eulogious

Ok, so after I installed the new modded TCU, it started acting up. It would basically reboot itself. I would be driving down the road and all of a sudden the power light and the atf temp light would come on and the car was in 3rd gear locked in 4wd. The TCU had decided it didn't want to work anymore. So after looking at the board some more, and with the help of presslab, I was able to determine that all the electrolytic caps where bad and needed replacing. So that's where I am at with that TCU. Now luckily for me I happen to find yet another TCU with that same part number, this time out of a car that was t-boned, so at least this TCU was functioning before the accident. It arrived in the mail on Saturday, thanks Uberoo!, so I was able to start playing with it. I plugged it in just to see what would happen, and it worked normally! Yay! It worked without any modification! So that meant that this TCU was already in better condition than my previous one. I then removed the solder from the holes to prepare it for a DIP socket. I unfortunately was unable to find a DIP 28 socket, but I got a DIP 40 socket and just trimmed it down to 28 and it that worked out good. I then left it in the stock configuration and test drove it around for two days and about 70 miles just to make sure that it was working before I tired to modify it. After putting it through the crucible, I decided it was time, all or nothing! So after practicing lifting the pins of the stock rom on my other TCU, I set out for some experiments. First I tried running a 10k resistor to pin 22 (OE Output Enable) from 5v, just to make sure that the OE was turned off. Tried it in my car, power light and atf light, so no go. I then lifted the CE (Chip Enable) pin 20, and the OE pin 20, and the same result, double lights. I then lifted the power and the ground for the chip to completely disable the chip, and I still got the double lights. So it was time for this stupid stock ROM to come off! So 30 min later I had the chip removed, I went down to my car, started it up, and the atf light went out, and then shortly after the power light went out as well! This is good. Turned it off tried again hoping for no errors and I got no errors! It seemed to work! After taking it for a test drive for about 20 miles all seems well. It shifts better, and feels more fun to drive. I have not tried out the paddle shifters, but its only running my modded ROM, so i should work fine. I have some caps on order for my other TCU, so when then come in I am going to try to figure out why I HAD to remove the stock ROM in order for my other ROM to work, so that should be fun. I want to be able to switch between the two ROMs, but really that's just to say I did I won't let this stupid thing beat me! This week I will have some time to be able to start running those wires, I have been busy trying to get this TCU to work, and then I will start building the paddle shifter assembly itself. I hope this all starts going smoother from this point forward thats for sure!

