T87 Swaps

[kidturbo]

Sweet. That certainly beats a saw and duct tape approach.. Now to get connecter angles lined up, and call that problem solved. So long as good inventory on the bracket.

Getting closer to locating a local 15.5 model in the area. Will certainly check the oem layout.

Sent from my SM-S901U using Tapatalk

 

[Cougar281]

So I completed my 'half octopus'



With the E35A, E35B and E39A, the T87 is still complaining about a loss of comms with ECM/PCM A, despite both the ECM and TCM being on the same bus and able to be communicated with...

Something I find... interesting... Both the T87 and T87A have DTCs for loss of comms with ECM/PCM A, but the E35A and E35B only throw a loss of comms code with the T87A (which is a 8sp gasser cal). With the T87 running the Allison calibration, neither the E35A nor the E35B throw a loss of comms with TCM code... My basic guess is the E35A and E35B are looking for 'X' and they're getting it from the T87, but the T87 and T87A are looking for 'Y+somthing', where the E35's are just sending 'Y' (or they might be sending 'T' lol)... I wonder if a T87A running an Allison cal would behave the same as the T87 with the Allison Cal...

 

[kidturbo]

So I completed my 'half octopus'

View attachment 119659

With the E35A, E35B and E39A, the T87 is still complaining about a loss of comms with ECM/PCM A, despite both the ECM and TCM being on the same bus and able to be communicated with...

Something I find... interesting... Both the T87 and T87A have DTCs for loss of comms with ECM/PCM A, but the E35A and E35B only throw a loss of comms code with the T87A (which is a 8sp gasser cal). With the T87 running the Allison calibration, neither the E35A nor the E35B throw a loss of comms with TCM code... My basic guess is the E35A and E35B are looking for 'X' and they're getting it from the T87, but the T87 and T87A are looking for 'Y+somthing', where the E35's are just sending 'Y' (or they might be sending 'T' lol)... I wonder if a T87A running an Allison cal would behave the same as the T87 with the Allison Cal...

Glad to hear you now own some Octopus too. They have been appearing on my desk every time I turn on the lights for the past month now.. Must be mating season...

Spent all of last night working on the speed sensor converter problem. My first off the shelf solution, yielded less than stellar results. Plus, what I thought was some 5v Ref lines on the TCM, looks to just be 5v pull up resistors on unused signal pins. Same thing with the Low Ref lines I found...

So I built another new octopus with external 5v power supply, and while it outputs 5v on the 2 signal pins, can't get the TCM to accept the patterns. Tracks perfectly with Uno Crank/Cam Emulator board sending a 5v pattern however. And while it shows valid speeds in our raw GMLAN data, neither speed sensor PID shows an Rpm signal in EFI OBDII data no matter what I feed the TCM with... This OS is crap compared to the T87 gasser version.. Need the SnapOn scope again to sort this one. Maybe I fried a chip, or that design is wrong for our task. Either way, that octopus getting tossed on the grill tonight..

Now that you are all caught up on the ECM mating, seems you are seeing the correct responses. Another thing I have noticed, the T87 only shows a No PCM, No BCM, and No ABS DTC's with key on power. Unlike the gasser OS which lit up all the missing transmission hardware DTC's at Key-On, this Allison version only seems to run a full diagnostics spread once ya feed it some some actual Rpm signal.

After getting the E38 running happy with the T87 gasser OS above, I decided to put that ECM and converter setup back on the bench to start out this Allison OS with. And that soon showed me the Required Message Set the Allison T87 OS needs to clear that No PCM code, is not the same set as the T87 gassers. Leading to the main reason I'm trying to track down an E86B 2015-16 Duramax truck.

Would speed things up if can capture it running happy once. That will determine what CANbus messages this TCM OS actually requires. LML's are the only trucks I don't have live captured data for on my PC. And I tried feeding it some L5P data, remained pissed off. So our msg set on the 15-16LML seems to be specific to this ECM/TCM combination. And that explains why they had to either modify TCM code, or inject the correct messages to make the T87A backwards compatible to a 2011-2015 LML.

Now we know. And was why I said liked the gasser option better. Guess tonight I'll start hacking out why it remains unhappy with this 87 Allison version, before my desk starts smelling like spoiled fish..

 

[kidturbo]

@Cougar281 Ya might also try matching up those VIN's just to eliminate that variable. I don't see the TCM spitting out a VIN #, but I know all the ECM's do. On the 87A gasser stuff I know that is a sticker, along with TCM wants to see a MAF signal or gets pissy I've read on other forums. There is actually a table and DTC the gasser guys write out if running speed density..

