Sal, thanks for the Euro map.
What unit are the fuel-map values, anything comprehensible?

dfink,

You have the best approach to tuning, simply any Ign map or Fuel map we discus is simply a ballpark idea and should only be used as such. The only way to know what WOT ign/fuel map works best in your car is simply via dyno time. And it's great to share such results with others that may have similar setup but even then what works in one motor may not in another even if built the same.

But if you are starting with nothing in a custom EFI then having some ballpark ideas would be very helpful.

AFR vs Intake Air temps?

The motronic factory setup is designed to enrich mixture at load and WOT as intake air temps go up. In my car I have set my AFR to about 13.0AFR (WOT AFR) on a cold 30F day and then this same exact chip setup on a 95F day will run a WOT AFR of about 12.8. By design the motronic has a compensation map based on Intake Air Temp that richens mixture as temps go up.

I think the richer AFR is designed to help keep CHTs cooler as Air Temps increase.

Now for how the IntakeAirTempFuelCompensation map works:
First let's set some basics:
- Cold air is more dense and thus needs more fuel
- Warmer air is less dense thus needs less fuel

So, to maintain the same AFR regardless of air temp you must decrease fuel as temps increase. And the Motronic does indeed do this but it just does not decrease fuel enough to keep the same AFR it simply only decreases fuel slightly as temps increase and thus results in AFRs actually getting richer as temps increase. Kind of hard to understand but in the end the engineers decided to simply let AFRs get slightly richer as temps increase.

If the system did not compensate for Air Temp then you would be in trouble because if you tuned on a warm 90F day and set your WOT AFR to 12.8 and then used this same map on a cold 20F day your AFR would be in the low 13s maybe even 13.3 and you just went lean! Detonation!

So here is the IntakeAirTempFuelCompensation from the motronic:
10F = 146 fuel trim, more fuel
58F = 136
88F = 128
147F = 121 less fuel
You can see how they take fuel away as temps go up. But they just don't take enough away and thus AFR goes down as temps go up.

Also keep in mind that these 3.2L cars use a AirFlowMeter (not a MAF) and the AFM does not measure Mass of air but simply volume and does not account for air temp and this is why they must include air temp in the fuel trim. MAF devices account for air temp and thus don't need to trim as much based on air temp but often they also do some trimming as well.

Sort of, the digital value is some number between 0 and 255 and they are simply fuel trim values. Where the number 128 means apply no fuel trim and means FuelTrim=1 then if you require extra fuel for the given cell the numbers go up and less fuel numbers go down. Here are a few examples:
128=1.0
255=1.99 (not a realistic fuel trim)
0=0 (also not realistic)
141=1.1 (this is a realistic fuel trim, 10% more fuel)
115=0.9 (also realistic fuel trim, -10% less fuel)

So values higher than 128 mean extra fuel and less than 128 mean less fuel. It's rare to see numbers higher than 150 or less than 110.

If you are trying to figure out what the AFR is for a given cell GOOD LUCK you can't do it! You simply need to use a WBO2 for this, so I look at what I have with the WBO2 and then tweak the map, very time consuming! This is why the modern EFI systems are great, you can setup a ballpark map and then setup the desired AFR table. Then the EFI system monitors the WBO2 and fine tunes the fuel trims on the fly, great stuff if you have a system with this auto tune ability.

I suspect:

"Load" for a Motronics system is different from a more advanced aftermarket or race style system. Motronicis only knows of load in two ways. If the WOT switch is triggered or if the "advanced algorithms" conclude such.

On a more advanced EFI system load is measured in different ways. It can be measured by changed in manifold pressure (MAP) or by looking at the throttle plate angle (TPS) relative to motor speed.

What we seem to be calling "load" on the Motonics map is the rate of air flow as measured by the AFM. If the throttle is opened significantly but not enough to trigger the WOT switch, we may reach for different cells in the cruse map that are more conducive to acceleration. However, the WOT map will not be triggered.

