The Subaru SVX World Network   SVX Network Forums
Live Chat!
SVX or Subaru Links
Old Lockers
Photo Post
How-To Documents
Message Archive
SVX Shop Search
IRC users:
SVXBot

Go Back   The Subaru SVX World Network > SVX Main Forums > MOD Mania > Transmission and Drivetrain

Reply
 
Thread Tools Display Modes
  #106  
Old 10-12-2019, 08:03 PM
Dispatch20 Dispatch20 is offline
Registered User
 
Join Date: Apr 2019
Location: New York
Posts: 9
Re: OBDII Mt conversion???

I appreciate the support, guys! It does help motivate me knowing people are interested in this.

So after I posted last time, I set up the Arduino and laptop to do more data logging. I was suspecting that I might be missing some sort of TCU message that happens at highway speeds (or at some other period). Well, I did a 10 minute drive and the pattern did not change at all. It stayed exactly as I've always seen it. So then I cleared the CEL light, and drove another 10 minutes and there was no difference in the TCU pattern.

So anyways, I'm thinking that perhaps it is something simple like a grounding problem. The ECU has a number of different ground pins with different labels (refer to the OBD2 pinout link a few posts up). It would be very unsafe to assume all the different labeled grounds are all common with the chassis ground. So my next step is to wire the Arduino ground to the more specific ECU ground starting with the "control systems ground" on pins ECU 17/18). To be honest, I can just check with an ohmmeter to see how the grounds work, but I haven't yet. It's not the most convenient area (under the dash & steering wheel)

Last edited by Dispatch20; 10-12-2019 at 08:24 PM.
Reply With Quote
  #107  
Old 10-12-2019, 08:23 PM
Dispatch20 Dispatch20 is offline
Registered User
 
Join Date: Apr 2019
Location: New York
Posts: 9
Re: OBDII Mt conversion???

Earlier this summer I bought the 1996 Service Manual OBD2 Supplement for $20 shipped on Ebay. I wanted to see what it said about the CEL P1702 code, and any debugging steps. See attached pics for the only 2 pages on this (although it also has the OBD2 updated wire diagrams which is always helpful).

Just like in my recent experience, it takes 2 trips to make the CEL go on. The 1st trip stores a "pending" 1702 code, and then the 2nd trip will turn on the CEL for the code. How the car defines a "trip" is interesting to me. It seems to follow the OBD2 drive cycle rules where the car must be below a certain temperature and then perform some level of driving before it will register the pending/actual 1702 code. I suspect this is why my brief drives around my neighborhood would never create a CEL code (even with the entire TCU disconnected).

The P1702 debugging steps in the tech manual just look for a short or open circuit. But they take the measurements relative to chassis ground, so I think my theory above probably won't help anything.

A second thought I have is that my Arduino starts outputting the signal in the ACCESSORY key position and the actual TCU outputs it in the ON position. I may be remembering this incorrectly, so I definitely need to double check that. I want to make sure that prematurely sending the pattern (i.e. in Accessory instead of On) isn't screwing something up here.

If I really can't get this method to work, then it's time to do a 5MT swap and try leaving the TCU installed and fooling it with load resistors (per the recommendation a few posts up). To be honest, I don't recall why people disconnect the TCU when doing a 5MT swap in the 1st place. Perhaps the ECU gets unhappy if the TCU sends the ECU a torque cut signal when it thinks it is shifting. Or maybe if the TCU doesn't sense proper A/B/C solenoids, it permanently causes a torque cut or some sort of limp mode. I don't think a code P1702 would be thrown, though. (Just capturing some random thoughts for now). FYI - I know people with 5MTs have sometimes had to override the torque cut signal into the ECU to prevent some stalling issues during shifting. Something about the torque cut signal disconnected/floating causing problems for the ECU.

Lastly - Yes, I do have a post on the SVX Nation Facebook page. It doesn't get a lot of interest, and I find this forum page an easier way to capture ideas, pictures, and links.

(One more thought - I should log what the TCU outputs in this ECU serial datastream when the tranny bulkhead harnesses are completely disconnected. I imagine this has to put out a different pattern than I've previously seen. If it doesn't, then what is the point of this new TCU serial data stream? I can speculate that it's required per OBD2 regulations to have the TCU->ECU signal but I really don't know. Subaru calls it a "diagnosis" signal. Since you can already talk directly to the TCU via SSM2 for diagnostics, I don't know what this new signal really does. Maybe it can capture freeze frame data in the ECU for certain tranny faults?)
Attached Images
File Type: jpg obd2_pic1.jpg (436.6 KB, 6 views)
File Type: jpg pbd2_pic2.jpg (504.4 KB, 5 views)

Last edited by Dispatch20; 10-12-2019 at 08:33 PM.
Reply With Quote
  #108  
Old 10-13-2019, 06:11 PM
Dispatch20 Dispatch20 is offline
Registered User
 
Join Date: Apr 2019
Location: New York
Posts: 9
Re: OBDII Mt conversion???

