[jpinto3912]

I've done a small controller that performs auto start-stop of the engine to save fuel. I've installed it on my BMW 320d and Ford Focus with no issues and while it's not a simple plug-in add-on, it's certainly doable for any DIY enthusiast.

I own a 220cdi 6-manual, and I'm just finishing up retrieving signal info to get it done. I don't have wis/ASRA, just an outdated Autodata v3.

Can anyone point out the pin and colors of the VSTAT_A signal from the SBC? I could go after vehicle speed info by looking at one of the wheel sensors at the ESP module, but this VSTAT_A signal is much more convenient.

If you're interested, I can elaborate on it works, which signals are monitored, where to actuate, whats the control rules/strategy...

 

[Diesel Benz]

Your projects sound a bit scary to me, interesting though. I hope you are implementing all safety features.

The VSTAT_A signal must be at the CAN-bus, you would need a CAN-BUS monitor circuit with proper software to read it I'm afraid.

I guess you are going to use the starter motor to start the engine? The OEM stop-start works differently, no stress on the starter motor and the engine fires up instantly too (also a CDI).

Welcome to the forum!

 

[Roadking999]

no pain, no gain
I thought the project to start the car by placing the crank at a specified spot BTDC failed and they went back to using the starter!! But may be I'm thinking of another OEM.

 

[jpinto3912]

Most (but not all) OEM solution are a bit different then what I do in two items:
Belt-driven Generator/Motor electronically controlled and an AGM battery

The first replaces the alternator and starter, and allows finer control of where to stop the engine, how much cranking power is required, how much current is needed to recharge the battery. I saw Valeo's and Bosch solution and both are based on a brushless 3phase 4 pole stator and a rare-earth magnets rotor. They are, simplifying, a 3X power design of what an electric bike uses.
BMW start-stop on the E90 passed on that and uses what they call a "beefed-up" starter and the exact same alternator I had on the E46. I'm yet to understand how much beef is in there... my E46 starter brushes gave up at 350kKm, a good 11yrs of service.
On the E90 vehicles with auto start-stop use AGM iso Flooded batteries. I know AGM for start-stop have lower cranking power, so they up'ed the capacity, and AGM copes better with high-current charging (results in fewer electrolysis). However, an AGM is double the cost of an equally rated Flooded one. So, my lead is, if you auto start-stop around 20x a day (10 traffic lights/busy roundabouts/intersections, and back), be prepared to renew your FLOODED battery every 3yrs"... it's still more economic that AGM. You read RENEW, which is different from BUY ANOTHER.

So, to sum-it up, on merc's and Bimmers, I'm sticking to the beef the starters have. On the E46 I replaced the brushes and the calling-solenoid for less than 40€ (at 350kkm)... the only nuisance really is the extraction and putting back... the W211 is also not friendly at all.

On safety:
I monitor clutch pedal, battery voltage, engine and vehicle speed.
Nothing gets shut-down unless 2secs of stable conditions are met {engine running & vehicle stopped & clutch up}.
Nothing turns on unless{engine stopped & clutch down & was auto-stopped}.
The wiring and relays are done in such a way that if you disconnect the module, your car works as before.
The module can be disabled/enabled by pressing the clutch 4 times in the first 6secs after power-up.
I monitor battery voltage and take notice of how it looks in {engine stopped & not cranking}... if under 12.5V, during this trip I won't auto-stop it again (I will start-it this one more time).
I keep track of whether it was the module that stopped the engine or not and whether the car has moved since the last auto-start... to deal with issues such as putting your key in pressing the clutch and it would start (bad); letting go of the clutch too fast and killing the engine (bad), 1st-rev-1st maneuvers where the care barely moves, letting the engine on just to keep the a/c on, etc.

The VSTAT_A idea came from a Mercedes diagram explaining the SBC, and it was separate wire from the CAN-bus... so it seams to be a discrete signal, not CAN messaging. Check http://bayhas.com/mercedes/w211/contents/IMAGES/sbc/sbc_can.gif

 

[Roadking999]

Nice writeup JP, just one question, what is the reaction time for restarting? I just drove a X3( F25) Beamer and it was dazzling quick.

 

[Diesel Benz]

Indeed nice write-up, I feel much better now that the OP has done a lot of homework and has skills for the task and takes safety issues seriously.

I had another look at the wiring diagrams and there are several wires between different control units that have not been names. One group of four wires between the SBC unit and the front SAM might include the VSTAT_A signal. I wonder if some other document had the wires named, otherwise it would need some measurements.

JP, I could print the WIS documents if you really feel you do not need the complete WIS for your task.

The same wiring info is included in this training document, but still without exact names for each wire (WIS at least shows the pins at the connectors).
http://www.mercedestechstore.com/pdfs/W211_advisor/219 HO SBC (WJB) 9-30-02.pdf

 

[jpinto3912]

On both cars I did, the E46 and a 99 Focus Tddi, the time it takes to press on the clutch and gear in 1st, is the time it takes to get it cranked.
You don't have to wait for anything, unless if you're in red-light-race mode.

On the W211, from manual experience, unless your injectors are shot, it fires up at the 4th full revolution, which roughly makes 0.5secs. I know common rails take a bit more time to create pressure (higher value), so I'm keeping modest hopes that 0.5sec will always be the mark.

BTW, if your engine sometimes cranks for several seconds and then starts with a big puff of unburned oil, that's most likely injectors with too much return, not allowing enough pressure build up.

 

[jpinto3912]

Diesel Benz,
Thanks for the PDF, nice info there! It's surely he "Vehicle Speed Status" signal no the SAM N10.
On Wis, can you make out which pin/wire color is that one on either the SBC or the N10 (btw, which module is N10?). I might be able to work it our with which wire colors are those 4 (from SBC to N10)... but it would be best to be 100% sure.