 

[Cougar281]

Well, I can now connect up to four modules to my harness at the same time... One thing I really can't wrap my head around is GM and their BCMs. If you search on Car-part for a 2017 GMT900 truck's BCM, what you get back is all GMT900 trucks as a match... BUT (and I forget how I stumbled on this), if you search for a 2015 Cadillac ATS, guess what shows up as a match? Various GMT900 trucks. Similar but different results searching for other vehicles that show up as a match under other parameters. It's so weird. They're all externally identical, and there's absolutely no technical reason for ALL GM vehicles to not share the same physical BCM module... you'd think the bean counters would LOVE that - you only need one SKU for this module to service EVERY vehicle from 2015 to now, the only difference is some software???

 

[Cougar281]

@Cougar281 Ya might also try matching up those VIN's just to eliminate that variable. I don't see the TCM spitting out a VIN #, but I know all the ECM's do. On the 87A gasser stuff I know that is a sticker, along with TCM wants to see a MAF signal or gets pissy I've read on other forums. There is actually a table and DTC the gasser guys write out if running speed density..

Good point. Although I haven't seen that being an issue in 'older' trucks, it COULD be an issue with newer vehicles. ECM and TCM are easy enough to change with EFILive.

 

[kidturbo]

From my gasser version testing, the BCM and ABS are not required to make the TCM happy. I started out with a VIN all matching 16 ECM and BCM, but was able to drop all the BCM data. Yeah it sets a no BCM code, but TCM wasn't in a limp mode because if it. All that would be missing is Driver Mode data. IE: Normal/Sport, Tow Haul, and our Manual Shift related stuff. Which is under my "Stage 2" development notes. For now, no-liimp, shift gears correctly.

 

[kidturbo]

Pulled my trusty old E35B stand alone harness and ECM out of the tote last night. Got it wired up on the bench, along with our A50, and T87 TCM's. Needs to add couple more POTs to clear all the sensor codes. Then connect Crank/CAM emulator to get it spinning. But it's pretty happy so far. Updated the same 2008 Express Van VIN into our Allison T87 OS.

First KOEO check shows same DTC results as @Cougar281 noted. The T87 doesn't recognize the E35B ECM. However the ECM sees the T87, and accepts what the sending to clear the No TCM code. Grabbed a little CANbus data off the two nodes, and now off to sort packets.

 

[kidturbo]

Ok think I may have cracked it. Depending on which PC I have my V2 connected, I can get it to clear the U0100.. On the V8.2.24 version, it clears whenever I feed it 4 or 5 missing ID's similarly to the gasser setup was missing. However on my other PC, with V8.3.30, I can't get it to ever show cleared. On that one, seen in list above, U0100 Lost Comms ECM/PCM is a "Permanent" status. When I looked up EPA Permanent, I get the following.

Permanent Diagnostic Trouble Codes (PDTCs) are very similar to regular Diagnostic Trouble Codes (DTCs). However, unlike regular DTCs, they cannot be reset by disconnecting the vehicle’s battery or cleared using an On-Board Diagnostic (OBD) scan tool. The only way to clear a PDTC is to fix the underlying problem with the vehicle that originally caused the PDTC and its corresponding DTC to set, and then allow the vehicle sufficient drive time to re-run the monitor that identified the problem in the first place. When the monitor runs without identifying a problem, the PDTC will clear itself.

I've tried powering it off several times, clearing with tool, and one shows it, while the other pc doesn't.... Gonna feed it some ECM RPM, then it's back to finding a running truck... lol

 

[TheBac]

Welcome to the wonderful world of permanent codes. Its sooooo fun.

 

[kidturbo]

Welcome to the wonderful world of permanent codes. Its sooooo fun.

Yeah.. We're about to see how permanent it is when I re-flash it's azz right outa this little aluminum box..

Funny thing, while it set same level for the missing valve body, and output speed sensor, it only sets a "Current" for the missing BCM and ABS nodes. Which I find pretty curious.

 

[kidturbo]

Ok, so now it wants to play hardball with it's codes...

The first new thing I discovered about these TCMs, they don't power down with key off..

Re-flashed a cal, and it didn't clear the DTC's upon boot. Figured it would at least take 30 seconds to set a permanent DTC after flashing. Yet the DTC was there soon as I could run a scan.. Then I turned key to OFF, which kills power to the Run/Start pin, and the Serial Data Wakup wire. Also the TCM GMLAN data stopped sending. I scanned it again expecting only the ECM to reply, and guess who spit out a DTC list??

Upon further investigation, canbus is remaining active with Batt pin only powered. That's new to me... lol @Cougar281 please verify this on your setup.

Then I hear a noise outside, and it's UPS truck with half of my Black Friday Amazon order. Broke down and bought a new soldering station, some good breadboard, pots and screw terminals. Tired of chasing loose wires on the bench rig, and soldering with a hot pitchfork..

So now who's got an Allison valvebody pinout handy?? Time to make this prickly little TCM think it's mounted in a truck again..

Sent from my SM-S901U using Tapatalk

 

[1FastBrick]

What year would you like???

