scarceller

In the stock chip it has this table to comp for IAT in your case MAT.

Temp C - % fuel correction
-------------------------------
-32c +9.38%
0c +6.25%
20c 0%
60c -5.47%

That should be your target for compensating fuel base on MAT.

Also, with TPS as the only source for metering air flow you need to understand you won't ever get your AFRs perfect. I'd simply target 14.0AFR for cruise and then richen up as loads get higher. For WOT conditions target 12.8-13.0

One other tip for determining load is to simply monitor and watch your injector pulse widths. You already said that idle is around 3.8ms and from this I can speculate you are firing injectors once per stroke or every 2 crank revs, am I correct? The stock setup has injector pulse widths down around 1.6ms at idle but that's firing injectors every crank rev.

Then you want to know what's the max injector pulse width at WOT, most likely around 16-18ms per stroke. What you want to know is the min and max pulse widths something like 3-18ms and from here you can simply assume 10% load for idle and 100% for WOT at 18ms or simply stop thinking about load and just start thinking about pulse widths as they make much more sense. Also most editors I've seen for the stock chips claim % Load but that's not how the DME works, the DME has NO concept of load. It only knows pulse widths. So don't trust those load numbers in any table showing load for the stock DME.

scarceller

And here's the injector offsets (dead times) for the stock injectors

Voltage - time millseconds
-------------------------------
7vdc 0.95ms
9vdc 0.81ms
11vdc 0.61ms
13vdc 0.45ms
15vdc 0.32ms

Your values should be at least close to these taken right out of the 89 chip.

scarceller

A little more theory about fueling on TPS alone, this is not a good idea if you plan to run this engine at different elevations above sea level and neither is MAP based! If it where my car I'd run a MAF, I actually run a MAF in my car with the stock DME but I re-wrote the entire air/fuel model software in the chip to properly support MAF. If your EFI supports MAF install one and save all the headaches! You have a very accurate indication of the mass of air being ingested by the motor and it's automatically compensated for air temp and altitude!

I run a Ford Racing MAF from the Ford 5.0L Mustangs it costs about $300.00 and works flawlessly. Just get the 76mm Ford Racing MAF for 19lb injectors.

Going to MAF is the absolute best for any street car and will yield rock solid spot on fuel metering.
Buy the 2 books I recommended as Greg explains in great details why MAF is the very best for metering fuel.

ML952

wow, this is exactly what i was looking for !
so, from what you wrote i would expect that the stock ECU drives the injector from approx. 1,6ms at idle up to approx. 8 ? ms ?? firing it every crank revolution.

Any hints for cranking, afterstart and especially warmup enrichment ?

What tool do you use for reading out that info ?
What are the axle values ? I tried to decode the eeprom using excel with all the informations i could get over the internet. RPMs are clear, IGN angles also, fuel values are typically 1.02 or so -> percent of base pulse width ?
What are the load values, are they AFM-transfer function related ?

scarceller

WHoooo slow down!
You won't find any tool to get this deep understanding of the DME. I spent 4 years de-compiling the entire chip, not just the 100+ maps but all the program code as well. I'm a software engineer by trade with a EE education so this was simply a fun project for me. I'm willing to help you as I can but I won't freely distribute all the details since I plan to market a MAF maybe in the future. I know exactly what every single map does and how the actual code uses them and I'll also say that some of what's often said on forums about how the DME works is very wrong and somethings said are correct.

Hint: Load does not exist in the DME! Folks make wrong assumptions! I already told you what the load axis is, stop calling things load think in terms of pulse widths and you'll be much better served!

Tell me what you want to know about for cold start? What does your EFI offer in terms of cold start compensation? It should offer to compensate Fuel and Ign for varying CHT and IAT(MAT) does it?

So in the stock setup the min inj time is down around 1.6ms and max in the 8-9ms range, that's firing injectors 2 times per stroke.

OH and the term 'AFM-transfer function' another made up internet forum mis-conception! The DME only has a AFM Air Flow to Voltage table, sure it's not easy to understand how it works till you actually understand the code. But the bottom line is it converts air flow into a 16bit value 0-65,535 it's that simple once you know the code. By the way in my MAF code first thing I did is build a new transfer table with 0-255 data points where each data point is a 16bit value.

scarceller

OK,

So here's how you fuel for cold start after the engine has catched and is running, it looks at CHT and compensates fuel based on CHT like this:
CHT - %FuelComp
-----------------------
-32c 55%
-16c 46%
0 27%
19c 25%
51c 5%
60c 0%

so that once you hit CHT=60C no more comp needed
At 0c it's common to have AFR down in the 11AFR range. You can do the math to see the AFR, if the target is 14AFR and you add 25% fuel then the new AFR will be 14*.75=10.5AFR

But here's the other missing piece, you also need to comp ignition during warm up. You need to add more ignition advance for cold CHTs! Like this:

Ign CHT Comp
Temp C - extra ignition advance
----------------------
-32c add 20deg extra advance!
-10c 14extra
65c 0extra

Cold air fuel in a cyl has very slow flame front speeds and needs extra ign advance. Those values above are in addition to whatever the ignition is. For example: if idle ign is set to 7deg and the CHT=-10c then the total advance would be 7+14=21deg advance!

Bottom line: it's not just fuel you need to comp during cold start.

Another hint: if the car has idle hunt during cold start then you have to much fuel and/or to much ign advance. But generally to much fuel causes hunting.

60c

scarceller

Accel enrichment is not really needed in the 3.2L engine at RPM above 4000RPM. And you'll be surprised how little is needed below 4000RPMs. You really need those books to fully understand why you even need accel enrichment.

ML952

OK, understood. If this is your brainware its your choice to keep it safe and secure.
I feel even happier you still help me so far.