Hello all, To start, a shout out to presslab, he really helped me worked through alot of my problems, thanks dude! A while ago I was browsing through this forum, and I came across this awesome thread about building paddle shifters for the 4EAT. Here's the link: http://www.ultimatesubaru.org/forum/showthread.php?t=93670 This was my inspiration. So about the beginning of February I PM'ed presslab to see what really needed to be done and if this was really possible for me to do. I guess I should also give a little background on myself before going any farther. Right now I am training to become a IBM Mainframe systems programmer, I have modded xboxes and other SOC (system on a chip) devices including routers, and I have a good background working with computer systems in general. So this whole car thing is new to me, but not electronics. So with that said... After talking with presslab, I came up with a list of things that needed to happen so that I could get paddle shifters in my loyale. Here's the list: 1. Get TCU 2. Build simple cable to interface with TCU 3. Download original TCU ROM using the cable that I just built and the linux software 4. Compare original TCU ROM Code to new modified ROM code to check for differences to make sure the ROM is compatible and modify any settings 5. Install DIP 28 socket onto TCU PCB 6. Burn new ROM image onto new EEPROM 7. Install new EEPROM into the new DIP socket on the TCU, and disable the original ROM 8. Splice in new plugs into exsisting TCU harness for the new TCU, but keeping the old as a backup 9. Modify the RPM signal going to the new TCU via the new harness. Use a 1k resistor and place it inline from the ECU to the TCU. On the TCU input install a 4.7V zener diode after the resistor with the cathode on the TCU input and the anode to ground. 10. Build the DCCD switch using a six position rotary switch and wire in some 1N4001 diodes in a series between each pole. The FWD switch pole is attached to ground and the common is attached to the atmospheric sensor input. 11. Build the paddle shifter assembly and install it With such a short list, it should take me no time to do it all Needless to say I have spent ALOT of hours just researching and trying to figure out how all this works, and I just got it working today, a month and a half after I started. So this project isn't for the weak hearted, that's for sure! Personally I don't want to modify something in my car, especially anything having to do with the transmission, unless I honestly can understand what has been modified, how it was modified, and why it was modified. Once I can figure all that out, I will feel much more comfortable installing this into my car. So I set out learning assembly code for the Motorola 68CH11 processor. Lucky for me I am already somewhat comfortable in a low level machine language due to my training on the Mainframe, so picking up the assembly wasn't that hard. It wasn't easy, but not that hard. The hardest part is knowing how to read a reference manual Man that crap can put you to sleep! Anyways, once I had figured out the jist of what was going on in the ROM, I set out on getting a TCU that had the same part number has presslabs. Once I found that, I needed to get the harness for the TCU, since the TCU I bought didn't come with the wiring harness. Wiring harness in hand, I set out to wire it up. Here's my thread on the wiring harness: http://www.ultimatesubaru.org/forum/showthread.php?t=110131 I broke this whole process up so that each step I did was more easily searchable, and more organized. Also these posts have a bunch of pics in them, so beware! Read through that thread to see the problems I had with that and how I solved them. Once it was wired in, I then had another problem, the AWD didn't work. I had a fried transistor on the TCU board that I had to replace. Here's my thread on how I fixed that: http://www.ultimatesubaru.org/forum/showthread.php?t=110121 After all the hassles and trouble shooting, trails and tribulations, I was finally able to install the moddified ROM onto the EJ TCU and got it working! Granted I skip a few steps along the way... I had a buddy of mine burn the ROM image for me, so that was already done. I ordered the cable that I need to interface with the TCU, but I am still working on getting the tcuscan software to work for me :-\ I just gave up and moved on. I figured that if it didn't work with the new ROM, then I would have to figure out tcuscan. Since the modified ROM worked, I am not in a hurry to fool around with tcuscan. I will in the future, but for now, it can wait. With that, I have also not downloaded the stock ROM to verify against presslab's ROM, but since it his ROM worked in my TCU, I can assume that the ROM is the same. Here's some pics of adding the new socket on the board, and with the new ROM on the board: Before Modifiing: After: Here's some "action" shots, or something like that... Half done: You can see where I started, and how my soldering skills got better as I went First time I have ever tried doing anything like this. Good times All done: Socket installed: So in order to get the DIP socket to work, you have to move a resistor... You have to move it from 421 to 422. The "000" on the resistor means 0 ohms, so you can just use a piece of wire as well, which is what I ended up doing, since that resistor is a huge PITA to move. I am going to install a switch so that I can flip between the two ROM's whenever I want. This way it makes this mod a little more reversible. To a point anyways At this point since the TCU is functioning "normally" in my car, I am ready to start running more wires for the paddle shifters and dccd knob, as well as the wire for the select monitor, and to start building and adding all the goodies and getting it installed (5 additional wires total). I have the DCCD knob already here, and I have some ideas on the paddle shifter assembly that I should be able to test out here soon, and I should be able to start wiring it all in starting this weekend. My first plan is to install the DCCD knob and get that working, then the paddle shifters. I figure one thing at a time, that way I don't get to overwhelmed, so stay tuned!

Notice something different? I can officially say that I was correct in my previous post. The new ROM("SST" chip) is in and is working perfectly with no errors and no flashing power light :banana: Took him for a drive and everything worked flawlessly. After the drive and during I stopped and checked for power light flashes/errors and... nothing! Woohoo!! Now that it works "normally", time to start adding all the goodies!

Kinda like the motorcycle wave as well. I give the peace sign to passing bikers just to say "Sup!". I like this idea, I will start doing it in the oly area, so lookout my fellow Olyians!

