It seems that one of the most challenging aspects of our cars, is getting that little silver box between the seats to agree with what we have done to our engines

In this thread I will share what I know about computers (like the one Im using now, and like the one in my car) and ask our resident experts (Darth?, Rockcrawl?....) for there advice or knowlege about the ECM and how to tune it.
I am not good at tuning EFI's by any means, but I know alot about the system, and would like to get familiar with tuning it. But In order to do so, I need to know everything I dont already.

I know that 1/2-2/3 of the computer is used to interperate data coming in, and data going out. Turning the data into a value between 0-255 in the form of hexidecimal (witch is represented with a $ sign) $00-FF, and turning the $00-FF into useable outputs, such as pulse width for the injectors. 0 or 12 Volts on and off for solenoids, lights and motors.

The smaller part of the computer is just that. A very basic computer. The microprosessor is a motorola HC6811 (not quite a Pentium, not even a 286!) and it operates at I think 8Kbps. The microprosessor is stupid. All it does is do what its told to do. And that information comes in the form of a program witch is stored on the later half of a PROM (Programmable Read Only Memory). That PROM also has a whole section in the beginning of the chip that stores fixed variables, such as timing advance curves, Idle, when to turn the fan on, the check engine light, and so on. There is also a special part, either on the computer board itself, or attached to the PROM held within a modular unit, its called the CALPAK. What this does is it stores a small amount of data that is used to tell the microproessor how to function. It tells the computer what to do if the microprosessor (We'll call it the uP from here on) fails to run properly. Also, if your car is equipt with a knock sensor, this calpak interperates the signals coming from it.

Now that I've told you the very basics about our GM computers, now comes the interesting stuff.

Hexidecimal is a numbering system. The name kinda stands for 6+10 or 16. Decimal is our standard numbering system, where when we get past 9, we start over at 0, but put a 1 in front of it, and you have 10. Where Hexidecimal is different is, we go 1,2,3,4,5,6,7,8,9,A,B,C,D,E,F, THEN we get to 10. (witch is really 16 now isin't it? thats were it got the name hexidecimal) Makes you wish you had 6 more fingers to count on though Then there is the numbering system we call Binary. Same consept, except theres only two numbers we are allowed to use. 0 and 1. this graph should help alot to compare the numbering systems
This is a VERY IMPORTANT thing to understand if you really want to know the nitty gritty about how computers work.
First colum is Decimal, second is Hexidecimal, third is binary.
,00 00 ,0000
,01 01 ,0001
,02 02 ,0010
,03 03 ,0011
,04 04 ,0100
,05 05 ,0101
,06 06 ,0110
,07 07 ,0111
,08 08 ,1000
,09 09 ,1001
10 0A ,1010
11 0B ,1011
12 0C ,1100
13 0D ,1101
14 0E ,1110
15 0F ,1111
16 10 10000
now your probably asking yourself "Why is this so important?"
The answer:
Computers don't have 10 fingers to count on. They only have "2 fingers to count on" The wire is either:
HI or LO
1 or 0
5 Volts or 0 Volts.
There are no exceptions. So what some genious thaught of is using binary to create numbers that computers can "use" to do things. The problem is, Binary is ugly. How can you look at 01101011 and know that it means 107? perhaps thats alittle too hard for the average person. (I admit I had to use my chart and a calculator) We can make it easier! So, thats why we have Hexidecimal to the rescue. Most uP's have 8 wires for data to pass through, Lets take that 01101011 number and put it through those 8 wires, the first one is 0 volts, the next 2 are 5 volts, the next 4 are 0 volts and the last one is 5 volts. So what we have accomplished is we put a number through 8 wires. Those 8 wires is what the uP uses to communicate with whatever it needs to. Now heres where Hexidecimal comes in. According to my little chart, 0110 is 6 and 1011 is B so 01101011 in hexidecimal is 6B. Where did I get 107? Simple, take the first number and multiply it by 16, then add the second number. 6 X 16 is 96, plus B (11) and youv'e got 107.
So, with 8 wires you can transmit anywhere from 00000000 to 11111111 witch in hex is 00 to FF or in decimal 0 to 255.

Bits and Bytes:
A "0" or a "1" Is called a bit. 01101011 or "6B" is a byte. your 60 Gigabyte Hard drive stores roughly 61,440,000,000 bytes. If one bit within those bytes is a 1 insted of a 0 or vice versa, your computer may mal-function. (Im not kidding!)

Knowing these numbering systems is the key to understanding how computers really work.

