason increasing fuel does little for performance gains is that within a spark ignition engine there is a wide range of AFRs where the engine will make power within in 5% of peak power. In other words, changing the AFR won't make significant gains. We're talking about 11.5:1 to about 15:1 AFR. We demonstrate this premise in the EFI classes using steady state on a dyno and a programmable engine management system.

Ignition timing will cause a wide range of power output since you are really looking for peak in cylinder pressure somewhere between 12-14 degrees ATDC. It also affects the EGT.

Red rooster

Jason,
My 964 ran for years with a O2 probe but with no cat converter. Annual emmission checks at idle always showed lamda 1 at the tailpipe . Dyno tests had WOT running around 0.9/0.85 .
A Hammer check on an idling 964 will show the O2 sensor volts cycling in a way normal for closed loop and it will also tell you that operation is closed loop.

The base idle fuel map I have had to alter on motors with big cams/injectors etc to bring fuelling back in range for O2 operation.

Just trying to add some real world stuff to the discussion.

Geoff

Lorenfb

"thought that there was an idle swtich on the TPS that selects the Idle map and a second switch that selects WOT map. Anything in between is using the 3d AFM/RPM map." - Geoffrey -

That's correct.

"Where is the barometric pressure sensor in the system?" - Geoffrey -

It's next to the DME unit under the left seat. It wasn't used on O2 cars because of the TRA function,
idle CO adaptation, in the DME. Try and get the Porsche manual WKD 495 121 which is the '89 964
technical manual discussing some of the issues on this thread.

"My 964 ran for years with a O2 probe but with no cat converter. Annual emmission checks at idle always showed lamda 1 at the tailpipe . Dyno tests had WOT running around 0.9/0.85 .
A Hammer check on an idling 964 will show the O2 sensor volts cycling in a way normal for closed loop and it will also tell you that operation is closed loop."

- Geoff -

That's it! Exactly what I've seen working on many 964s over the last 10+ years.

JasonAndreas

I'm using IdaPro so the disassembly is not the problem just the interpretation.

That is basically the code I posted minus the actual fuel correction part.

I posted the idle fuel map. What do you mean by "idle target map"? The RPM or the amount the ISV is open or ?


Without O2 correction the graph I posted is the AFR that the ECU is shooting for. The corrected output from the AFM is multiplied by the injector constant(s), ECU tick/timer duration, the idle fuel map and some additional temperature compensation.

That's what I was hoping for! I know you and a few others have been down this road before so I thought maybe something in the disassembly of the O2 processing code I posted above might stand out?

Geoffrey

So you are saying that Porsche designed the idle fuel map to run the engine at .875 (12.8:1) AFR? I think that is suspect. On the 3.3 Turbo engines with Bosch fuel injectors I can get them to idle smooth at 900rpm at a 14.1-14.2:1 afr. Any leaner and they begin to become rough. I would expect that with the higher compression and therefore better efficiency at idle, as well as emissions requirements they would idle in the 14.7:1 range.

RE the graph, The ECU has a 2d idle map which shows RPM as the x axis and the values in the table represent pulsewidth. I understand that point. The graph you posted looks like it also comes from a 2d table where the X axis is RPM and the values in the table represent AFR. So, is that the target AFR table the ECU uses in closed loop lambda control? I'd be suspect of that because the narrowband sensor isn't going to be able to read that rich and its purpose is to hold the engine at/near 14.7:1 so I doubt it can make the correction from 12.8:1 to 14.7:1.

Red rooster

Geoffrey,
From memory the idle graph is injector duration , pre modifiers .
Jason is going to force me to hook up my bench 964 system !

Too partially answer the question , disconnecting the O2 probe at idle would show all. If Jasons understanding of the dissasembly is correct there would be no fuelling change .
I am pretty sure that is wrong.
I would do that myself but my 964 has a 993 motor/DME.That would cloud the issue.

All the best

Geoff

Geoffrey

"From memory the idle graph is injector duration , pre modifiers .
Jason is going to force me to hook up my bench 964 system ! "

I would assume that would be correct, just like the main 3d fuel table (AFM vs RPM) would contain the injector duration "pre modifiers" or as I would call it, the base pulsewidth prior to any compensations.

I just don't understand the graph posted above.