Ok so I wasn't quite happy with my previous answer since I didn't know exactly what was failling so I dug a little bit deeper, and then I found this little handy chart saying what to look for under certain conditions. According to this chart, and it's from a 1995 FSM so things could be different, if front wheel slipage occurs, one of the things to check in VSS2, which would fall into line with what I was saying in my previous post. Since the TCU is recalculating VSS1, it's assuming that VSS1 is correct (it's not in this case, VSS2 is correct because it's not being recalculated, but the TCU doesn't know any better) so when the numbers are off, it defaults to saying the VSS2 is broke because it's giving a false speed reading (since it's different than VSS1), and then it puts the trans into FWD mode because of this. I think that answer all the questions on why my trans is acting the way it is and why the trans is putting itself into FWD mode. Here's the code that I found and if you follow the code for the VSS2 error, it leads to the same chunk of code that handle the FWD fuse, so I can back up this all up with the code. Of course this is assuming that I am reading the code correctly as well First chunk shows the VSS2 Error and FWD fuse functions and where to branch to handle those: E25E loc_E25E: ; CODE XREF: CalcVehicleSpeed+7j[B][COLOR="Red"]E25E brset *errorFlags, $20, loc_E27A ; ' ' ; VSS2 Error[/COLOR][/B][B][COLOR="red"]E262 brset *misc_Inputs, $80, loc_E2C7 ; '€' ; FWD Fuse[/COLOR][/B]E266 cmpb byte_C04BE269 bhi loc_E273E26B ldaa byte_C04EE26E aslaE26F staa byte_AF Here's section E27A (VSS2 Error): E27A loc_E27A: ; CODE XREF: CalcVehicleSpeed+CjE27A ; CalcVehicleSpeed:loc_E25EjE27A ldaa byte_17E27C brset *errorFlags, 4, loc_E2B0E280 ldab byte_C04EE283 stab byte_AFE285 cmpa byte_C04BE288 bhi loc_E2A1E28A cmpa byte_C04C[COLOR="Red"][B]E28D bcs loc_E2C7[/B][/COLOR]E28F ldab VSS2E291 cmpb byte_C04B[B][COLOR="red"]E294 bhi loc_E2C7[/COLOR][/B]E296 bclr *errorFlags, $10E299 bclr *errorFlags0, 2E29C bset *byte_2, 2[B][COLOR="red"]E29F bra loc_E2C7[/COLOR][/B] As you can see (highlighted in red), if it errors in anyway it branches to the FWD Fuse function... And section E2C7 (FWD Fuse): E2C7 loc_E2C7: ; CODE XREF: CalcVehicleSpeed+12jE2C7 ; CalcVehicleSpeed+21j ...E2C7 brclr *errorFlags, 2, loc_E2DFE2CB ldaa VSS2E2CD ldab VSS1E2CF brclr *byte_30, 8, loc_E2D8E2D3 bclr *errorFlags, 2E2D6 bra loc_E336 So at this point, the code is running the same branch instructions (to location E336), so it's the same for the VSS2 error, and the FWD fuse. So I think this gives us a solid answer as to why this problem is solved via software and why tire size and diff ratio actually matter when dealing with calculating the AWD. If I am wrong on this, someone please correct me!

Ok, so after reading what Chux said, I decided to dive into the code to try to figure out what the hell was going on. Here's the two parts of the code that calculate the 2 speed sensors, VSS1, and VSS2: Calc VSS1: D6ED CalcVSS1: ; CODE XREF: SPIE-1044PD6ED ; SPIE-FB5PD6ED ldd word_24D6EF std word_1BD6F1 ldx #gearRatioMapD6F4 ldab currentGearD6F6 aslbD6F7 abxD6F8 ldaa 1,xD6FA ldab word_1B+1D6FC mulD6FD tabD6FE clraD6FF std byte_BBD701 ldaa 1,xD703 ldab word_1BD705 mulD706 addd byte_BBD708 std byte_BBD70A ldaa 0,xD70C ldab word_1B+1D70E mulD70F addd byte_BBD711 std byte_BBD713 ldaa 0,xD715 ldab word_1BD717 mulD718 tbaD719 clrbD71A addd byte_BBD71C std inputShaftRPMD71E ldd word_26D720 lsldD721 lsldD722 std word_1B[SIZE="3"][COLOR="Red"][B]D724 ldaa VSS1toKPHconst ; (862 * 4.11) / (899 * 3.7) * 0xE2 = 0xEF[/B][/COLOR][/SIZE]D727 ldab word_1B+1D729 mulD72A tstbD72B bpl loc_D72ED72D incaD72ED72E loc_D72E: ; CODE XREF: CalcVSS1+3EjD72E tabD72F clraD730 std byte_BBD732 ldaa VSS1toKPHconstD735 ldab word_1BD737 mulD738 addd byte_BBD73A std byte_BBD73C lsldD73D bcs loc_D745D73F lsldD740 bcs loc_D745D742 lsldD743 bcc loc_D74AD745D745 loc_D745: ; CODE XREF: CalcVSS1+50jD745 ; CalcVSS1+53jD745 ldaa #$FFD747 bset *byte_30, 1D74AD74A loc_D74A: ; CODE XREF: CalcVSS1+56jD74A staa VSS1D74C rtsD74C ; End of function CalcVSS1D74C Calc VSS2: DFFB CalcVSS2: ; CODE XREF: SPIE-1041PDFFB ; SPIE-FB2PDFFB tst byte_AEDFFE bne loc_E006E000 clraE001 clrbE002 std word_1DE004 staa VSS2E006E006 loc_E006: ; CODE XREF: CalcVSS2+3jE006 brclr *byte_2F, 1, locret_E061E00A bclr *byte_2F, 1E00D tst byte_AEE010 beq loc_E05CE012 seiE013 ldaa byte_2DE015 staa byte_BCE017 ldd byte_2BE019 cliE01A ldx #2E01DE01D loc_E01D: ; CODE XREF: CalcVSS2+2AjE01D asl byte_BCE020 rolbE021 rolaE022 bcs loc_E029E024 dexE025 bne loc_E01DE027 bra loc_E02CE029E029 loc_E029: ; CODE XREF: CalcVSS2+27jE029 ldd #$FFFFE02CE02C loc_E02C: ; CODE XREF: CalcVSS2+2CjE02C std byte_BBE02E xgdxE02F cpx word_C012E032 bls locret_E061E034 ldd word_C012E037 fdivE038 stx word_1DE03A addd word_C014E03D ldx byte_BBE03F idivE040 xgdxE041 addd word_1DE043 std word_1DE045 ldd word_C016E048 ldx byte_BBE04A idivE04B xgdxE04C tstaE04D beq loc_E054E04F ldab #$FFE051 bset *byte_30, $10E054E054 loc_E054: ; CODE XREF: CalcVSS2+52jE054 addb VSS2E056 rorbE057 stab VSS2E059 bset *byte_30, 8E05CE05C loc_E05C: ; CODE XREF: CalcVSS2+15jE05C ldaa byte_C018E05F staa byte_AEE061E061 locret_E061: ; CODE XREF: CalcVSS2:loc_E006jE061 ; CalcVSS2+37jE061 rtsE061 ; End of function CalcVSS2 As you can see, VSS1 is recalculated, and VSS2 is untouched. I have highlighted in red the important part of VSS1, the part that recalculates VSS1 to read the same as VSS2. So what I found out is that the TCU has to take the input from VSS1 and then recalculates it so that it matches VSS2 in order to figure out if slip is occurring (DUH!). So it has nothing to do with the tire size directly, but how the TCU recalculates the VSS1 speed to make sure that it MATCHES what the VSS2 should be so that the two sensors are reading the same speed in order to calculate slip correctly. Now this makes total sense. So then if VSS1 and VSS2 are reading different numbers (they are reading two different numbers because the VSS1 is being recalculated for the EJ tire size and 4.11 rear diff, not the EA tire size and the 3.7 rear diff (I don't think that the rear diff matters, but it could), and VSS2 is not being touched so it's the actual speed the car is traveling, while VSS1 is recalculated based off what the tire size should be coming from the factory, and the factory rear diff ratio for the car) the AWD will not be able to calculate the amount of slip correctly, and then it faults, putting it into FWD mode. Basically it runs the calculations on both speed sensors and then compares VSS1 and VSS2, which should be the same, and if they are not, it throws out an error which then affects the AWD. The rest of the system is fine, since none of the other components require a direct comparison of the 2 speed senors in order to function properly, it's either one or the other. So I believe that's why I am having the problems that I am having, and also why changing that one variable in the software fixes everything right up. Does that all make sense? Chux: Thanks for the info man! I didn't know all of that before, so thanks for that!