So again, with Motronics "load" the output of the AFM. With an add on EFI it is the throttle angle or manifold absolute pressure.

Dose this make for different strategies between the two types of systems?

:)

OK.
So does 128 correspond to 50% duty cycle of the injectors or something like that?

How large are the injectors by the way, do you know?

I'm not trying to figure out AFRs, just a rough baseline. My EFI will adjust it self, as you say, with the help of a wideband.

You are close on 'load' for motronic but not exact. Load is not just AFM signal, it's the combination of given AirFlow vs RPM. The motronic has an algorithm to calculate load that uses the RPM vs AirFlow to come up with what Load row to use from the PT tables. It really does calculate load for a give RPM. But I never did exactly figure out the formula yet and it's on my wish list.

Not really sure about the duty cycle question.

The stock injectors are 24lb injectors at 3bar FP. But these are run in Batch injection mode that squirts fuel 2 times per crank rev. If you are using some new breed EFI system then you may be using Sequential injection that times the fuel delivery to the opening of the valve and in this mode sequential systems tend to use higher flow injectors like in the 30-40lb range because the really want to get the fuel in quickly right at the exact valve time. Also as RPM increases the amount of time to get fuel delivered will decrease meaning that at some RPM sequential starts to look a lot like batch. Many folks say that sequential is most useful for hi-torque lo-RPM engines and that it really is not that beneficial to hi-RPM engines. In the end sequential injection works best with larger injectors and not driving them much over 65% open. Hope this makes sense.

Idle Oscillaltion
 

Gents,

The roads cleared up yesterday so I took the P-car out for a spin. Wasn't planning to do any tuning, just enjoy the sunny day for a couple of hours.

I noticed that I've picked up an annoying idle oscillation. The oscillation cycle is about 2-seconds. The engine would idle fine, then it would jump about 100RPM, and then settle back down. While watching the VE and timing maps, it appears the fuel mixture and timing maps are fighting each other at idle.

It was a pretty cold day and MAT temperatures never got much past 2degC.

The idle AFR is approx 14.0.


T...

Trog,

How much headroom do you allow before you try to correct idle with ign or fuel? Meaning: if your target idle is 850RPMs then I would not try to correct the idle till it got at least 50RPMs or more off target (>900 or <800) I also found that simply correcting idle with ignition works best and maybe you should not try to correct with fuel. If you wish to also try fuel then kick the fuel in later but don't do both at same time.

So does your setup try to correct with ignition and fuel?

Also when you say it jumps 100RPMs does it jump 100 above target or does it also drop below target. For example 850-950-850-950-...

My car sometimes hunts for idle as you describe but only for few moments during cold start, never once warm.

Sal wrote,

The stock injectors are 24lb injectors at 3bar FP. But these are run in Batch injection mode that squirts fuel 2 times per crank rev. If you are using some new breed EFI system then you may be using Sequential injection that times the fuel delivery to the opening of the valve

Sal,

"Batch" mode is two injections per engine cycle. OEM's time Sequential Injection to the closing of the intake valve. The OEM acronym is based on injector closing X degrees before intake valve openning. For them, it is all about emissions and economy, the idea about firing the injector when the intake valve is closed is about having consistent puddling. Injecting into an open valve has less consistent results because of vaporization issues. Above 40% load it is close to irrelevent, injectors are sized to 85% duty cycle so they are spraying for most of the time available in the engine cycle (2 crank revs), intake valve open and closed.

The jump is +100RPM.

Currently I'm not using ignition to correct for idle drift (hasn't been a problem). The AFR values around idle may be the problem. I've limited the closed-loop O2 range to 10%. I may be hitting the limit.

T...

Paul,

Thanks for catching my error: for batch it injects the fuel in 2 shots per complete cycle (2 crank revs not 1).

So the 85% duty cycle seems compatible with Batch Mode, this means that you really don't need larger injectors for sequential?

I'm not that familiar with these new EFI systems or sequential injection but am very interested. I'm thinking of toying with a megasquirt install in the near future.