Quick update; sorry about all the text and no fun pictures!

I wired a ground wire from the ECU harness pin 17 direct to a ground pin on the Arduino. I think this might have actually helped! I took two long drives today (separated by 11 hours) and did not get a pending P1702 or CEL. I noticed that wiring this ground also made my datalogging Arduino setup work better. I want to test for a few more days before I claim success, though, but I'm feeling pretty good about it now.

I also datalogged and disconnected the large transmission harness by the starter to see what would happen. When you disconnect the big connector, the TCU just sends all '1's to the ECU which guarantees the P1702 will go on. The TCU must fail to see the right loads/solenoids/switches in the transmission and send the alert to the ECU via the pin 80 "diagnosis" signal. This lends credence to the theory that adding load resistors to the TCU could fool it into thinking it has a proper automatic installed despite having a 5MT swap.

Strangely, after filling up gas on the second half of my 2nd test drive today I got a pending 1101 Neutral Safety switch code. It seems unrelated to this testing but maybe some dirt got in the tranny harness when I plugged it back together, or maybe I didn't plug it all the way in. I'll keep an eye on that.

Lastly, the TCU does, in fact, start sending the serial data when the key enters the ON position (not Accessory). My Arduino starts sending data to the ECU when Accessory power comes on but it doesn't seem to matter. I also noticed that the power cuts out to Accessory devices when the starter is engaged, but it does NOT cut to the TCU. This doesn't seem to matter in terms of P1702, either.

Last edited by Dispatch20; 10-13-2019 at 06:13 PM.
Reply With Quote
  #109  
Old 10-14-2019, 07:28 PM
Sean486's Avatar
Sean486 Sean486 is offline
Happy SVX
Subaru Silver Contributor
 
Join Date: Jul 2005
Location: Mechanicsburg, PA
Posts: 3,373
Registered SVX
Re: OBDII Mt conversion???

Quote:
Originally Posted by Dispatch20 View Post
Quick update; sorry about all the text and no fun pictures!

I wired a ground wire from the ECU harness pin 17 direct to a ground pin on the Arduino. I think this might have actually helped! I took two long drives today (separated by 11 hours) and did not get a pending P1702 or CEL. I noticed that wiring this ground also made my datalogging Arduino setup work better. I want to test for a few more days before I claim success, though, but I'm feeling pretty good about it now.

I also datalogged and disconnected the large transmission harness by the starter to see what would happen. When you disconnect the big connector, the TCU just sends all '1's to the ECU which guarantees the P1702 will go on. The TCU must fail to see the right loads/solenoids/switches in the transmission and send the alert to the ECU via the pin 80 "diagnosis" signal. This lends credence to the theory that adding load resistors to the TCU could fool it into thinking it has a proper automatic installed despite having a 5MT swap.

Strangely, after filling up gas on the second half of my 2nd test drive today I got a pending 1101 Neutral Safety switch code. It seems unrelated to this testing but maybe some dirt got in the tranny harness when I plugged it back together, or maybe I didn't plug it all the way in. I'll keep an eye on that.

Lastly, the TCU does, in fact, start sending the serial data when the key enters the ON position (not Accessory). My Arduino starts sending data to the ECU when Accessory power comes on but it doesn't seem to matter. I also noticed that the power cuts out to Accessory devices when the starter is engaged, but it does NOT cut to the TCU. This doesn't seem to matter in terms of P1702, either.
Wow! This is awesome news. Hopefully, it doesn't come back.
__________________
1992 SVX LSL 170k [I]4.44 Transmission, Ecutune TCU, Ecutune Stage 1, transmission cooler, silver STI BBS wheels, Sony Bluetooth double DIN head unit, Infinity front speakers, Lamin-X protected headlights
2010 Mazda 6 i Grand Touring - 120K
2007 Honda Odyssey EXL - 100K
2004 Mazda 3 S - 106K
-----------------------------------
1995 SVX Lsi Polo Sold at 118K 2007
2006 Mazda 5 Sold at 60K 2012
2005 Mazda 6i Sport Sold at 206K 2013 great car[/COLOR]
Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -6. The time now is 09:53 PM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2019, vBulletin Solutions, Inc.
2001-2015 SVX World Network
(208)-906-1122