5EAT TCU Reverse Engineering

[utc_pyro]

Oct bump!!

 

I kind of forgot about the resistor mod, Ill check and see if my local audio shop has one for cheap and may try it out.

 

A nabisco guy with 4eat said he was having cel with the ipt resistor (he had full ipt trans built) so ipt removed it and he said that made the most difference (that he can feel)

 

If you want to try it, I'll buy and ship you the resistor out of the TCU test fund.

 

PM me with shipping info and how agressive you want to go for. If you're hesitant on soildering I can prep it to be a quick cut-and-crimp install.

Edited October 6, 2010 by utc_pyro

[LGTONCOKE]

my problem is that the tcu can't think fast enough so I have to shift it myself.

[underground000]

If you want to try it, I'll buy and ship you the resistor out of the TCU test fund.

 

PM me with shipping info and how agressive you want to go for. If you're hesitant on soildering I can prep it to be a quick cut-and-crimp install.

 

Its ok, its not that expensive to do

Ill be delayed again cause I have to go to Colorado next week:redface: Unless I do it by the end of this week

[fishbone]

my problem is that the tcu can't think fast enough so I have to shift it myself.

Trust me when I say the TCU "thinks" in microseconds, but executes slow because the Legacy is not a GT-R so the Subaru engineers wanted to make sure they don't freak out Grandma Slow. What behavior are you referring to? Gear shifts? Yeah, painfully slow in stock form, becomes obvious when the car is modded.

[LGTONCOKE]

In my specific situation. I have had clutch packs cryo treated and valve body ported. But if I put it in sport and then smash it to the floor it can not think fast enough to shift before it hits red line then hits face and kinda dies then picks back up.

[g00fy]

Visiting from the forester community here to say keep up the good work! I'm workin on pushing some seriouse numbers with my new 35R setup... Unfortunately even with my built IPT trans the TCU is still too stupid to actually take advantage of everything.

 

My biggest gripe is Torque converter Lock-up, I have an extremely loose Torque converter and the thing wastes soo much power cause it never locks up under load, it also puts the hurt on my fuel economy.

 

Second is shift points/shift control, the TCU is always so slow to shift and usually it shifts too late and drops me into the rev limiter, also on the dyno it'd be nice to be able to hold the car in 3rd to get a true 3rd gear pull.

 

Last is line pressure, the shift firmness is acceptable, but it can definately be much better.

 

I'm hoping to break into the 10's with my new setup, and i think i may have the horsepower, but I have a feeling the transmission is really going to hold me back.

 

So anyway good luck on the project, if you need donations i'd be happy to contribute!

[utc_pyro]

Its ok, its not that expensive to do

Ill be delayed again cause I have to go to Colorado next week:redface: Unless I do it by the end of this week

 

Anyone else want to try it? Again, I'll make you a quick install one if you promis to install it within a few days of delivery.

 

So anyway good luck on the project, if you need donations i'd be happy to contribute!

 

There is enough funding in the fund still to cover more test equipment. We need someone that knows some one at Hatachi that can tell us what the heck the chips are.

 

though that said, if some oen chan crack open a newer forster TCU and post pics it could be helpful. Maybe they use some normal braded components?

 

I was looking at my SpartinIII board (FPGA running a microblaze core for a class I'm in), and NSFW's thinking it's a FPGA agin is looking quite possible. The smaller chip on the right only has a few wires going to the main processor, and nothing else. The layout is quite simular to flash arangment to program the FPGA on my devboard. If this is the case, it may take some one FAR smarter then me to take the bitstream and create a netlist. That or pay a reverse engineering house in China to take it, but I think Cobb would be the only ones with the budget to do that.

Edited October 13, 2010 by utc_pyro

[g00fy]

Anyone else want to try it? Again, I'll make you a quick install one if you promis to install it within a few days of delivery.

 

 

if some oen chan crack open a newer forster TCU and post pics it could be helpful. Maybe they use some normal braded components?

 

 

What do you need tested? And i'll see if i can get at my TCU this weekend and get some pics, is 05' new enough?

[underground000]

What do you need tested? And i'll see if i can get at my TCU this weekend and get some pics, is 05' new enough?