Trog,

You may have found your issue: very cold air usually needs more fuel even if the engine is fully warm. If your base fuel map is a bit to lean for the cold air intake temp then the WBO2 target AFR of 14.0 may simply be outside the 10% range.

Does the EFI system have intake air temp compensation maps? to allow more fuel for very cold air? If so you may need to add more fuel into those compensation maps. Don't simply add more fuel to main fuel map as this may work now for you in these cold temps but later in warmer weather it will run rich.

Sure does, currently set at 0 for all MAT temperatures below 20degC.

What about at other higher temps? Is it 0 across the board? The idea is to get more fuel at cold temps less as Air Temps go up. In the end you want to achieve a Intake Air Temp compensation scheme that keeps AFRs about the same as AirTemps change. I say about the same because in general you want your AFR (like at WOT) to get slightly richer as air temps increase. Read prior post I did on this in this thread.

Sal,

Injecting half the fuel once every revolution goes back to the old analog systems that ran off the coil pulse rpm signal. Four injections per engine cycle would be twice as many opening and closing times and they get in the way. Sequential injectors are not twice as large, big injectors are too hard to control below 2ms. For road cars, the emphasis is emissions, economy, and drivabilty. With batch injection, different things, some good some bad, are happening in different cylinders when the injectors all fire. Sequential allow the good and bad events to at least be consistent. Firing at a closed valve heats the fuel more consistently.

Minor reduction of fuel at higher ambient temperatures:

http://forums.pelicanparts.com/uploa...1233003297.jpg

Sal, You are a great resource on Motronics, thank you.

So are the two axis on the map RPM and AFM voltage or postion? Or are you saying the second axis somthing other than AFM voltage?

I understand some type of correction factors that take the active cell and modify it for O2, temp, etc. But is sounds like there is somthing else going to degermin the second axis???

Not getting it I guess.

If the cell value is just RPM by AFM voltage I can understand how we could set a line of cells that represent cruse load at cach RPM level. Then the cells just past that line where the AFM voltage is a bit higher could be a more aggressive settings. Then again eventually approaching values to achive max TQ levels as on the WOT map. How fast the cell values approach aggressive values would have an impact on part throtle response.

Sal, thank you for the ignition maps.

Does the Motronic WOT ignition map differ much from the high-load sites on the part throttle map?

The 'load' Axis is not simply AFM signal. Let's try another way to explain.

Lets say Air Flow is a value of 0 to 100, 0 being no air (AFM door closed) and 100 being door fully opened.

So we can have an air flow of 50 (door 50% opened) under many different RPM values.
For example:
- it could be 50% open at 3000RPMs under light load
- or 50% at 2000RPMs under moderate load
- or 50% at 5000RPMs under decel condition

This is why you can't just drop the AFM signal into the load axis - it must be compensated along with engine speed (RPM). The AFM door being at a given position does not alone indicate load. But if you combine the door position with RPM you can deduce load.

Also think of it this way: for a given RPM say 3000RPMs the AFM signal can vary by a significant amount based on throttle angle. At lo-load the door could be 50% open but at WOT fully loaded with a Load dyno at 3000RPMs the door could be 90% or more opened. Then on decel down a steep hill you could maintain same speed and RPM of 3000 but throttle would be closed and door only 25% open. As you can see the AFM signal can vary by 75% even though RPM has not changed.

You got it, the hi-load row in the PT Ign map does indeed match very closely with the WOT ign map. They do this so the transition from the PT ign map to the WOT ign map goes smoothly. Most chips I've seen have these set very close or the same.

Go look at the prior post with the Euro Ign Map you'll see this is the case.

Great question.

Thank you Sal. I assumed this was the case. All the transitions should be smooth.

I am getting ready to integrate spark into my EFI system and have been curious bout how much advance one could run at medium loads. I am running a twin-plug system and think I can start the map process by scaling the single-plug values down. I will have a knock sensor and use this as feedback for maximizing advance at different load cells.

