Designing a Device to Simulate a Functioning Automatic Transmission (For Manual Swap)

Thread Starter

pt500

Joined Apr 6, 2018
8
I'm in the process of designing a plug and play device to make a modern (early 2000's) Mercedes Benz think it's automatic transmission is present and functioning when instead it has been removed for a manual transmission conversion. The concept is to plug this device into the connector originally meant for the transmission, and not trigger any fault codes. This device will also connect to two buttons, activating "Neutral" and "Park" conditions according to the ecu. Park signal is required to remove key from vehicle, and neutral is driving mode with manual transmission. I have a complete service manual for the transmission, designation 722.6. I also have theories on how half of the system would work, but am here to request clarification and assistance. Attached is a photo of the "conductor plate", the electronic board found inside the 722.6 transmission and communicates with the Transmission control module (TCM), and the wiring diagram that shows how the TCM connects to the Transmission. My device will be prototyped using a spare conductor plate, and potentially streamlined to a small PCB inside an enclosure in the future, that could be tucked inside the car.
The components are as follows:

- Torque Converter Clutch (TCC) Solenoid (listed as pulse width modulated)
- Modulated Pressure Control (MPC) Solenoid
- Shift Pressure Control (SPC) Solenoid
- 1-2/4-5 Shift Solenoid
- 2-3 Shift Solenoid
- 3-4 Shift Solenoid
- Transmission Fluid Temperature (TFT) Sensor
- Input Speed Sensor N2 (Hall Effect sensor)
- Input Speed Sensor N3 (Hall Effect sensor)
- Park/Neutral Switch
**********************************************************************
Separate from Conductor Plate:
- Parking Lock Interchange (overcomplicated switch designed to tell the car the transmission is actually in park and it is safe to remove the key. Car will star and shut off in neutral, but park signal is necessary for key removal)

The service manual lists resistance values for all the solenoids and the TFT sensor. The P/N switch is a magnetic reed switch.
Service Manual Links (with complete diagrams and resistance values):
http://www.all-trans.by/assets/site/files/mercedes/722.6.pdf
http://www.all-trans.by/assets/site/files/mercedes/722.6.1.pdf

Where I'm at currently:
- Bypass all solenoids with appropriate resistors
- Bypass TFT sensor with resistor set to random temperature (TFT sensor signal is ignored when car is in park or neutral so T value is irrelevant)
- Rewire P/N switch to switch in Center console to toggle drive modes (Along with Parking Lock Interchange, which allows key to be removed in park)
***- Alternately, wire Parking Lock Interchange in line with Emergency brake, eliminating the necessity for a "Park" mode button. Car would be started with E-Brake on, creating Park condition for the TCM, and once the brake is released the neutral button could be pressed to "select neutral position" according to the TCM.

My questions are:
Will an appropriate resistor behave the same as a solenoid valve (electrically speaking) to fool the TCM into thinking the solenoids are present and functioning?
How would one create a pulsed or intermittent signal to fool the two hall effect sensors? According to the manual the speed measured by each sensor individually must match. Some Math needs to be done to calculate proper pulse rate that needs to be simulated. Some online research has shown a 555 timer could possibly be used to replicate the hall effect signal, but I am unclear on how that would work.

I appreciate any and all input, thanks.
 

Attachments

Thread Starter

pt500

Joined Apr 6, 2018
8
****Edit: Forums #1 Rule, SEARCH FIRST!!! My apologies. I turned up this article here on ACC:
https://forum.allaboutcircuits.com/threads/vehicle-speed-sensor-simulator.137175/
And it made me realize, the signals emulating the two hall effect sensors will need to be RPM dependant, or a fault will be triggered due to mismatched engine and "transmission" speeds. I'm a little confused though, as the TCM reads different sensors in different gears. Here's an excerpt from the transmission service manual:
"The 722.6 transmission uses 2 input speed sensors referred to as N2 and N3. Both speed sensors are located in the electrical conductor plate, as shown in Figure 19. The speed sensors are Hall Effect speed sensors that are used by the TCM to calculate the transmissions input speed. Since the input speed could not be measured directly, two of the drive elements are measured. N2 records the speed of the front sun gear and N3 records the speed of the front planetary carrier. Two input speed sensors were required because both drive elements are not active in all gears. The input sensors N2 and N3 will report the same input speed in 2nd, 3rd or 4th gear. If the N2 and N3 input speed signals are not the same in these gears, then there is an issue with the transmission and the DTC for "Input Speed Sensors Mismatch" will be set. The N3 input speed sensor is not reported in 1st and 5th gears. The N2 input speed sensor is not reported in Reverse. The Input Speed Sensor Overspeed is a rationality check that is intended to indicate a major transmission failure and will cause a loss of drive, with transmission going to neutral."
I'm now trying to find out which speed sensors are triggered in Neutral.
 

bwilliams60

Joined Nov 18, 2012
1,442
What you are asking to do is near impossible. The sensors etc are the easy part. The hard part is duplicating load factors, gear shift patterns, temperature ranges etc for every driving condition and exercise that your automatic transmission would come across. Why not just go out and get an ECM for a manual transmission and run with it. Best of luck to you. This is a major undertaking.
 

