E67 to T87/T87A TCM interface needed

[Horshack]

Hi Ed -

I've found A solution, just not the one I really wanted. Right now, today, I can get an adapter to put a Dodge 8HP90 transmission behind any Chevy engine, and get a controller that requires a TPS and RPM signal. It will control the transmission and allow transmission tuning via HP Tuners as though it were in a Dodge Charger. This is my current #1 solution, #2 being to get a Dodge 6.2L along with the transmission. But, if I could avoid all that and run an 8L90E or a 10L90E behind an LS3 and skip the adapters that's how I'd really like to go.

Ideally, a controller for the transmissions wouldn't necessarily require an E67 though. If a standalone controller could be made to work that just took the appropriate signals and let the transmission handle the shifting, that would be the best way to go. EFI Live can segment swap a manual trans on just about any engine, so if the controller that worked with the trans just took in the necessary inputs as a standalone signal and CANNED them for the transmission, that would be the best solution. That would allow you to put an 8/10 speed even behind a carbed engine with no computer so long as you sent the trans an RPM and TPS signal and there are standalone TPS signal generators out there that can be fitted to a carb. The market would be wide open that way.

If anyone wanted to build that product, I still have a ways to go before I'll be forced to make a final decision. You wouldn't believe how hard it is to find a shop willing to do radical suspension work that isn't already up to their eyeballs in work. It's getting to the point that I might have to learn how to weld and build a friggin' suspension table because I expected to have my project back and be prepping it for the engine by now but there's yet to be a single bolt turned. I don't mind saying though, that even if anyone did take on this project, there's no guarantee that I'd go with it. When I'm ready to buy a transmission, your product would need to be tested, in production and on the market. If it isn't, then I'd go with the ZF solution and that would be it for me. Well, at least until my next project, anyway.

Still lurking around.

[Horshack]

Hi Ed -

I've found A solution, just not the one I really wanted. Right now, today, I can get an adapter to put a Dodge 8HP90 transmission behind any Chevy engine, and get a controller that requires a TPS and RPM signal. It will control the transmission and allow transmission tuning via HP Tuners as though it were in a Dodge Charger. This is my current #1 solution, #2 being to get a Dodge 6.2L along with the transmission. But, if I could avoid all that and run an 8L90E or a 10L90E behind an LS3 and skip the adapters that's how I'd really like to go.

Ideally, a controller for the transmissions wouldn't necessarily require an E67 though. If a standalone controller could be made to work that just took the appropriate signals and let the transmission handle the shifting, that would be the best way to go. EFI Live can segment swap a manual trans on just about any engine, so if the controller that worked with the trans just took in the necessary inputs as a standalone signal and CANNED them for the transmission, that would be the best solution. That would allow you to put an 8/10 speed even behind a carbed engine with no computer so long as you sent the trans an RPM and TPS signal and there are standalone TPS signal generators out there that can be fitted to a carb. The market would be wide open that way.

If anyone wanted to build that product, I still have a ways to go before I'll be forced to make a final decision. You wouldn't believe how hard it is to find a shop willing to do radical suspension work that isn't already up to their eyeballs in work. It's getting to the point that I might have to learn how to weld and build a friggin' suspension table because I expected to have my project back and be prepping it for the engine by now but there's yet to be a single bolt turned. I don't mind saying though, that even if anyone did take on this project, there's no guarantee that I'd go with it. When I'm ready to buy a transmission, your product would need to be tested, in production and on the market. If it isn't, then I'd go with the ZF solution and that would be it for me. Well, at least until my next project, anyway.

Still lurking around.

And, FAIL. The solution I found would work great if I were building an engine making less torque than the transmission can handle, but I'll be making quite a bit more. Realized that the standalone solution would not be able to talk back to the PCM, so there's no possibility of torque management with the standalone. So, back to hoping someone here can work a little wizardry to make the E67 and the T87/T87A friendly to each other. Otherwise I'm back to the original problem, LS3 and a 6 speed auto or LT engine and an 8 speed auto.

[Horshack]

Bump...

[kidturbo]

Solved this issue last fall, and tested it against an E38 on the bench, and in a vehicle.