So far, the fuel tuning has been very straightforward. I will wait to say the same after experimenting with spark.

Sal,

So the table is RPM by Load ("Throttle %"?).

Load is derived from the expected Air Flow (AFM V?) against some type of base map or formula.

If the base map or formula perfectly fit the motor all the cells would be the same number (128?) and / or there would be no need for the Fuel Trim table.

Also, if one piloted cruse at each rpm point, it should stay close to one of the Throttle Percentage lines through the RPM range (something near 26%?).

The numbers in the cells are used to "correct" or "adjust" what is expected from the base map or formula to fit the motor. Thus "trim" the base map or formula cell.

Am I getting warmer?

Big Thanks.

Jamie, I think the 95 993 car is twin plug car would those ignition maps help?

YES! you got it. At cruise on level ground you are generally in the 2nd or 3rd row across all RPMs.

And again YES if everything was perfect then all cells would be 128. But Bosch designed this system to fit many different setups and be very adjustable so that's why the trim tables.

You have the basic understanding for these maps!

BTW - if someone knows the exact formula for LOAD I'd love to have it. I have seen references to the formula made and it does seem to be some sort of AFM signal divided by RPM.

Sal,

Again, thank you for taking the time to teach me/us.

From what you are saying it sounds like the heart of the fuel deliver system then is basically a math formula to compute the quality of fuel needed for any size and performance level of motor relative to RPM, load, injector size, fuel pressure... But that is just a guess.

Might start with something like ((motor capacity / number of cylinder) x (RPM/2) x (VE))/target A/F ... and have a bunch of other variables that can be changed to fit any make up of motor.

This formula then computes the expected duty cycle at each point of operation.

Then it sounds like the Trim Table modifies the theoretical fuel delivery to the unique specifics of a motor.

:)

Yes, the 993 spark maps would help a lot. More compression and twin-plug in the 3.6. What I need to figure out is how the spark map relates to peak torque rpm and shift accordingly to the cam choice. My cam should have a peak torque approximately 800 rpm higher than a 964 or 993 cam.

I am running megaquirt. You have the choice of specifying the number of squirts per injection cycle as a multiple of the number of cylinders. For a 911 this can be 6,3,2,or 1. I use six smaller injections compared with one large batch fire. I have my injectors wired with alternating squirts such that 162 and 435 are grouped together. This represents the two rotations of the crank.

MS fueling is based upon an algorithm that determines the pulse width of injection (ms) based upon multiple constants.

PW = REQ_FUEL * VE * MAP * E + accel + Injector_open_time

Several of these values are constants. Req_fuel and the open time are injector constants dependent on the actual injector (flow rate and impedance) MAP is read by the engine, VE is the fuel trim number in the load vs rpm map. Accel is the enrichment from changing throttle positions. Accel can add or subtract fuel. The correction factor takes into account pressure, temperature, closed-loop O2 and warmup corrections.

E = gamma_Enrich = (Warmup/100) * (O2_Closed Loop/100) * (AirCorr/100) * (BaroCorr/100)


Megasquirt is very capable. Currently with the exception of sequential it has everything one could want including the rpm activated switches and triggers. ONe day I will install a varioram top section on top of small ITBs and control it all with MS for the best of both worlds.

Prototypes of the sequential exist and are being beta-tested. Target delivery if later in the spring. I have considered the upgrade but probably will not since my car drives great at low rpm and emissions are not an issue. Its fantastic at high rpms.

Jamie,

Here are the 95 993 ignition and fuel maps. The theory that twin plug needs less advance is sort of out the window if you look at the 993 Ign maps. In the PT Ign Map it does seem to back off on ignition for part load as compared to the 86 Euro. But the 993 WOT Ign map pushes into the 30BTDC and one wild spot is at 2480RPMs it pushes 67BTDC! I keep looking at it saying it must be wrong, but it appears to be by design. Any idea why push so much advance at WOT 2500RPMs?