Thread Starter

pt500

Joined Apr 6, 2018
8
Thank you for the input! Out of curiosity, what makes you think of load factors and gear shift patterns? As far as I can tell, the TCM just verifies that every solenoid is working properly and that the other sensor inputs are correct? The car doesn’t even send the speed readings from the sensors in the trans anywhere but the TCM to verify that the transmission is functioning properly, instrument cluster speed is taken from wheel sensors. Or is there way more to it than that, not apparent from the diagrams? I totally agree with your suggestion of a manual ECM, unfortunately it doesn’t exist for this platform :/
 

shortbus

Joined Sep 30, 2009
10,045
Out of curiosity, what makes you think of load factors and gear shift patterns?
Those are what is used to control things like up shift and down shift of the transmission. Comparing throttle opening and engine speed to gear selection. Even though the transmission only "reports" to the TCM, both it and the ECM "talk" back and forth.
 

bwilliams60

Joined Nov 18, 2012
1,442
Thank you for the input! Out of curiosity, what makes you think of load factors and gear shift patterns? :/
About 40 years as an automotive/Truck and Coach technician/Professor.
In a mechanical transmission, we used road speed to create governor pressure and modulator valves to assist governor pressure to actuate shift valves. Modulator valves are a measure of load placed on the transmission.
Like I said, I applaud you for what you are trying to do but there are a lot of moving parts in this equation and I don't see how you are going to be able to recreate all of the variables that occur between the ECM and TCM. I have spent a lot of time on the electrical/electronic side of cars and have built many test benches for duplication and simulating sensors and temperatures/pressure is the easy part. It's what happens when the thing starts to move and everything is varying that is the hard part to recreate. This is all done nowadays by a microprocessor and that my friend, is where the fun begins. If you can do it, please send me your creation so I can build a test bench to emulate it.
As for your thoughts on the TCM, like Shortbus says, that is just one piece of a much bigger world. ECM, TCM, ABS etc all communicate together to make things happen in your vehicle. We have the easy part. Just drive it.
The best you can hope to do is put in your manual transmission and start removing codes one by one but I think you are going to be unhappy when you can't put that light out and you have to get a safety or E-Test. Again, I wish you luck.
 

shortbus

Joined Sep 30, 2009
10,045
I would ask, if the car doesn't support a manual transmission, what is the transmission out of? Maybe, just maybe using the TCM that worked with that vehicle could be made to talk to the ECM?
 

bwilliams60

Joined Nov 18, 2012
1,442
The TCM is going to be looking for speed inputs from the input shaft, intermediate shaft and output shaft. It will also be looking for governor pressure, modulator pressure and line pressure as well as temperatures.
There is no problem getting a TCM to talk to the ECM. The problem comes with getting the right information, at very precise moments to keep the idiot light off. I am not a programmer but I would think that rewriting the EEPROM to not look for an automatic transmission would be easier than trying to go the other way?
 

Tonyr1084

Joined Sep 24, 2015
7,852
For the sake of argument, and I'm no expert here, suppose in an automatic transmission you're accelerating from a red light. Going easy, not trying to set any land speed records. The TCM tells the ECM what gear it's in and how much torque is needed. The ECM will respond by applying the appropriate fuel mixture for the amount of air flow going into the engine. Now, suppose you stomp on it. Suddenly the ECM has to provide more fuel. If the ECM is thinking you're driving easy and you're out to climb that steep hill and accelerate from zero to 60 in 5 seconds then you're going to starve your engine for fuel. That's what others are trying to tell you. The whole thing is a balancing act. The transmission needs to down shift to accommodate the driving needs. You're going to be shifting when you need to, but the ECM isn't going to know that it's you that's shifting the vehicle.