Now. When it all comes down to it, all we care about within the whole computer is the first half of the PROM. This is where the settings are stored. These settings can be modified to tune the car, but it is at your own risk. YOU CAN BLOW YOUR CAR UP! If you do something really wrong. The PROM chip is small by todays standards, it only stores about 4,000 - 32,000 bytes. You will be looking at your PROM chip in hexidecimal, you will never get to see it in just 0's and 1's.
Settings are stored in four different ways. Flags, Values, 2D (curves), 3D (Charts or maps).
Flags:
A flag is an on or off setting, a 1 or a 0.
Since you programing in hex, The Flags are grouped into 8. You may see 6B, but really it turns on or off 8 different things. So really, Flags are binary. You have to disect them using the chart. If you want to turn one item within the byte from a 0 to a 1, then your new Hex number could be fairly different. Lets say you wanted to turn the third bit of 6B (01101011) to a 0, then your new hex number would be 4B (01001011)
Values:
These are the simplest things to understand, basically, its anything that is set at 00-FF (0-255) lets say you want the fan to turn on when the temperature value gets to 195. you would change the value to 195 or "C3" . Simple really!
2D (curves or Arrays):
These are a series of values that form a curved line if you were to plot them out:
1E 2B 3A 48 57 66 78 89 would form a curve, but it doesn't really have to.
3D (charts or maps)
these are basically a series of arrays in sequence. They look like a ski slope when plotted in 3D.
These are used for the main spark control, And volumetric efficiency amongst other things.

To do tuning of your ECM, there is an assortment of "must have's" before you even contemplate attempting this.
First, you have to know everything about engines, how they work, what spark knock is, fuel pulse width, timing advance, Air to fuel ratios...
You need tools. At the very least, you need a prom burner, a prom eraser, a scan tool, software and some chips to burn. You would be much more sucsessful if you have a Wide band OČ sensor, a laptop, Data logging, and a PROM emulator. These things cost $$$.
Your software is VERY important when it comes to tuning. You can extract a Bin file (the information stored on the PROM) from the PROM and have a mess of useless numbers and letters, but with the software, you can make heads and tails of the information easily, and proceed to modify as necessary. Decoding the PROM chip is an extreemly difficult task and you probably don't want to go there. Thats why you spend the money on the software. The most popular editors I know of are Winbin, and Tunercat. Tunercat is supported, you get the definition files directly from them, they cost money though, but thats better then trying to make the definition files yourself.

If you find any mis information let me know. Feel free to add your knowlege and experiences as you so desire. I'm hoping that by sharing what I know about tuning and listening to what others know, I will learn how to go about taking on my specific tuning challanges and help others take on thers.

I'm not a tuning expert, I just know how to do it. There are others here who are far more knowledgable than I am.

quote You will be looking at your PROM chip in hexidecimal, you will never get to see it in just 0's and 1's.

Depending on the software you have you can look at the data in binary or hex. Needham's EMP software even lets you directly edit in binary which makes setting a single flag much easier. The only time you would need to do that is if you do not have a definition file for your ECM, and then you woudn't know where to go to change the flag unless you are doing your own hacking, and that's way beyond the scope of any discussion you'll find on PFF. The average Joe Schmoe who is just getting started will have the software and will be looking at real world decimal numbers that he understands.

For anyone who thinks they might like to get started in tuning their own car, a shareware version of GME Pro is available for download from my website. It includes the definition files for the 85 and 86-88 V6 Fieros. A full version and definition files for just about any GM vehicle through '95 are available by contacting the creator via the email link on the same page.

http://www.fieroaddiction.com/EPROM.html

Jon

Gotta grab that!

Latley I have been messing around with my 3.4L Pushrod motor. Im trying to get it running good. So far, It has 3.4 DOHC injectors, a 1227730 computer running a modified AUAF $8F TGP Code with code 24 shut off. It has DIS ignition and a knock sensor. I have no provision for an EGR either. So far so good, it runs ok. Idles alittle rich though. Whenever Im driving with a steady throttle, the AFR is fine. But as soon as I give it alittle more accelerator, the AFR dives hard (goes lean) and the car will loose power, and sometimes even buck. But within 2-3 seconds the AFR returns to normal, but If I give it more accel, the engine will go lean again, unless I pass threshold for Power Enrichment, witch it will then go a bit rich (perfect). It does this no matter what the Temperature the engine is, or whatever RPM's its at. Could this have something to do with the EGR not being there? Also, what table controls Idle AFR's?

I am setting my 3.4 PR engine up with this computer so that I can learn to tune the ECM to the motor. Down the road, I will swap in the 3.4 DOHC with the supercharger. It will be easier to tune with what I will learn with the 3.4 PR.

What software can I use to monitor my ECM? I have a 1227730? I don't think I can use WinADAL? I am having similar problems as you are having with the accelation. It seems to have learned each of the different RPM levels by driving at lest say 2500 RPM for 2 min. then 2100 RPM for 2 min. But it will then shurge when going to 2300 RPM. Need to see what is happing at these different RPM ranges. I would appreaciate any help in this area.

Wow, Fierobsessed, that huge but of information almost overloaded my brain! No offense but the way you explained it, although very informative and correct, is way overly complicated even for me at this late hour! Let me see if I can confuse you .

I am with rockcrawl in that I don't profess to be a tuning expert by any means, but I do the best that I can using the tools I have available and the knowledge I have accumulated both thru personal experience and training. With that being said, allow me to add a bit of my brainpan on this subject. Most of the better tuning programs available today, some of which are free and some of which cost money, basically do all the hard work for you, so long as there is a definition file available for your platform. Personally, I use tunercat (www.tunercat.com) which I find to have a good balance between user friendliness and tweakability.

