View Single Post
  #717  
Old 05-04-2008, 01:52 AM
b3lha's Avatar
b3lha b3lha is offline
Phil & Belha
 
Join Date: Aug 2001
Location: Alcyone Limited, Buckinghamshire UK
Posts: 2,671
Smile

Quote:
Originally Posted by b3lha View Post
However, there seems to be a bug in my software changes. The power mode is stuck on permanently. So I have to do some debugging and re-burn the eprom. But that shouldn't be too hard to do.
I fixed the bug in the TCU, but I'm not really sure what was wrong:

Testing with my tcuscan program revealed that the mode switching was working as intended, but for some reason the power light was stuck on whichever mode the TCU was in.

I made a dump of the TCU rom through the select monitor interface and verified that it was the same as the file I had written to the eprom. Then I reloaded the eprom with the original unmodified UK image and it worked properly. That proves there is no hardware issue.

I made two changes and one or both of them together evidently fixed the problem.

Firstly, the eprom is 64K, but the TCU only uses the last 16K of that. If you envisage the chip as four 16K blocks, then it looks like "blank, blank, blank, data". But when you read it through the select monitor it reads as "blank,blank,data,data", even on a factory TCU. This suggests to me that the A14 address line, that selects between blocks 3 and 4 is not connected, so you always get block4 when you ask for block3 or block4. Straightforward enough. Just as an experiment, I copied the data into all 4 blocks, this shouldn't really make a different because the TCU only looks at block 4.

The other thing I did is to set the checksum correctly. I didn't think it mattered because the self-test routine seems to be dead code. AFAICT it is never executed. Even if it was executed, it doesn't seem to do anything different whether the checksum is correct or incorrect. Either that or there is something screwy with my disassembler.

Anyway, two changes that shouldn't logically have made any difference seem to have made the problem go away. It is now switching modes correctly on my car, but obviously I can't test drive it because of the speed sensor incompatibility between JDM and Euro gearboxes.

Joe, I'll mail your ECU and TCU back to you first thing on Tuesday morning.

The new TCU functionality is as follows:

1) The economy shift map is disabled.
2) When the economy switch is on, the TCU will use the normal shift map and kickdown will give you power mode.
3) When the economy switch is off, the TCU will use the power shift map.

Everything else is same as factory.

Please test it out and post your results. I'm confident that it will work, but don't panic if it doesn't. I'm going to be away for a week from Friday so I'll catch up with you when I return.
__________________
Subaru ECU and TCU Website
1992 Alcyone SVX Version L
1992 Alcyone SVX Version L
1994 Alcyone SVX S40-II
2004 Subaru Legacy 2.5 SE Sports Tourer
1996 Subaru Legacy 2.2 GX Wagon
1988 Subaru Justy J12 SL-II

Last edited by b3lha; 05-04-2008 at 02:02 AM.
Reply With Quote