5l40e as340r / as390r control requirements

[gurov]

this is an all electronic gm transmission. mine is in my 325xi bmw attached to an awd transfer case. same trans is used in holdens, solstice, some Cadillacs, and other vehicles.

this is a 5 forward speed transmission, reverse, neutral, and 'sports mode' which is electronically different.
the bare minimums to control the trans:

TO trans

3 output on/off solenoids for shift pattern
line pressure PWM solenoid
TCC torque converter


FROM TRANS:

5 on/off (12v high, gnd low) channels for lever position
INPUT SHAFT speed (compare with engine RPM to get TC slip)
OUTPUT shaft speed

from the interior of the vehicle there is a mechanical linkage that goes to the actuators to put the trans into R, N, D, P, and steptronic mode.

steptronic mode is a on/off switch that goes high when on (12v again)
there's paddle up
paddle down (active high 12v)

gears 1 2 3 have 2 different output patterns each, one is defined as "engine braking" other is defined as not. 4 and 5 are always engine braking (according to docs)

now, for the control to happen, i assume there would have to be a load signal (i.e. manifold pressure)

to be continued

[Fred]

you would probably want tps and map.

tps tells you what the driver wants and map tells you how much load the engine is under, these are different depending upon RPM and throttle size.

can you define all of the shift patterns and how the code would progress from one gear to the next?

just assume i know nothing about slush boxes.

is TCC torque converter controller? if so, what type of control does it need?

what about the engine braking modes? just switch from 2nd accel to 2nd brakign when the driver lifts?

Fred.

PS for anyone wondering, the idea is to possibly create a FreeShift firmware variant with similar comms and structure but different guts. Before you suggest that this is derailing the main EMS effort, consider that to make them share the code base, the code will have to be clean(er) and modular so we dont duplicate the maintenance. I'm unsure how it would go at this stage, project structure wise but it could potentially be part of the same git project with a different make file target and actual shared files. there are other options, too.

[gurov]

these are my notes from 5l40 research:

Code: Select all

PIN ECU   CONN

L1 - 17 -  11
L2 -  1 -   2
L3 - 14 -  16
L4 - 15 -   7
L5 - 16 -  19

selector
L1 - L5

PIN ECU   CONN

L1 - 17 -  11
L2 -  1 -   2
L3 - 14 -  16
L4 - 15 -   7
L5 - 16 -  19



         1  2  3  4  5

P        1  1  1  0  0
R        0  0  1  1  0
N        1  1  0  1  0
D        0  0  0  1  1
4 (3)    1  0  1  1  1
3 (2)    0  0  1  0  1
2 (1)    1  0  0  0  1

-------------

steptronic:

downshift - 20
upshift   - 19
mode      - 18
input 1 = 17
input 2 =  1
input 3 = 14
mode    = ?
sol1    = 20
sol2    = 10
sol3    =  9
temp    = 19 or 18 ?

LUF     =  6
TCC     =  8
brake   = ?
input speed
output speed




output
-----
  r
P   0 0 0 
R   0 0 0 
N   0 0 0 
1   0 1 0  
1   0 1 1  engine braking <--
2   1 1 0  
2   1 1 1  engine braking <--
3   1 0 0  
3   1 0 1  engine braking <-- 
4   0 0 1  engine braking <-- 
5   0 0 0  engine braking <--

i will have to look into how the trans actually wants the inputs to progress as far as torque converter clutch locking/unlocking and line pressure.

also notice, there's now a brake input, which i guess could be used for engine braking switch. mshift does not do this now, so the definition would have to be for either one or the other.

in my opinion, engine braking would be the default mode to keep the things more locked and more tied together to avoid the slushbox feel.

the TCC and line pressure solenoids have specified frequencies that are recommended, these are outlined in the technical docs of the trans that i would have to dig up.

i fully think this should be a part of the EMS. on the BMW, this is a separate box that talks via CAN to the main controller, but combining this into one could potentially add value. Look at AEM, their boxes include such features already and just have to be enabled, configured and wired.

[Fred]

Absolutely not! There is NO way this should be part of the EMS. No way at all. Ditto wideband. Each discreet piece of functionality deserves a full chipset to work with, and not have to make compromises to co-exist. As it stands the EMS will have the chip maxed out feature wise. Which is fine, but if you take it past maxed out, that is not. Dedicated control is the only way to roll.

Otherwise, great info! :-)

One question, park, reverse and neutral all have 000 pattern, how do you select reverse? Or am I blind? What have I missed? Wait, I see something. There are two lists of codes, one with three values, one with five, what does each mean?

Fred.

[thebigmacd]

Absolutely not! There is NO way this should be part of the EMS. No way at all. Ditto wideband. Each discreet piece of functionality deserves a full chipset to work with, and not have to make compromises to co-exist. As it stands the EMS will have the chip maxed out feature wise. Which is fine, but if you take it past maxed out, that is not. Dedicated control is the only way to roll.

Otherwise, great info!

One question, park, reverse and neutral all have 000 pattern, how do you select reverse? Or am I blind? What have I missed? Wait, I see something. There are two lists of codes, one with three values, one with five, what does each mean?

Fred.

The one with 5 inputs is the feedback from the selector switch sensor. Tells the TCM what position is selected. A more traditional way of doing this was with a resistor network, giving different voltages for each position. I like this binary way better.

The one with 3 outputs is the solenoid shift pattern. The shift selector manually enables hydraulic circuits to get park, reverse, neutral, and drive, and in the drive position the combinations of the three solenoid signals control which gear/mode the transmission is in.

Based on the pattern in the table, it looks like the third solenoid signal would enable the torque converter lockup solenoid, but the fact that 5th gear has braking without that line enabled (no lines enabled, in fact) refutes that deduction.

Page 11 in this link tells you a lot more about the operation:
http://www.transtec.com/tech_insert/IS-5L40E-1.pdf

If I glean anything new from this document, it is the point that no matter what the selector position it is important to send the correct solenoid pattern to the transmission based on speed and load, so that it the shifter is bumped into drive from neutral, for example, nothing blows up.

This too, is a good resource: http://www.crossedflags.com/forum/showthread.php?t=458
The TCC solenoid operates at 35 Hz PWM
The pressure control operates at 292.5 Hz PWM

Here's the official GM manual
http://www.scribd.com/doc/34140538/5L40E

[gurov]

the "safe" output pattern is ALWAYS 000, which places the trans into 5th.

[Fred]

Why is safe not neutral? If you're at 7k rpm and you slam 5th, something is going to give and it aint gunna be the tyres.

[thebigmacd]

Why is safe not neutral? If you're at 7k rpm and you slam 5th, something is going to give and it aint gunna be the tyres.

Keep in mind there is a torque converter, which helps dampen the "slam"

[Fred]

But at 7k the torque converter is in full stall, or do I fail at understanding TCs?

[thebigmacd]

But at 7k the torque converter is in full stall, or do I fail at understanding TCs?

Full stall is relative to the applied torque. If you increase the input torque the stall speed increases. ie "slip"

That's why "stall speed" is a somewhat arbitrary rating. Unless the TCC is locked, the shock of a slam-shift will partly be absorbed by the converter.