I think you hit the nail on the head here with this. If I remember reading correctly the legacy speedo is controlled by the TCU? Or something to that affect. The TCU somehow influences the speedo in the legacy's, I believe. I think I remember reading something like that. I am too tired right now to research this, but I will tonight when I get up. But I think that this is where the TCU comes into play and why changing that one variable fixed it for presslab and hopefully me. Even though the TCU doesn't control my speedo, the sensor that does control it might have to be calibrated, and that's where the software comes into play. Do keep in mind that I am only going off of what someone else has done, and my reading of the code and understanding of the code. Presslab said that he changed one variable in the ROM when his trans was acting like mine, and then all was well, and he hasn't had a problem since. I have the chip in front of me with the moddified ROM burned on it already , the DIP socket, and the solder holes clear on the board, so it will be as simple as "throwing it in" to see if it works or not, then I will be able to tell for sure I am just trying to get my ducks in a row before I dive in, so one could say I have a pretty good handle on it... The important part was making sure the TCU actually worked and that also ended up being the hardest part :-\ I don't know much about OBD II at all so ya. I do know that in an OBD II trans there are more signals going to and from the TCU to the ECU, but not in mine. That's a Phase II 4EAT if I remember correctly... There is only 1 input from the ECU and that's for RPM on the Phase I, but I am sure you already knew that. So there are some differences, but I don't know enough about them to be able to say anything about it... Right now I am having trouble getting the tcuscan software to work, but I am to tired to care, so I will look at it some more soon. So once I can manage that and confirm that my wiring is correct via the tcuscan software (I think it is already, but I like to double check), then it's off to the fun part of the project

Congrats on making it home! It just seems like something was bound and determined to not let you get that car home! Glad it's fixed man.