 

The resister mod the bump up the line pressure

[ClimberDHexMods]

I'm fairly certain all you need to do is scale up your Requested Torque table in Romraider. It tells the TCU how much line pressure to apply at any given mph and rpm per reasonable and unreasonable trans fluid temp.

[underground000]

I'm fairly certain all you need to do is scale up your Requested Torque table in Romraider. It tells the TCU how much line pressure to apply at any given mph and rpm per reasonable and unreasonable trans fluid temp.

 

 

 

Unless I'm getting a road tune by Infamous I'm not going to buy a $150 cable for my 08 car just for increased line pressure

[seabass07]

So it can't be done with an ebay cable? I thought there was a CAN-BUS ebay one.

[hadvw]

I'm fairly certain all you need to do is scale up your Requested Torque table in Romraider. It tells the TCU how much line pressure to apply at any given mph and rpm per reasonable and unreasonable trans fluid temp.

 

Any idea if AccessPort has a similar table? I currently have an AP tune..

[utc_pyro]

I'm fairly certain all you need to do is scale up your Requested Torque table in Romraider. It tells the TCU how much line pressure to apply at any given mph and rpm per reasonable and unreasonable trans fluid temp.

 

Give it a shot. Every one seems to cap out at 320(?) on the request torque, I've never seen someone take it higher*. I'm hesitant to going to 100% uncharted teratory with my only way to work .

 

With a stage 2 tune, scaling it up in the lower revs will probably help duribility. We have a LOT more torque down there than stock. Even if you dont go past 320, you'll get more pressure.

 

*I'm fairly new to all this, maybe someone tested this in the past?

[g00fy]

The resister mod the bump up the line pressure

 

Oh, I had that on my car after IPT rebuilt the trans, I don't know if it really made that much of a difference, but i can tell you it kept throwing a trans CEL so it was pulled and the CEL hasn't returned. IIRC the problem was the resistor caused the solenoid to be over-loaded which triggered a fault code.

 

Give it a shot. Every one seems to cap out at 320(?) on the request torque, I've never seen someone take it higher*. I'm hesitant to going to 100% uncharted teratory with my only way to work .

 

With a stage 2 tune, scaling it up in the lower revs will probably help duribility. We have a LOT more torque down there than stock. Even if you dont go past 320, you'll get more pressure.

 

*I'm fairly new to all this, maybe someone tested this in the past?

 

My Torque table is maxed out at 400, again I can't tell if it really makes that much of a difference.

[utc_pyro]

My Torque table is maxed out at 400, again I can't tell if it really makes that much of a difference.

 

The 4eat has a direct input from the throttle, so you may be out of luck on that one. If you scale your torque request and throttle angle tabels up or down by 30% you could test if it changes anything.

 

On the resistor throwing CEL's, you might want to try putting one back in with a smaller value. If it was too high, the TCU will think the solonoid has failed. It works opposit of how you'd think, the less power going to it, the more pressure you get.

[subarutech77]

I'd be willing to install the resistor in my 05 OBXT, if you still need testers. How do you want it installed? parallel or in series? are you just looking for an increase in line pressure duty? I can get actual line pressures, but that may have to wait until later next week.

 

I also have an 05 Baja sportshift TCU, that I was going to try in my Forester to help with the wheel speed difference issue, I can take pics of it tonight. I have opened it up and does not look anything like the 5eat tcu pictured in the 1st post of this thread or the earlier 4eat tcu's of the SVX pictured on Phil's site

Edited October 14, 2010 by subarutech77

[g00fy]

The 4eat has a direct input from the throttle, so you may be out of luck on that one. If you scale your torque request and throttle angle tabels up or down by 30% you could test if it changes anything.

 

On the resistor throwing CEL's, you might want to try putting one back in with a smaller value. If it was too high, the TCU will think the solonoid has failed. It works opposit of how you'd think, the less power going to it, the more pressure you get.

 

 

Well, I'll see if the wiring is still in place, if so what size resistor would you recommend? I think the code was High Duty Cycle Solenoid C or something along those lines.

[utc_pyro]

I'd be willing to install the resistor in my 05 OBXT, if you still need testers. How do you want it installed? parallel or in series? are you just looking for an increase in line pressure duty? I can get actual line pressures, but that may have to wait until later next week.

