Newb OBDI question for OBDII Vet... LOL

[3400-95-Modified]

How in the hell do you determine which definition you need for your ECU... That's my first question. The next is the GM OBDI datastream definitions, once again, how do you know what you need?

I'm asking for two instances... my 1990 Beretta Indy will need to be tuned when I start tweaking so I need to know what to expect. Prom is AUPC 2004, The other instance is a friend who is trying to get a chip burned for his 1990 Beretta GT and every one that's been made wont even run the car... I'm wondering if the wrong definition is being used when burning which may or may not be changing the pin out functions.

[Guest]

To find the correct definition, or at least the mask, you look in a certain byte or pair of bytes in the binary file that is read from the PROM. Most of the bins will have that information at location $08. From that you find a definition file (or set of definition files, since datalogging requires a second file for the datastream specifically) that has been made for that mask, for the software you are using.

Many times people have already figured this out and you can look up the BCC, AUPC in your case and see what that matches up to.

The definition has no bearing on the programming of the E/EPROM, it's ONLY used for modifying, or at least looking at the binary file. Programming is done by directly writing the binary to the E/EPROM. Depending on which programmer you use there are different steps needed, but in no case does the definition file become a factor.

It sounds like either the E/EPROM is being installed incorrectly, or they are being programmed with the wrong bin.

Check out junkyard-efi.com for more specifics about tuning and getting started.

[3400-95-Modified]

I've tried looking up the AUPC but that has led me nowhere and even on large ECU application lists I can't find that as a listed Prom. The individual who is having the chip issue has sent a known working chip to another individual and every copy he has made will not start the car. It also has no sensor data when he attempts to connect to it. But the source of the "copy" starts the car. It doesn't have a perfect tune on it and that's what he was trying to get on the different chip. I wasn't sure if running a $A1 mask on a 90 car with a digi dash would cause issues or not. All I know is he's not putting the chip in wrong but something is happening to the file after it's being saved or re-burned to the new chips.

[neophile_17]

Is the person burning the chip familiar enough to know the right offset to read and write to if using EEPROM?

[3400-95-Modified]

I would think so.. I have no clue. He's made his own chips for his 92 Beretta which is why I ask if the 90 eprom would need to be a different mask or something of that sorts.

[padgett]

First off just saying '90 Beretta GT does not tell me anything, need to know which engine and transmission it has.

Second AUPC sounds like the right year ('90) but not what it is for. First place I look is here but it is not complete.

If not there then I turn to a factory P&A Grp 3.670. Unfortunately I only have Pontiac W and Buick E from that period

The engine family is the ninth byte of the PROM or a .bin file(starts with zero as the first so is x08 (or 08h or $08 for hexadecimal) e.g.
000000 05 A1 58 C2 0F 51 2F 04 DF C0 0C 0C 0B 61 43 01 engine family ($DF = LQ1)

$A1 is for a 1991-1993 LH0 (FWD) family but includes all variations and transmissions. The ALDL datastream should be the same for all. The sensors and error will vary depending on application.

So best to start with the year, engine, and transmission and work from there.

Also from the question I suggest you start with background on the GM P4 computers and what is needed to extract and program. This is a good place to begin.

[3400-95-Modified]

The configuration that is not working is in a 1990 Beretta GT, It currently has a 3400 V6 and 3t40 transmission. This car is also equipped with a Digital Dash display. The individual who is burning these chips has a 91 Beretta with a 3400 and 5-speed in it.

I know very little on what he is doing and I do not have the hardware to perform Chip reading and writing yet. I was trying to figure out what he could be doing wrong during the read/write process that would take a working chip and make a dud duplicate. He was starting to think the data he was modifying was the reason it wasn't working and that's why he just tried to make a duplicate of the chip he was starting with, the original chip he's using as the source will start the car but the tune is not ideal for it's configuration.

My situation is different, I have a 1990 Beretta Indy, 3.1L MPFI 3t40 digital dash. This car starts and runs perfectly fine since it's all stock right now. In the future I will be modifying this car and I'd like to know what bin I should be starting with and what definitions I will need to do that. The only info I have to go off of right now is the chip which has AUPC on it and that is not coming up anywhere that I look. The ECU is a ACDelco variant/reman unit, PN 88999196.

To be 100% honest its like the 90 L-Car 3.1 MPFI just simply doesn't exist in the tuning world...

[padgett]

Since I have a Pocket Programmer III (USB) from xtronics with a Memcal adapter the first thing I would do would be to download the AUPC contents into a .bin file and analyze that.

BTW a quick online look found this: p/n 16190177 EPROM,ECM(CODE 2699ALXB,6656ATTA,0291AUBJ,2004AUPC) L(LH0,MD9,NA5)
1990 Chevrolet Beretta 6 Cyl 3.1L MFI; 3-Speed Automatic Transmission, THM125C/3T40

Just a matter of knowing what question to ask.

Now I did find the AUPC on this site: look here and you can download the bin. No idea why it is not in the download list.
But 000000 03 37 01 04 04 15 D1 A0 88 C0 08 89 0B 61 3C 01 indicates you want an $88 file and not an $A1.

$88 1227730 or 16197128 90 - 92 F CAR LH0 V6 90 - 94 Cavalier/Sunbird LH0 27C .

Personally I'd pull your Memcal and see if the same but suspect this is why it did not run

[neophile_17]

It's been awhile but I remember having to use a different offset to read a stock memcal than to read or burn one of the SST 27SF512. This is because the SST 27SF512 has much more memory and you need to select the proper sector. That may be why you can't copy a chip.

[3400-95-Modified]

Well I'm getting more information as I ask for this individual, apparently his car is a 89 GT V6 auto the one that he can't get to run with this modded chip... And I am not the one trying to make these copies... I don't have chip burning hardware/software yet.

We will all be getting together on the 23rd for BerettaFest so we'll look at it then and see what we can come up with.

[padgett]

Easy: a 27C256 .bin is 32k (32x. -512 holds 64k. So you can just load the same program twice, once at 0000h and again at 8000h and it does not matter which side the ECM picks up.

Or you can load two different programs with different IDs and run a 5v switch to the MS address line. On is one program, off is the other.

ps why the 3.4 is not running depends on the year and code of the engine. Different years had different configurations of MAP vs MAF or both and whether or not it has an EGR or single vs dual ring crank sensor. With the wrong program it could be easily confused and carry on cranky.