 

[kidturbo]

15-16 be good for starters.

Sent from my SM-S901U using Tapatalk

 

[pl_silverado]

15-16 be good for starters.

Sent from my SM-S901U using Tapatalk

 

[Cougar281]

I am not seeing the same behavior. One note - I thought I had checked this before, but maybe not. With the E39A and T87A connected, I am not getting a loss of comms with ECM/PCM A code in the TCM... I do with the T87.

On turning the 'ignition' off, the TCM (both the T87 and T87A) disappears off the bus and stops responding to EFILive after about 8 seconds (although the first time I timed it with a stopwatch, it was over 30 seconds, then the next time it backed up my 8 seconds). The E39A does eventually disappear as I left 'battery' connected with the 'ignition' off while I picked my daughter up from school and an hour later it wasn't responding (it seems to be inconsistent - one time it seemed to shut down after about two minutes, maybe a less - but definitely more than 15 seconds - most of the time it seems the E39A is 'still there' more than five minutes later ). Turning the 'ignition' on brought everything back up.

With the E35A and E35B, upon turning the 'ignition' off, the ECM disappears pretty much instantly, with the TCMs following suit roughly 8 seconds later.

What exactly is the 'serial data wakeup' pin? Is it just a 12v signal or is it actually some sort of structured data signal? I left that one off my T87/T87A/T93 connector because the name suggests structured data, although I've been told the T93's won't program without it...

 

[kidturbo]

Thanks guys those schematics should help.

May have located a complete Allison 6spd valve body close that I can borrow for bit. Nobody beating the door down for a couple cores back setting at my buddies shop.. Since I have a couple full chassis harness laying around here, if this TCM gives any trouble faking the sensors on the bench, I'll go full tilt and wire up the complete valve body. Then lay a big hammer on the bench, and dare it to complain..

I'm used to the older model stuff shutting down soon as ya cut the key power. So this delayed sleep mode approach is new to me. Not uncommon in modern electronics, just never seen it used in GM before. If ya trace the wire, that Serial Data WakeUp originates at the BCM. Not sure if it actually passes any sort of serail data, have never checked.. However it provides a 12v signal to all nodes on the network to WAkE UP when the BCM tells it. Typically at Key On or Key Inserted.

Since our PCM power and Crank relays are controlled by the ECM, based on VATS related commands sent over canbus from the BCM, they needed some way to tell each node to come alive. IE: WAKE UP listen for data. I only know this because in E35B stand alone setups, you have to power that wire to wake up the ECM/TCM before you crank the engine with a direct starter relay to a key. IF not, it won't fire.. Also you'll find it listed as 12v Key On in EFIlive Bench Harness diagrams. And if you forget it, that's how ya brick an E35. Ask me how I know that... LOL. I have helped several people with swaps who missed that wire and had a crank / no run issue.

So by design, GM brings the CANbus nodes up first by powering the Serial Data Wakeup off BCM. You'll notice each node has a Batt+ wire, which is always hot. Then the ECM gets a Key On/Run command msg over the bus if immobilizer data checks out. At that ECM Grounds the PCM relay, which powers up all connected nodes. Same goes with Start commands. I've never watched the serial data wire on the Shut Down process. But guessing it's hot anytime the IC is active., or key hasn't been removed,

This new sleep delay doesn't seem to be tied to our wake up wire anymore. Which again isn't uncommon as things move towards FD CAN and Ethernet based systems that can use Wake On Data. However it's something to keep in mind as we work through the issues with this T87 TCM.

Also interesting you don't see the No PCM Comms message in the T87A version.

 

[Cougar281]

Hmm. I wonder if my E35 is wired with the serial wakeup as IGN... If not having it will brick one on programming, what would explain why I haven't bricked one lol.

I only don't get the loss of comms with the T87A and the E39A connected... My guess being the 2016 E39A out of a Malibu is close enough or identical to the E80 out of a 2017 Camaro that it was paired with (I haven't gotten the E80 working because I haven't gotten the bench connector made for it yet).

Hmmm..... That gives me an idea.... Maybe I should borrow the E39A out of my wife's 15 Malibu and see if it plays equally nice with the '15 T87.....

 

[darkness]

Forgiveness > permission.

 

[TheBac]

Hmm. I wonder if my E35 is wired with the serial wakeup as IGN... If not having it will brick one on programming, what would explain why I haven't bricked one lol.

I only don't get the loss of comms with the T87A and the E39A connected... My guess being the 2016 E39A out of a Malibu is close enough or identical to the E80 out of a 2017 Camaro that it was paired with (I haven't gotten the E80 working because I haven't gotten the bench connector made for it yet).

Hmmm..... That gives me an idea.... Maybe I should borrow the E39A out of my wife's 15 Malibu and see if it plays equally nice with the '15 T87.....

"But Honey, its for scientific research".