Now let me address your situation specifically. You posted that you have some lean-out problems with your 1227730 system. I have the same computer on my 87 GTA 350TPI car, but it is a slightly different code ($8D). I had this same lean-out problem with my car because I have aftermarket heads, intake, exhaust, and a fairly large cam. What I have found with this computer is there is an issue with BLM cell placement settings on the stock chip. What was happening is for some reason the decel BLM cells were ending up at the same cells as part throttle accelleration. The 7730 determines BLM cell limits by RPM and MAP or load value. These cell limitations are adjustible on the chip so long as your tuning program is set up to adjust them. Well since I am running such a large cam, my MAP pressure during part throttle accel and decel were both about 60kpa. Well it just so happened my BLM cell border limit was set so high that the ECM was having to relearn the BLM every time I went from idle to accel and decel back to idle. Hence the drawback of speed density compared to a MAF based system. (MAF systems generally don't have this problem because they measure air flow directly)

Let me explain this another way. I could tune the main VE table so I would get a 124-132 INT and BLM count readings at idle and part throttle. Well what was happening was when I decellerated, the computer was learning new INT and BLM settings below the 120's (leaning out) because on decel the engine needs to run leaner than normal due to the reduced VE (lower amount of vacuum) of my larger cam at lower RPMs. Well the problem I had was when I went to accelerate again my RPMs would come back up thru that same BLM cell that just got leaned out during the last decel and therefore the engine would temporarily go lean and it would take a second for the computer to fatten the fuel back up thus relearning the correct fuel delivery (INT and BLM at 128). If I widen my decel fuel cutoff window, the BLM would not be affected but there was a problem. Decel fuel cutoff cannot be set this low, at least on my car otherwise it will have a tendancy to stall due to the loose torque coverter and quickly falling rpms. The reason why stock engines don't seem to exhibit this same problem is because stock cams have such a tight overlap that they will produce more vacuum on decel over an aftermarket performance cams. Therefore, the BLM cell will be different cell during decel compared to accel.

Well looking at the scan tool, I was noticing that during a decel when my RPMs were falling back thru a certain RPM range, the computer was in the same BLM cell as during a low-accel thru that same RPM range due to similar manifold pressure readings. Well my tuning software does not have a table for closed throttle VE, nor, looking at a disassembly of the $8D code, am I able to find one. There could be a table for closed throttle AFR vs. RPM but I haven't spent much time looking for it. My solution to the problem was to put my BLM cell boundry right where my accel and decel RPM range is where I had the problem. What that did for me was to force the ECM to constantly switch back and forth between the BLM cells thus no major change had a chance to happen to any one BLM cell in the short time of the decel. Now there are other ways to work around this. You can change your BLM update rate to a slower setting, or you can extend your decel fuel cutoff parameters to cover the RPM range where you are experiencing issues. You could also pour over a disassembly of the $8F code and see if you can find anything related to closed throttle AFR or VE; or the BLM update rate or cell boundries as I described earlier.

I hope that helps you with your issue. The way I explained it makes it sound like the stock computer is more overly complicated than what it needs to be but you have to remember these systems were set up to work on stock engines with small cams and there had to be some kind of fuel curve learning ability built into them for emissions compliance. This probably wouldn't even be an issue on an aftermarket ECU but then again most aftermarket ECUs don't really care about fuel curve learning and decel enleaning because they are not oriented towards emissions control. So what do they care if you are running too rich during decel anyway? Its not something that was intended to be built into those aftermarket systems that were designed for "racing" use.

------------------
power corrupts. absolute power corrupts absolutely.

Custom Chip Burning | Fiero Engine Conversions | Turbocharging | www.gmtuners.com

[This message has been edited by Darth Fiero (edited 09-01-2004).]

That is the reason I like the GM ECM's, they have SO many overly complicated parameters. In other words, you can adjust almost any little detail you want about what your engine is doing. Im just glad someone did a good dissassembly of the $8F code. I feel that it is the most useful code out there. It encompasses two major types of EGR's, has wastegate control, Knock sensor, DIS, OBDI, Lean cruise compatible, batch fire and is Speed density. It could be tweaked to control anything from a turbo 4 cylender to a supercharged Northstar, (if anyone can handle the re-programming for that one!) Come to think of it, there is almost no engine that isn't compatible with this setup (if you have a crank trigger, or aftermarket crank trigger and sensor)

I do see what you mean about the BLM cells being used for two different purposes and how that doesn't work well with a modified engine where the two purposes requirements are no longer the same.
I think I have a different problem, but I'm sure that it has alot to do with BLM cells too. I must get a WB OČ and a Scantool. Because right now, I have a A/F guage, while it's hardly useful for tuning an engine, it's pretty clear what the problem is when the A/F goes off the guage lean and the car looses serious power. (I bet its KR'ing to no end too, further killing the engine)