Don't be fooled by the name TOD (tick of death), the tick is fine. They can tick all day and never hurt the motor, or so I have been told anyways... Mine does the same thing on startup, and I am not worried about it at all. Sometimes I will add some MMO (Marvel Mystery Oil) and it seems to help with sir ticks alot, but I don't add that alot of the time. I have notice that when my oil gets low it will tick more, but other than that i wouldn't really worry about it. I drive from Olympia to Seattle and back every weekend and have racked up about 10,000 miles since I got my tickin' subie and I have had no problems at all. You can replace the pump, but just for a little tick, I wouldn't fuss about it until something else comes up and you need to dive in a replace the oil pump since if you replace the pump you might as well do the water pump and the t-belts as well since you are in there. It's kinda a big job. As for seafoam, just do a search for it. There are tons of threads about how to use it. I use 10-40w chevron oil and don't plan on changing that anytime soon I don't really think there is a need to run synthetic, but to each their own. I think that just about answers your questions, good luck!

Sad times But I think you get the idea of where this is all heading, it was probably the same thing that you wanted to do... Ok, you got me. It's not truly JUST the rear diff. The TCU actually expects the tire/rim size to be a certain size and rotate at a certain revolutions per minute while the car is traveling at 60mph. So for the EJ cars, it expects E2 in hex, or 226rpm at 60mph. For the EA cars it needs EF, or 239rpm at 60mph. So it works out to be about 6% difference, just in the tire size, and that doesn't include the 3.7 rear diff which throws things off as well. So I think that is why you can switch trans' with the EJ series (SVX, Legacy, Impreza) with no problems and without switching the TCU, but when you try to move things across to the EA series, you run into a little more problems concerning tire/wheel size. The XT's might be different as well since they have 5 lug, but I really don't know, I haven't looked into that. But I think this is the main reason for the AWD not working. Now I also know that Gloyale swapped in a EJ trans into his EA car, but kept the EA TCU and all is well, so ya. I haven't looked at the code on the EA TCU to see the differences but obviously that works fine, but you are keeping the wheel size the same for the most part (EA wheels with EA TCU). It could also be that the code is completely different and more forgiving about the tire/wheel size as well, but I don't know since I have not seen the ROM code, so that is just speculation. It would make sense though because the XT's are 5 lug, so they probably just used the same TCU and made the code more forgiving, and then locked it down more with the legacy revision. Again speculating, but it would make sense. As it sits right now, when I start the car twice to check for errors, I get no errors, but as soon as I go for a drive, I get an error the next time I try to start it, so something is going on after the car starts moving, which falls in line with what presslab said his was doing before modifying that variable from E2 to EF. Once he changed that variable in the ROM, the AWD starting working again. The next step that I am going to do is to interface with the TCU via the alcyone.org linux software tcuscan while it's running and really see what's going on. I have the cable made already, I just need to get the actual TCU to function "properly" before I tried doing anything else to it, which I have finally succeeded in doing, so I will probably get to this in the morning. Once I get that done I will be able to tell what is really going on with the TCU and to see if all my sensors are working the way they should. Everything else in the car functioned properly on my test drive, just not the AWD. But the tcuscan software should be able to tell me if everything is good or not. At least I hope so

For those of you who are curious, I have created a thread showing how I made the harness to fit an EJ TCU into an EA car (4EAT). Here's the thread: http://www.ultimatesubaru.org/forum/showthread.php?p=910845#post910845 This was good times for sure :-\