<font color=red>EDIT - EDIT - EDIT - do not trust these maps! they are likely not correct!</font>

95 993 Ign Maps
EDIT: This map was wrong I removed it, see post below for correct map.

95 993 Fuel Maps
http://forums.pelicanparts.com/uploa...1233064743.jpg

Sal,

The 67 BTDC must be an error. There are plently of threads on twin plug ignition timing to read. Most published info suggests that you need 7-10 degree less max timing for optimum power with twin plugs. However, if you look at the ignition timing curves from the 911SC to the 993, this is not what the factory did. The 993 curve I saw showed a max advance above the 26 BTDC max used in almost all cases. Up until the 993, it almost looks like Porsche was married to a 26 BTDC max and altered the compression ratio to suit the fuel spec for different markets. Compression went from 8.5 to 9.8 to 10.3 to 11.3.

Paul,

It could be the chip bin I have if faulty. But if you look at the rest of the WOT timing it shows 30BTDC after 5500RPMs.

I'll double check the chip map by hand with a hex editor to see if those numbers make sense.

These curves are most helpful. Maybe it has to do with the large jump in Fuel trim at WOT between 2000 and 2480. Could it be compensation for a low-speed acceleration event based upon timing and not fuel enrichment.

I am sure that some will not share their information and I understand the reason why but I will ask the question anyway. AFtermarket tuning chips make changes to the fuel and ignition maps. In what areas are the factory specifications too conservative? Where are the areas that need improvement and how much can you improve safely? These can be approximations without specific numbers. One might say that after peak torque the engine is too lean(or rich) or that the timing in this area is too advanced for peak output. I realize that once the fuel trim is changed that timing should be altered to take advantage of the increased (or decreased) AFR.

I am currently running a twin-plug distributor. I run 12 deg at 1000 and max of 27 coming on at 4000. This is a converted SC distributor. It runs well but I think I can gain some smoothness for more lean cruise.

I have been reading the book "Internal Combustion Engine Fundamentals" by John B Heywood. There is some very good information regarding spark curves in there. I will summarize once I finish digesting.

Jamie

Hello Jamie,

OEM's are conservative because their product has a warrantee. Engine failure recalls are expensive and embarrassing. So ignition timing curves have less advance and fuel tables are too rich at high load to provide a margin of safety. An air cooled engine needs a wider margin of safety because the temperatures can vary more than a watercooled engine. Aftermarket tuners can take more risk with your engine. Higher octane fuel than the factory spec can balance this risk.

Just a though, the 993 has a knock sensor to pull back timing depending on the fuel in the market is was sold. With our fuel, we may not be seeing those numbers in real live.

Just looked at the 993 chip bin with a hex editor to be sure something was not wrong with graphical editor I use. And the ignition at 2480RPMs does indeed take a wild jump in the raw map. The hex value is a '0x74' or 116 decimal and if I apply the ignition formula to this value it comes out to be 67BTDC.

Could be that I have a faulty bin file but I doubt it.

Could that be built in by the Porsche to test the knock sensor?

Sal,

Some of the other numbers in that map look fishy to me. Part throttle timing numbers should not be lower than the WOT map at the same rpm. Did the 993 have a full range TPS and do acceleration enrichment based on TPS ?

According to this map, the highest compression, twin plug engine had the most WOT ignition timing....(3.0-3.6) They would not go to 30 BTDC, even with a knock sensor, unless output was higher.

Paul,

I know very little about the 993 cars, I just had the chip bin and know howto find the maps but that's about it. I'd just be guessing.

993 does have a full-range TPS that feeds into the ECU (I run that signal to my data logger...). Dunno how it's used since I'm no tuner.

Great thread. I run a protomotive stage 1 tt on my 1995 993, so I'm fascinated. I'd love to find a way to custom tune myself.

Something does not makes sense with the 993 WOT ignition.

Here is another site showing the 964 and 993 WOT ign side by side:
http://911chips.com/ignition.htm
Then scroll to bottom
I trust the info on this site.