 

In series for now. This is really a mod untill we figure out something else and long term as just a "budget" mod.

 

Well, I'll see if the wiring is still in place, if so what size resistor would you recommend? I think the code was High Duty Cycle Solenoid C or something along those lines.

 

If you can measure the resistance I can give you a suggestion. i dont have the wiring diagram on hand for the Forister, if you PM me a link I'll let you know a suggested value.

 

 

 

 

Guess what I found thanks to a tip:

 

http://www.renesas.com/media/products/mpumcu/m32r/m32r_ecu/32185/child_folder/e32185_pinar.pdf

 

Compare that to the pics in post 1. We got JTAG

[dschultz]

There's a bit of info here on the M32R.

[ClimberDHexMods]

Unless I'm getting a road tune by Infamous I'm not going to buy a $150 cable for my 08 car just for increased line pressure

 

You can lead a horse to water...

 

Give it a shot. Every one seems to cap out at 320(?) on the request torque, I've never seen someone take it higher*. I'm hesitant to going to 100% uncharted teratory with my only way to work .

 

With a stage 2 tune, scaling it up in the lower revs will probably help duribility. We have a LOT more torque down there than stock. Even if you dont go past 320, you'll get more pressure.

 

*I'm fairly new to all this, maybe someone tested this in the past?

 

I am in the process of trying it out. Mine maxes out at 416 for now. I think that is appropriate given I have a 20G on pump gas.

 

I really think this is a huge unturned stone. I can explain all sorts of shift firmness behavior based on my old requested torque map. Hopefully it's not just a coincidence. For instance, at low-to-mid throttle and low RPMs, I get a slam from 1to2 (probably from slipping clutch, speaking of not enough line pressure). RT is relatively high in this area. Conversely, at high RPMs, it won't shift fast enough anywhere. Perhaps more RT will increase LP before it hits its physical or ROM-limited max.

 

I am at the beginning of my retune. Once my tune is finished, I am happy to mess with it in dramatic ways to see what I can feel. My inclination is to bring it up as high as I can until it starts slamming, if it even will. I would guess that there is a peak requested torque value above which you will not get any line pressure increase.

 

Final note, IPT modified my valve body, so I should be able to make more line pressure than a stock unit, if I understand it correctly.

[utc_pyro]

There's a bit of info here on the M32R.

 

 

jackpot

 

The 4 pads on the right of the JTAG are the real-time debug ports. Subaru/Hatachi left us most of the tools we'll need to get serrious with this thing.

Edited October 14, 2010 by utc_pyro

[underground000]

Any idea if AccessPort has a similar table? I currently have an AP tune..

 

In AccessTUNER Race, theres 'Gear Determination' - Dont think it has anything to do with line pressure.

 

Under 'Throttle Tuning': - Again I dont think it controls line pressure, numbers start dropping after 2800rpm

'Requested Torque A - Intelligent'

'Requested Torque A - Sport'

'Requested Torque A - Sport #'

'Requested Torque B - Intelligent'

'Requested Torque B- Sport'

'Requested Torque B - Sport #'

http://i596.photobucket.com/albums/tt49/underground000/acc.png?t=1287099720

[utc_pyro]

That's just how the ECU figures out what gear it's in. It does not mess with the TCU at all.

 

Still cant figure out what that second controller is. I was thinking 78K0, but the Cristal is in the wrong place. Hum.... Do I2C flash memories normally have there own clock?

Edited October 15, 2010 by utc_pyro

[ClimberDHexMods]

In AccessTUNER Race, theres 'Gear Determination' - Dont think it has anything to do with line pressure.

 

Under 'Throttle Tuning': - Again I dont think it controls line pressure, numbers start dropping after 2800rpm

'Requested Torque A - Intelligent'

'Requested Torque A - Sport'

'Requested Torque A - Sport #'

'Requested Torque B - Intelligent'

'Requested Torque B- Sport'

'Requested Torque B - Sport #'

http://i596.photobucket.com/albums/tt49/underground000/acc.png?t=1287099720

 

That is the same table I am referring to. It says so right in the part of the vacation pics that talks about trans control. TCU references 4 values including "REQUESTED TORQUE" to determine target line pressure. That is the exact table.