I have added the adjustable tables here:
http://www.vems.hu/wiki/index.php?pa...%2FScreenshots

CHT IGN adjustment is also available in newer firmwares.

How is the control strategy of the IAC-valve ?
Currently i´m driving it with 150Hz and a steady 42%-PWM.
As the effect is not seen for the TPS i must keep it steady, there are some IAC-value-conversion to TPS value, but i feel the IAC-control fighting EGO will not work at all.

This would be another point for switching to speed/density or MAF (which is actually not 100% supported afaik.

scarceller

ICV is setup so that the valve is in it's center position all the time if the idle speed is on target. If the idle dips the valve opens slightly to correct idle. If idle goes up then valve closes a tad. I did not concentrate much on the ICV portion of the code, I know where it's located and the basics but don't recall all the details. But I will say that the DME has a ton of code for the ICV! And the DME is always feeding the idle control valve pulses at all times. I believe a 50% duty cycle keeps the valve parked at 1/2 open.

You may want to simply park the valve 1/2 open and do nothing else at all. Then just try to achieve idle control with just ignition corrections.

scarceller

Looked at your maps and noticed something: During Warm up set target idle to 1200RPMs not 1000RPMs, stock chip idles at 1200RPMs then feathers down to target 880RPMs map looks like this:

CHT idleSpeed
40c 1200RPM
70c 880RPM

and in my personal setup I do this
40c 1400RPM
70c 920RPM
I like my setup better than the stock.

ML952

Do you know if IAC-valve closes from middle position to zero when the throttle opens ?

scarceller

I'm not sure, but when I have a chance I can easily check.

But let's think about the question for a moment: all the ICV does is bypass air around the throttle plate and the setup is set such that idle is achieved with the valve 50% open. So my educated guess is that you'd park it 50% when the throttle opens. Here's why: if you close it just because the throttle opened slightly it would significantly effect air flow, for this reason I'd guess it's best to park it 50% open. Otherwise the transition from idle to part throttle would be significantly effected.

Not sure what your doing now but I'd say simply park it 50% when you transition off idle.

ML952

OK, i´ll try to keep it at half, no matter what happens.
Two open questions:

Is there a MAT and/or "load" dependent IGN retard ?
Is the ICV changed to help start and/or warmup ?

Thanks.
Marcus

scarceller

The ICV is used at warm start, for example even in your ECU you can set a higher idle during warmup (like 1200RPM instead of 900RPM) this is done by the ECU opening the ICV to increase the target idle.

As for MAT based ign retard the factory DME does compensate ign based on IAT (MAT). If IATs go up then they pull out ign degrees but only under WOT conditions, table looks like this:

IAT WOT Ign Comp
TempC - Degrees Removed
---------------------------------
10c 0deg
45c 2deg
65c 4deg

But in my opinion I think you should have a strategy that takes ign out by % based on IATs and only for moderate to hi loads. If I designed a system it would have a threshold injection pulse width that if you go above this threshold you then employ the IAT Ign Comp table. Assume max pulse width is 18ms than I'd set the threshold to like 12ms and if above this you then used the table to remove ignition on a percent bases. You would do this anytime you are above 12ms or about >66% load. Extra free bonus with a setup like this is you can allow for more Ign advance on very cold days (ambient temps 0c or less) because at very cold intake temps these motors are less prone to detonate and Air Density is also higher. The idea is you would tune the engine in the Winter and then again in the Summer to find the best WOT ign for those 2 conditions and then use this IAT IGN Comp table to model the 2 extremes.

The other idea is to also comp ignition based on CHT so that if they really increase like above 200c you could remove ignition as well. Danger here is that this may not really help CHTs since removing ignition from the optimum WOT ignition only elevates exhaust temps and super heats the exhaust valves and exhaust, this is exactly what you don't want to do. My advice is don't comp IGN at WOT based on CHT since it just results in more problems. The best solution is to get off the gas if CHTs climb more than 250c. HINT: you may want to add fuel if CHTs go extremely high at hi-load as this is a much better strategy to to help CHTs come down. The other way to bring CHTs down quickly is reduce load to 50% or less and run at 4000RPMs since here your forcing tons of air via the huge engine fan past the cyls. Also don't lug these engines in extreme heat! If IAT temps are >35c run these motors above 3000RPMs always! If you run below 3000RPMs on very hot days the head temps really climb even at moderate load.

While on this topic of ignition you have one more condition where retarding ign can be good. During accel fuel enrichment it's also nice to back out ignition as it may help throttle response (Greg's book talks about this) and the factory DME has such a table! But they don't remove much at all. Then the same is true for decel condition where you want to increase ign to help slow the motor down (engine braking). This is usually only available in the high end EFI systems like Motec. But even back in the early 80s the Bosch engineers thought of this stuff for the Motronic, amazing what they knew even back then!

ML952

OK, all you mentioned is adjustable allready.
I will retard IGN for high loads like in your table.
What is the max retard value for accel enrichment ?

I made all my adjustments, here are my findings from my drive just some minutes ago:
keeping the IAC fix when opening throttle does not work yet, i´ll talk to the developers.

Everything turned out as improvment but I can´t use the stock CLT/Fuel table you send me.
I need to use my warmup table in the link above (which is much richer). Otherwise it is way to lean
and engine speed is too low due to that.

You were right i had bad settings and was only alternating the injector groups. Can you show me a typical trace of a wideband lambda run ?
So every injector group was only firing once per cycle, not once per crank rev.
I fixed that and get now some pulse width of approx 1,8ms in idle and up to approx. 8,2ms in WOT.
My signal was jumping a lot with the wrong settings, now it is better. I´m asking because i want to know if some signal jumps are normal for the (not perfect) batch fire injectors.

However some nice improvements !
Maybe you find some more stuff that would help rebuild the stock settings.