Currently two possible ways of doing it. Both requires fudging some CANbus data. The main issue is, GM added a few more ID's to the HS GMLAN protocol over the years. And the T87 /T87A have a "Required Message List" in the firmware that it must see or it sets a U0100 hard code, and goes into a limp mode. So if we just fabricate the missing messages, inject them into the bus, all is well. Almost...

Above I said GM added ID's to the HS GMLAN protocol over the years, well they also added some byte length to some existing messages, and our new TCM's not only wants to see the ID's, they also check the Byte Length in each ID. And one of our main Engine Data ID's, 0xOC9 has grown from a 6 Byte length message on the E38 to a 8 Byte length message on the E92. They added Boost data..

So now for those of you who enjoy playing with CANbus, you'll understand the retro issue GM created into the new new hardware. It's easy to fudge some data and inject it onto a bus, but ya can't intercept and change what's already sent by another node.. Unless you put a gateway in the middle. Then your need to RX every packet, on the main bus, and filter, rebuild, and TX the required packets on the second bus that the TCM is looking for. And then do the same going back the other direction. Not something super difficult to code, just takes a lot of time, and testing to make sure it's 100% accurate handling message timing, and matches ISO and GM standards. If not, ya might have a brick fist time you go to load a tune thru a gateway and it fails. Or ceases to talk while programing the ECM.

Which bring me to the first way I did it, down and dirty hack a bin, and avoided using a full gateway. While you couldn't just write out all the "missing" required ID's from the T87 code, because some are torque related messages and rolling counters to verify everyone is in time. I was able to replicate all required messages from existing ECM CANbus data, using some captured E92 data as a guide. However the TCM still wasn't happy, and couldn't clear the U0100 code while connected to the E38. That's when I noticed some message lengths were shorter in the E38 stream than the E92. Sure enough, the TCM was counting...

Next I hit up a buddy who's pretty handy in the bin department, and ask if he though we could "Add" some bytes to an exiting CANbus ID message length for the E38. He did a little digging, sent me a couple bin's and on the second try we had it. T87 cleared the U0100 code, and runs happy on the bench with a Crank/Cam emulator firing the E38 ECM feeding all the throttle, torque, and rpm data over CANbus. So if this approach work for the E38, it should work the same for the E67. And the only hardware required was a teensy with a CANbus transceiver..

So now that we know it can be done by modifying the ECM bin, next we are going to try it by modifying the TCM bin. Couple simple tweaks there and it should work just the same, but requiring a customer TCM bin rather than the ECM bin. And for the CANbus hardware, believe we need to add some extras like P/N output signal, VSS output signal, reverse light, and probably a mode and tap-shift input channels. Think that should cover the stand alone builds. And the same hardware should work with T87A on a 10L90. Actually that's what I tested the CANbus code on..

Let me know what ya think. And if anyone has a setup and wants to beta test some hardware with us, be happy to set ya up.

[gmtech825]

very cool stuff!

[Tre-Cool]

very interesting stuff about the can byte counts needing to be longer.

a week or so ago, i got tagged with something on facebook with guy selling 10L90 kits with T87 controllers etc to go in place of the 6l80e, so perhaps these guys are doing the same.

they say it's plug & play.

http://www.powertrainswapsolutions.com/

i'd love to give the 10l80/90 a go, but the only cars we got them in were the 6th gen camaro's here in australia & would need to pick 1 up from a wrecker.

[In-Tech]

So now that we know it can be done by modifying the ECM bin, next we are going to try it by modifying the TCM bin. Couple simple tweaks there and it should work just the same, but requiring a customer TCM bin rather than the ECM bin. And for the CANbus hardware, believe we need to add some extras like P/N output signal, VSS output signal, reverse light, and probably a mode and tap-shift input channels. Think that should cover the stand alone builds. And the same hardware should work with T87A on a 10L90. Actually that's what I tested the CANbus code on..

Let me know what ya think. And if anyone has a setup and wants to beta test some hardware with us, be happy to set ya up.

I'm interested in learning/helping. I have the crank/cam signal sim too. I really want to learn the 10spd but I have no use for it yet. My only dilemma is being able to set aside some bench time since I am still recovering from some bad sickness.