If you really want to eliminate codes then remove the ECM and go with a carbureted engine. Go old school. You'll cause more pollution while doing so, but you won't have those codes. Now, how much of the ECM is needed for the rest of the vehicle function - that I don't know.
 

shortbus

Joined Sep 30, 2009
10,045
If you really want to eliminate codes then remove the ECM and go with a carbureted engine. Go old school. You'll cause more pollution while doing so, but you won't have those codes. Now, how much of the ECM is needed for the rest of the vehicle function - that I don't know.
Not in this day and age. There are 'stand alone' EFI and digital ignitions out there. There are companies that build/make special control units to put motors and transmissions in vehicles they don't belong in. Others that can flash the controllers to fit two different makers parts work together. But since it took them time and money to get things to "work" it costs to be able to do it. You just need to figure out which is better for you, pay them and drive or take the time to do it yourself.
 

Thread Starter

pt500

Joined Apr 6, 2018
8
I do understand what y'all are saying. Not to be contradictory, but I do know it's possible to drive the car with it thinking it's in neutral. Engine load is determined through the dozen sensors in the engine itself and via the ECM. The resistors in place of the shift solenoids just need to mimic the transmission being present and in neutral, not actually shifting gears. The current holdup are the two hall effect sensors, and I'm still looking into which sensor is read when the car is in neutral. The actual engine computing is handled by the ECM and the rest of the car. If I can't figure out how to trick the hall effect sensors, I'm thinking possibly tapping the rpm signal between the engine and ECM and feeding it into where those sensors would normally be. If I can figure data on the ratio between whatever is spinning in the transmission and what the engine rpm is, could I feed that engine rpm signal into some kind of microcontroller to modify the signal? Let's say the engine is seeing 1000 rpm at the crank and the sun gear read by the trans sensor is supposed to see 500rpm, how complex would modifying that signal be? I've never messed with microcontrollers. Also attached is a link to a YouTube video showing what I believe is the world's first manual swapped Mercedes of my generation. The guy who did it says the car thinks it's in neutral, and drives fine. I messaged him for details on the wiring but he said he forgot and has no notes on the swap. I'm not sure if he's holding out on me or genuinely forgot, but he wasn't any help haha.
 

Alec_t

Joined Sep 17, 2013
14,280
Even if you could fake the Hall (and any other) sensor signals, won't the TCU and the ECU be talking via the CAN bus? If so, won't you need to know the (probably proprietary) protocol used?
 

Thread Starter

pt500

Joined Apr 6, 2018
8
So looking at the pins on the TCM, my first post second photo, it seems that there is only analog communication between the Transmission itself and the TCM. Now, there does appear to be CAN communication between the TCM and the automatic shift lever. I don't have a schematic for that, may just have to pull one apart and cannibalize it to hook up in conjunction with my trans spoofer.
 

bwilliams60

Joined Nov 18, 2012
1,442
I believe the signal from the transmission shift lever would be for the neutral safety switch or back-up lights? I would be more inclined to sniff the CAN BUS and see what code has to be changed to simulate park and neutral. Same could be true of the solenoids. (If I were building a simulator). Neutral is easy to simulate. You just need to know what solenoids are activated in neutral and that is what the TCM will send the ECM as information. Keep in mind that in neutral, it will only see a very light load so may be under-powered.
The Hall-Effect sensors are digital signals which are speed based and this is where I said you Can't fool it. It is looking for shaft speeds and unless you plot a graph of the current shaft speeds and write a simulation program, you are not going to fool anything. It probably can be done but you would really have to know what you are doing. Keep in mind that all the time I am talking, I am basing my opinions on building a simulator to replicate your scenario.
 

shortbus

Joined Sep 30, 2009
10,045
What manual transmission is being used? The Mercedes G56? Or something else? The G56 is also used in Dodge trucks and some Jeeps so if that is what your transmission is couldn't the TCM from one of those be used?
 

Thread Starter

pt500

Joined Apr 6, 2018
8
As of right now the only transmission I know bolts up 100% is the Chrysler Crossfire NSG370. The bell housing pattern, transmission cross member, and driveshaft all fit my platform as well as many other Mercedes that were delivered with the 5G-Tronic. I'm working on possible other crossover transmissions, but nothing confirmed so far.
 

shortbus

Joined Sep 30, 2009
10,045
I'd look at getting a new ECM that was used for a manual tranny and that engine. Automatics aren't as popular in Europe so they should be able to be bought from there through Mercedes. What are you going to do for all of the other stuff involved, flywheel, clutch pedals and linkage etc?

Have you checked with any of the engine swap places? There are some companies making stand alone modules for stuff like this. Painless is one of them, or maybe Jaguars that Run.
 
Top