So I just wanted to start out and give a shout out to presslab who has hooked me up with alot of info so that I could make this happen, thanks! So I got fed up with my front wheels spinning on wet days, so I decided that I wanted to upgrade my loyale 4EAT TCU to an EJ TCU. It's a new processor, it's faster, more "up to date", just better all around, so why not. So after hunting down a 1990 legacy AWD TCU, I went to my local JY and grab a harness, and then got ready to start splicing in wires. So to start off, I had this document of the EJ TCU pinout, which is a good starting point. I then updated it with info that I have gleaned from talking to other and what not to try to get all the pins labeled. I also had some wiring scans for the loyale 4EAT, so that I could match up the wires. Well, I actually have the FSM, so I have another document that I haven't scanned yet that made it WAY easier to create a pinout like the one above. First the 4EAT wiring diagram: Now here's the diagram like the EJ that I have created, this is looking down at the plug, after it's been unplugged. So it's looking straight at the plug, if that makes sense. I will get around to making the connection like the EJ, one the is for the TCU itself, but for now this works: Ok, so here's the pics of it all said and done: And then here's my work area :-\ Nice and roomy… So basically using the two documents above I sat down with masking tape and labeled the EJ TCU wires and then I got into the car and I labeled them as well. This way the wires were labeled and I didn't have to worry about the diagrams anymore. I kept the labels on to help troubleshoot down the road, and I am very glad I did. Made it much easier to try to find the problems I was having. The one soldering tip that I used to speed up the soldering process and create a more solid mechanical connection was to create a hook on the wires end after stripping the insulation off, and the hooking the two wire together. Does that makes sense? This made it easier to solder with one hand, and add strength to the solder joint itself. After that I then matched up the wires, spliced, soldered, heat shrinked and moved onto the next one. I took a while, since my back started to hurt, so I spread it out over 2 days to relieve some back pain. It's pretty cramped in there and there are about 30 wires to splice, so it took some time. I have also created a RPM clipper circuit to limit the voltage going to the RPM wire, just so that it falls in line with what the TCU is expecting. I have a 1992 legacy FSM and it shows the wire coming right on the coil for the carbed EJ22, but just to be safe, I will add in the clipper circuit. I have not done that now because of diagnosing problems up to this point. Here's the circuit diagram I made: So I now have an ej TCU harness spliced into my exsisting loyale harness, so that I can swap either TCU in, so I will never been down because of that. I also know that this works as well. I have gone on a 20 mile drive and the EJ TCU shifts smoother, better, and is just nicer all around driving, and it shifts just like it should and acts just like my other one. The one problem that I do have is that the AWD fails (doesn't go into torque bind mode, but doesn't work, FWD all the time) because it's expecting a 4.11 diff, and I have a 3.7, but they will be fix via software soon At first my AWD was locked into 4wd mode all the time because of a burned out duty c circuit in the TCU. Here's my thread on how I fixed that: http://www.ultimatesubaru.org/forum/showthread.php?t=110121 So in doing all this I have tripple checked my connections and I have verified all my documents to make sure they are ok. So ya, this is the first step in my latest project. More to come on that! I will also scan in some more docs that I used to make this easier when I get a chance, and I will update this thread with those. If you have any questions, feel free to ask and I will help were I can!

Thanks! The real reason I stuck with it is that I need this exact TCU part number for my project to even have a chance at working, so I kinda needed to attempt to fix it to make my life easier. Now I have some experience in fixing circuit boards as well This was my first attempt at repairing a circuit board, if you couldn't tell by my ripped out traces

Hello all, I am posting a link the the new gen forum post where I outline how to fix a fried tcu due to a bad duty c. I thought that I would post a link here, since some of the older cars (like mine!) have a 4EAT in them. Thanks all! http://ultimatesubaru.org/forum/showthread.php?t=110121 Feel free to ask me any questions, in the linked post of course

Hello All, So I am posting this in this in the New Gen forum, even though this is going into my loyale because there are more 4EAT in the new gen's than in the older gen. To start off I would like to thank presslab and Gloyale for answering my questions leading up this whole deal, thanks guys! So I have acquired a 1990 legacy AWD TCU so that I can adapt it into my 1990 loyale. But when I got the TCU and plugged it in I got instant torque bind, and I mean instant. It was apparent to me that the TCU wasn't going into limp mode, the hardware had to be bad somewhere. I was also getting the 16 power light flashes that mean that the TCU has a fault somewhere. So I took it for a drive, and man it was binding up like a no other in the corners, so the AWD was not working. I tried putting a fuse in the FWD fuse holder, nothing, still binding. I went and tripple checked my connections and verified all the voltages, and everything was fine except for the duty c circuit. I got nothing out of it. So after saying a few curse words and after contemplating pissing on the stupid TCU, I sat down and PM'ed a couple people to see if they could help me. Presslab suggested that I check the board's transistors, and to start by following the duty c solenoid pin and tracing the circuit. Well low and behold I found the culprit! Here's a diagram showing the TCU pinout, the Duty C solenoid is pin 3 on connector B33: Blown transistor: Duty C Solenoid Circuit Path: The red highlight is the path, and the orange circle is the blown transistor, just in case you all didn't catch that. As you can obviously see, this transistor was blown So my hopes shot up, I might have found the reason it wasn't working! So the next day I went to my local electronics wholesaler and picked up a new transistor and a solder sucker so that I could remove the old part. New transistor: Solder sucker: So after getting home and only spending $6 it was time to attempt to get this bad transistor out. So I would highly recommend getting a good soldering iron before you attempt this, heat will KILL the board, so be very careful. Use no more than a 30 watt iron. I didn't take pictures of me actually taking it out, but I will explain what I did. The first thing was to add a little bit of solder to the existing solder joints to heat it up so you can suck it out. You don't want to be hanging out too long heating up the solder, you could damage the board by getting it hot, so all this has to be fairly quick, which is a PITA. So heat up the exsisting solder and suck it out. Repeat for the other 3 holes. I had to repeat this process about 3 times on each hole to get most of the old solder out, so take your time, and repeat, it will come out. So once I had removed enough solder, I was able to wiggle the blown transistor out. My Soldering Iron, Sears $50 special but it's a Weller Here's the empty holes: Here's the blown transistor: The stuff circle in red is what you don't want to do I was very lucky that the contacts were still on the bottom, so I was able to salvage this, but THIS IS WHAT YOU DO NOT WANT TO DO! I rip the traces out. Very BAD. So take your time! I didn't get enough of the solder on the top of the solder joint, so when I removed the transistor, I tore out the top part of the traces. So don't do as I did on this one Man, I got lucky. Okay so now that it was out, it was time to get the new one in: Yellow circle shows the new transistor: The bottom of the board: Once that was done, all I had to do was add a little bit of solder and hope for the best! All soldered up on the bottom: Topside: As you can see I didn't get solder all the way up to the top of the board. This is due to not using any flux. If I would have dipped the end in flux, the solder would have flowed better. But this doesn't bother me, since I ripped the tops of the traces out anyways, it really doesn't matter So then it was time to test it. So I went and plugged it in, started the car, turned it off, and then started it again. I did this because the first time the TCU won't show any lights, it's only after you turn on the key and turn it back on does it flash. And no flashing lights It worked! So to verify I took it for a drive. No torque bind, it shift smoothly and went into lockup and all. Everything worked great! So in conclusion if your duty c solenoid fails and you fix it but the car still has the same symptoms, you might want to look at the TCU, since it can be blown by a bad duty c! So with a little patience and a steady hand you can fix it for about $1.50 and go on your marry way. I hope that this helps someone else out. If anyone has questions, just ask, I will help the best I can! This is just the first part of my project that I am doing, so I will be posting more in various forums sections as I get more done!

