Notices
944 Turbo and Turbo-S Forum 1982-1991
Sponsored by:
Sponsored by: Clore Automotive

86 944turbo DME/KLR question

Thread Tools
 
Search this Thread
 
Old 03-28-2002, 03:39 PM
  #1  
2zeroes
Racer
Thread Starter
 
2zeroes's Avatar
 
Join Date: Jan 2002
Posts: 308
Received 2 Likes on 2 Posts
Post 86 944turbo DME/KLR question

How many pins does the DME and the KLR chips have for the 86 951? Anyone know what year they switched computers? Thanks!
Old 03-28-2002, 04:52 PM
  #2  
Danno
Race Director
 
Danno's Avatar
 
Join Date: Jul 2001
Location: Santa Barbara, CA
Posts: 14,075
Likes: 0
Received 2 Likes on 2 Posts
Post

The 4k 24-pin chips were used in '86-87. The '88+ computers used 8k 28-pin chips. However, the extra address lines and extra data storage weren't used for anything. In fact, F.R. Wilk actually stores 2 sets of data on his chips and you can flip between stock and aftermarket with a toggle.
Old 03-28-2002, 06:04 PM
  #3  
2zeroes
Racer
Thread Starter
 
2zeroes's Avatar
 
Join Date: Jan 2002
Posts: 308
Received 2 Likes on 2 Posts
Post

Hey Danno:

So 4k is the size of the chip? And both the DME and KLR chips are 24 pin-chips?

Now on the 89 951, the KLR is a 24-pin chip and the DME is a 28-pin chip. Is that correct?

What I am trying to do is copy the program that is on a 86 chip and transfer it to the 89 chipset. So the blank chips have to be 8k 24-pin chip for the KLR and 8k 28-pin chip? This can be done, right, as long as you have the necessary equipment and software?

Muchas gracias...
Old 03-28-2002, 06:49 PM
  #4  
Danno
Race Director
 
Danno's Avatar
 
Join Date: Jul 2001
Location: Santa Barbara, CA
Posts: 14,075
Likes: 0
Received 2 Likes on 2 Posts
Post

Yeah you can do that. But the data from an '86 won't be what an '89 needs to run optimally.
Old 03-28-2002, 07:07 PM
  #5  
2zeroes
Racer
Thread Starter
 
2zeroes's Avatar
 
Join Date: Jan 2002
Posts: 308
Received 2 Likes on 2 Posts
Post

These chips and the early computers are plugged in an 89, if that makes any sense...I just want to use this program without having to buy the old computers for my car.

If I were to read the program on a chip, would there be any mention of who wrote it...for example if I were to read an Autothority chip, would there name be somewhere in the program?

Thanks again Danno, you're a pal
Old 03-28-2002, 07:09 PM
  #6  
Konstantin
Addict
Lifetime Rennlist
Member
 
Konstantin's Avatar
 
Join Date: Jul 2001
Location: Germany/Braunschweig
Posts: 1,937
Likes: 0
Received 0 Likes on 0 Posts
Post

[QUOTE]Originally posted by 2zeroes:
[QB]Hey Danno:

So the blank chips have to be 8k 24-pin chip for the KLR [QUOTE]

you said that the 24-pin chip is a 4K how can you make a 8k 24-pin chip?

Konstantin
Old 03-28-2002, 09:10 PM
  #7  
Danno
Race Director
 
Danno's Avatar
 
Join Date: Jul 2001
Location: Santa Barbara, CA
Posts: 14,075
Likes: 0
Received 2 Likes on 2 Posts
Post

"If I were to read the program on a chip, would there be any mention of who wrote it...for example if I were to read an Autothority chip, would there name be somewhere in the program?"

Well, you'd have to disassemble the code (reverse of compiling). Don't know of very many Intel disassemblers laying around (they're vintage mid-'70s chips). I guess there are still vendors out there selling the DME programming softwares since BMW & Volvo used similar DMEs with different chips.

Best bet is to just read the raw data from the '89 chip and dump it onto a 24-pin EPROM.

"you said that the 24-pin chip is a 4K how can you make a 8k 24-pin chip?"