I am curious if anyone knows if I can use my E92/T87A 2017-2018 8spd transmission with a combo 2018 E92 and 2016 T87 or if I MUST use the newer T87A? This is my own personal transplant and not even close to being ready.

Thanks for the advice and files you have been posting. A 10 spd T87A file would be cool to ponder

[kidturbo]

very interesting stuff about the can byte counts needing to be longer.

a week or so ago, i got tagged with something on facebook with guy selling 10L90 kits with T87 controllers etc to go in place of the 6l80e, so perhaps these guys are doing the same.

they say it's plug & play.

http://www.powertrainswapsolutions.com/

i'd love to give the 10l80/90 a go, but the only cars we got them in were the 6th gen camaro's here in australia & would need to pick 1 up from a wrecker.

That's actually the same guy I tested this with live last November. Same one I supplied the custom Teensy board with the running code, test drove his truck with the 10L80, then suddenly claimed he wasn't interested perusing this project any further.. Didn't even offer to kiss me...

But since this cat had talked me into spending a couple hundred hours mapping this T87/A operation inside and out against several other 2004-2020 GMLAN PCM's and nodes, figured I should also do the A40/50 Allison versions. Which is where I've been focused until I saw that same link you shared above. Then I stumbled onto this thread, and noticed I haven't posted anything of value here in past 9 years.. But have been playing with GMLAN heavily about every other day.

So figured I'd toss it up here, that first I didn't solve this alone, as some other very bin-wise guys pitched in with me. And second that it's possible to do these swaps several different ways depending on your build. This TCM can run about any GM transmission built to date. And some Fords also I hear. But matching one of these new T87's up with any older GM hardware shouldn't cost more than a few hundred bucks to accomplish. They just tried to make it more difficult for some reason..

E38-T87-1.jpg (122.06 KiB) Viewed 73 times

Depending on the TCM OS, it wants to see a few other nodes on the bus like BCM and EBCM data. And we've ran into data differences between years on those nodes. IE: 3 sensor or 4 sensor ABS data. For some things we can just flip a switch in the bin, it seems, but haven't fully tested all these tweaks. So I'm continuing to approach this from a 100% stand alone deal also, based around a custom OS on the TCM side for starters. Modifying the default DTC list and trimming down that required ID list best we can to match up with earliest GMLAN releases. Then the rest should fall into place. Still requires an extra piece of hardware to fill in gaps like P/N switch and such, as I mentioned before, but here is example of T87 running on my bench against an E38 for several hours. No limp, no comm codes. So lets find something Holden to test one against..

E38-T87-2.jpg (161.73 KiB) Viewed 73 times

[kidturbo]

I am curious if anyone knows if I can use my E92/T87A 2017-2018 8spd transmission with a combo 2018 E92 and 2016 T87 or if I MUST use the newer T87A? This is my own personal transplant and not even close to being ready.

I'm gonna say the 2018 E92 against a 2016T87 should run fine outa the box. Best I can tell, there isn't any major data differences between the 87 and 87A on the gasser OS between the years. Was some fuel pump issue I recall based on ECM OS? Best I can tell the TCM hardware is the same even. The T87A code clearly started out life as the T87, and is basically what I have been working backwards from. I see no reason to not run the cheaper 16 model if ya can. So far I've swapped between the Allison and Gassor OS versions a dozen times on a T87. It's a trick since nobody has a decent fully open bin loader yet, but we figured out how to trick an existing tuner.. Now I'd like to try and load a 2019 OS onto the 16 hardware. I bet it would load just fine either way.

I haven't cracked open a 87A yet, but have a 87 in front of me now that I've been getting mid-evil on for a few months, and just busted it free from it's enclosure. Pretty rugged little board, with a good STM processor. The hardest thing to hack has been the new style shaft speed and VSS parts. But I got it going now where I can run a old VR sensor against the new Square wave pulse the T87's require.. And that wasn't the 20 minute converter chip fix anyone who knows hall sensors would have guessed. So be keeping that one as my secret for now.. LOl..

[Tazzi]

That's actually the same guy I tested this with live last November. Same one I supplied the custom Teensy board with the running code, test drove his truck with the 10L80, then suddenly claimed he wasn't interested perusing this project any further.. Didn't even offer to kiss me...

Seriously?
So.. monitored what you were doing then just took off ?