Man, I love lmgtfy!! Good times

Ya I basically have "factory" cruise now! At least most of the "factory" parts The original owner spent $550 to have the dealer install the cruise control back in 1990 when they bought it new, I have the original invoice For that much money, especially back then, I have no idea how it was acceptable to have your $550 cruise control system NOT function properly If I spent that much money to have cruise, you damn well better bet it will work the way EVERY other cruise system works, or I will raise hell. But not the PO, they just didn't care or something. Seriously lame :-\ It cost me about $40, a little bit of R&D, and time to fix the system. I just don't get some people I should do that. I will look into what it takes to get an article submitted and what not. Then I can take better pics as well Thanks for the suggestion!

So to sum up everything... I started out with having cruise control that "worked" but wouldn't hold speed up a hill, wouldn't accelerate, etc because I have a turbo motor and when I hit boost the engine no longer produces vacuum, therefore causing the vacuum operate cruise control not to work. So first I had to come up with a vacuum pump to put the cruise control on it's own vacuum subsystem. I then modified a factory vacuum pump by removing all the controls from the pump, just leaving a pressure switch, reserve tank, and pump. I then proceeded to wire in a relay to kick on to turn on the vacuum pump when the cruise was turned on. Then I had to add in the cruise control steering wheel controls. That was easy. Just splice away and you are good to go! So all said and done I am very happy with this setup. I have solved all my problems with the cruise and it's WAY better than it was when I first got it. I hope that this helps someone else out. If you have any questions about it, feel free to ask!

Ok, giving a final update with pics finally Lots of pics, you have been warned... So I got the ring connector I was talking about earlier, it happen to come with a new switch assembly that I found that had intermittent wipers on them, so I pulled the wheel, installed the new turn signal/headlight switch unit, found out how the dealer system tells the cruise what to do, wired in my new buttons, and BAM! I got intermittent wipers AND cruise controls on my steering wheel all with one stone Here's a the pinout for the dealer cruise control unit: Here's a couple photos of my modified vacuum pump assembly: Here's what the original looked like from a factory installed cruise system: Now, the dealer unit is a real simple unit and it was easy to figure out how it worked. Basically there is a small control unit for the cruise that is located where the power mirror control is normally located: Here it is out of the car: All this unit has is four wires going into it. One main power, and 3 others that turn it on/resume/accel, set/decel, and turn it off. All the switch does is apply a positive voltage to the correct wire to get it to do that function. Meaning that if I want to turn on the unit, all I have to do is apply 12v+ to the on/resume/accel wire, and it will turn on/resume/ or accel. Easy enough. Subaru also made it easy for me with the factory cruise buttons, they are wired with power (12v+) already! All I had to do was take the wires for the steering wheel buttons and splice them into the wire coming from the existing switch and bingo! Now my steering wheel controls work for the cruise Here is the connector I spliced into (I haven't spliced it at this point): And here it is installed in my car Here's where I moved stuff around in the engine compartment and where I spliced into the cruise harness in the engine bay to install my relay: Here's a pic of the dealer vacuum valve that is powered by engine vacuum: Here's a shot of the vac pump unit installed, the red wire is the power to the motor of the vac pump, it buried but you get the idea, it's locate in the stock location: Sorry some of the photos aren't so good, they were taken with an iPhone

I believe that there is some sort of amplifier in there and you have to find the inputs to that, like from the tape player. I bet that's why it didn't work just wiring it up to the speakers... Once you find those inputs to the internal amp, then it's just figuring out which input is left and which is right. The problem is finding those inputs. Trial and error with a DMM is really your best option. I would tend to lean that this is going to be a little difficult to achieve, since you don't have a wiring diagram for it. I wish I could help you out more, but that's all I got. Good luck, it will be interesting to see what you come up with!