Half of the data on the 8k 28-pin chip is the data, the rest is the program code.
Old 03-29-2002, 01:55 AM
  #8  
TT
Racer
 
TT's Avatar
 
Join Date: Jun 2001
Location: Huntingtown, MD
Posts: 363
Likes: 0
Received 0 Likes on 0 Posts
Post

This info applies to the '86 and later 944 and 951 only (944S and 944S2 have a totally different DME).

The code used in the 944/951 DMEs is 8kB total. The microprocessor in all of them is an 8751 with a 4kB internal ROM. For the 24-pin models, the high part of the 4kB is in the 24-pin external socketed ROM. For the 28-pin models, the 8751 is jumpered to disable the internal 4kB ROM and all running code is contained in the external 8kB ROM. Converting a 24-pin to 28-pin model is possible by moving a soldered jumper on the processor board from one position to another, removing the 24-pin socket, and soldering in a 28-pin socket. I have successfully done this using a 24-pin DME from an 86 944 that is now controlling an 88 944 engine with the 28-pin ROM.

I checked to see if the 944 and 951 DMEs were interchangable, but the 944 DMEs lack the components that communicate with the KLR (the traces and layout are there, the actual components are not).

You can use larger memory parts because manufacturers added address lines in the larger parts using previously unused lines in the older smaller parts (even to the point of using a 28-pin device in a 24-pin socket with a board such as F.R. Wilk sells).

The upper 4kB of ROM contains the constants and maps along with a jump table that allows manufacturers to change certain operations in the 24-pin DMEs even though the internal 4kB ROM in the micro is untouchable. To successfully transfer information from a 24-pin DME ROM to a 28-pin DME ROM would require transfer of only the constants and maps along with pointer adjustments if any maps were moved, not the code in that area (3 part throttle fuel inj maps were moved when comparing the 86 and 88 944 DME). To go from a 28-pin to 24-pin, it is easier to just convert the DME from 24-pin to 28-pin and use the 28-pin ROM.

[quote]If I were to read the program on a chip, would there be any mention of who wrote it...for example if I were to read an Autothority chip, would there name be somewhere in the program?
<hr></blockquote>

The aftermarket chip in my 951 is a 28-pin 32kB device with the code duplicated 4 times. This aftermarket chip has an identification string of "(C)1992 DINAN". The KLR aftermarket chip has "TC-8220 KLR/COPYRIGHT (C)1992 DINAN ENGINEERING FOR AUTOMOTION". Not sure what they were copyrighting since they were modifying code that is more likely copyrighted by Bosch.
Old 03-29-2002, 04:49 PM
  #9  
2zeroes
Racer
Thread Starter
 
2zeroes's Avatar
 
Join Date: Jan 2002
Posts: 308
Received 2 Likes on 2 Posts
Smile

Thanks for the additional info TT! Would it be alright to contact you via email if I have further questions? Thanks again...
Old 03-29-2002, 09:07 PM
  #10  
TT
Racer
 
TT's Avatar
 
Join Date: Jun 2001
Location: Huntingtown, MD
Posts: 363
Likes: 0
Received 0 Likes on 0 Posts
Post

Sure no problem.
Old 03-29-2002, 11:10 PM
  #11  
Rob
Addict
Rennlist Member

 
Rob's Avatar
 
Join Date: May 2001
Location: Melbourne, Florida
Posts: 942
Likes: 0
Received 0 Likes on 0 Posts
Post

Hey TT,

Would you mind giving some details on what you use to disassemble/assemble the binary codes that are on the chips. I haven't looked at any yet, but was planning to borrow a reader and check some out.

Thanks,
Rob
Old 03-30-2002, 12:03 AM
  #12  
TT
Racer
 
TT's Avatar
 
Join Date: Jun 2001
Location: Huntingtown, MD
Posts: 363
Likes: 0
Received 0 Likes on 0 Posts
Post

I personally like IDA Pro, they have a free version available. Disassembling is the easy part, next comes reverse engineering. I am close to completing my analysis of the code. What is amazing is out of the 8kB, the 944/951 only use about 2kB of actual executable code and 2.7kB for constants and maps. I have written a working version of the DME code that reduces the executable size down to 1.6kB.



Quick Reply: 86 944turbo DME/KLR question



All times are GMT -3. The time now is 09:59 PM.