Ok, scratch my first idea... After thinking about this I made sense of the system, and here's what I would do. It looks like in order to disable the ABS, you remove fuse FB-5 fuse 19. When you do that, the ABS unit then grounds the ABS warning light wire, turning on the warning light. So I would use a combination of a DPDT switch and a 40a bosch style relay. You will want the relay to handle the "fuse pulling" since a normal "heavy duty" switch can only handle up to 10a MAX, you want something that can stand the amps, and I bet you the ABS unit fuse is more than 10amps, and even if it was 10amps I would want something that could handle more just so I know it wouldn't fail and fry something in my ABS/brake system causing a massive failure while driving So I would use nothing but a relay for that, even if you don't disable the light. When messing with the brakes, one must be very careful and overdo it on everything to make sure nothing fails. The ABS light will pull very light amps on the other hand, so using a switch for that should be fine. So wire in a relay to cut the wire coming from FB-5 fuse 19, then wire in a DPDT switch: Circuit 1. "Common" to relay trigger wire (ground); "on" side grounded; "off" side open. Circuit 2. "Common" to ABS light; "on" side open; "off" side to ABS Unit ABS light output. This way when the switch is "on", the power is cut to the ABS just like removing the fuse, and the ABS light is disabled at the same time, but when you turn it off, the ABS is enabled, and the ABS light control is returned to the ABS unit, just like stock. That's what I would do. I don't think the TCU really cares about the ABS, so I wouldn't worry about that. Let me know what you think!

Sounds like an interesting idea... What type of trans do you have? I have a 1995 legacy FSM and looking at the TCM output for the auto, and it says that for the TCU to think that the ABS is ON, it needs to see less than 1v, so that means that if you ground that pin, it will think that abs is on, and then you can "pull the fuse" and the TCU will still think it has ABS, even though I don't know if there's any advantage to that or not. Something to think about. Here's the diagram for the TCU: I don't know what will happen if the TCU thinks the ABS is connected when it's not, but I don't think it will hurt it. Here are the FSM wiring diagrams for the ABS unit itself for the 1995 lego, just because I am feeling nice So I would use a DPDT switch, one circuit to disable the ABS by splicing into the power wire for the ABS light AFTER the light and grounding it and leaving the end going to the ABS Unit open effectively disabling it while leaving the light off, and the other circuit ground the TCU pin allowing the TCU to think the ABS is still working. That way one flick of a switch and the ABS is disabled, but no lights, and the TCU won't be all sad, flick it the other way and all is back to stock. This of course has not been tested and is just my thoughts looking at this, and I don't even know if splicing the wire after the ABS light would work since I don't have an ABS car, but that's what I would do to start. It's a little more work, but if you are going to go through all the trouble of installing a switch, you might as well do it right! I hope this helps!

If you do decide to remove the dash, you should also replace stuff while your in there, like the heater core, dash lights, etc. Anything that you might have to remove the dash again later to get to if it breaks. Might cost more now, but save a huge headache in the future. I have to replace my heater core, and I am not really looking forward to it at all :-\ If you want to know how to remove the dash, search for "heater core replacement" or something like that, there are a couple howto's for that and they tell you in detail how to remove everything. Miles did a great job of condensing it down though

I agree 100% with you, but if you only spend $500 on your subaru, I doubt you will want to spend anywhere near half that for cruise. You can pickup a used system for pretty cheap, and around here you should also be able to find any parts that might break fairly easily. Mine actuator is 20 years old and works fine. When it breaks, I will get a new one for cheap. In fact I think I have one in a parts box coming my way anyway, so I should be good to go Just my opinion...

Well, technically it pulls the pedal, but same idea. It's attached to the top of the peddle assembly, so therefore when it pulls/sucks it accelerates (it's vacuum operated ), and then when it release the vacuum, it decelerates. It's a really simple but effective system. Works great and has for the last 7,000 miles since I fix it, I just used it about 10 min ago, and I am about to use it again on my way back to work I really like cruise, especially at 2am, that way I don't speed and my speed stays the same