The weekend is coming, if my kids allow I might get some work done for this.

 

[isstay]

This is some info on connectors wiring between A7/3 (Traction system hydraulic unit) and A47/5 (ESP control module) and N10/1 (Driver-side SAM control module with fuse and relay module), I believe that this can be the info you need. Hope it helps.

 

[jpinto3912]

isstay, thank you!
So VSTAT_A seams to be pin 4 of A7, "Operating mode signal output". That leads to pin 7 of M5 on N10.

A7 is a part of SBC module, right? It's a bit of routing from there to get near the ECU...
The M5 connector of the N10 is just there near the ECU, right?

 

[isstay]

isstay, thank you!
So VSTAT_A seams to be pin 4 of A7, "Operating mode signal output". That leads to pin 7 of M5 on N10.

A7 is a part of SBC module, right? It's a bit of routing from there to get near the ECU...
The M5 connector of the N10 is just there near the ECU, right?

Yes, N10 is the driver side SAM sitting on top of the left front fender and feeding power to different components. ECU also plugs directly in to front SAM or N10. A7/3 is the High pressure pump of SBC unit mentioned as part of the Traction Control. Let me know if I can help w/something else and I will try to dig it out.

 

[jpinto3912]

isstay,
Can you check on WIS where does the EIS (Electronic Ignition Switch) gets its power?
Autodata only shows the VAN and CAN pins. If I can do without a CAN and/or VAN driver, the better.
The EIS is a bit buried, so it would be great if it comes from the N10...

I want to test if it's enough and good practice to cut power to EIS to emulate a key-off situation, without tripping any codes... this being an EIS, might be not as simple. I have plan B.... and C...

I'll also test if getting EIS powered again, and the key already at position 1, does the ECU sequence injection if it gets crank+cam sigs.

On "regular" key-switches it's a breeze... I hope the EIS doesn't require me to "black-op" the CAN/VAN.

 

[isstay]

isstay,
Can you check on WIS where does the EIS (Electronic Ignition Switch) gets its power?
Autodata only shows the VAN and CAN pins. If I can do without a CAN and/or VAN driver, the better.
The EIS is a bit buried, so it would be great if it comes from the N10...

I want to test if it's enough and good practice to cut power to EIS to emulate a key-off situation, without tripping any codes... this being an EIS, might be not as simple. I have plan B.... and C...

I'll also test if getting EIS powered again, and the key already at position 1, does the ECU sequence injection if it gets crank+cam sigs.

On "regular" key-switches it's a breeze... I hope the EIS doesn't require me to "black-op" the CAN/VAN.

Here is the info I have:

 

[jpinto3912]

Isstay, thank you once more!

It seams I can take power from outgoing N73-C-2, just before it reaches N10 C1-3.
I use about 25mA, N73 should source that without issues... but I might play it safe by biasing a transistor to take power from N10 C1-12 (colored RD, right?), driven by N73-C-2 (colored PKRD?) signal.
This would give me an indication of whether the key is >0.

I'll test if interrupting N10-C1-3 (coming N73-C-2) turns off the engine (simulating pos 0). In the actual montage, this would be done with a normally closed relay.

If the above works, a second relay would Y switch from 12V (from N10 C1-12) or N73-C-3 (colored PKYE?), relay normally at N73, toward N10 C1-10 to energize the calling-solenoid of the starter.

If that all works, I'm all set. None of these signals will require high current driving (judging from the 0.5 section), which makes wonder if this merc version could be transistors only (no relays). The upside of relays is that I can disconnect the module and the car is working as original... the downside is that relays have an MTBF of 1000 to 100000 of cycles... Any thoughts?

I can't have "play" time today... hope that tomorrow I can make some progress.

 

[jpinto3912]

Position 2 is not what I think it was... and there's no simple signal for the start position (position 3). Key 2 it's the position where the key is left when the engine is running. I'm able to shut the engine off interrupting N73-C-2, so I'll have to energize the starter via a relay (it's what's done in my other two start-stop mods; I was hoping that the EIS would allow simpler ways).

I need to work backwards from the starter. Could anyone point me the wire and connector coordinate that comes from the starter-relay at N10?

I need to test that by simulating key2->key1->key2, if then I energize the starter the ECU will make injection cycles... just as a rolling start (from N73/N10/ECU none gave the starter any order).

 

[isstay]

N10/1 connects to the starter motor M1 through fuse 58.

 

[jpinto3912]

This EIS is a pain! Cranking up is easy, but getting the EIS-ECU to unlock its ass and perform engine control is a bitch!

Interrupting Key position 2 will not shut the engine off... it will however prevent that at pos 3 the ECU commands cranking.

Fiddling with Key pos 1 signal coming from EIS does engine shut-off. However, after de-interrupting that signal, forcing the starter will not correctly fire the engine (ECU seams locked).

So, I abandoned that Plan and went with the next simplest approach: crank sensor.
I simply shorted out the two crank sensor signals (don't ground or 12V them, the ECU circuitry may not be that well protected). That caused the motor to shut-off (duh). However, and again, the ECU locked-up and energizing the starter does not get it fired-up.

The scope tells me they're biasing the crankshaft sensor differentially, with a constant current sinker on one end and sourcing it through a simple resistor:
- on both, DC is quite low, <2V.
- both signals present the same AC component (2-3Vpp, depends on rpms) but in opposition
- there is a constant DC difference when there is no sensor activity.
The coil in the sensor is ~600-700Ohm.
Autodata lists wiring/sensor errors on this sensor, so it's a given the ECU uses the DC value to diagnose them.
I think the shorting trips the error detection, which then locks the ECU requiring a full key-operation cycle.

CONCLUSION OF THIS RUN:
Plan a: fiddle with the key-pos N73-N10 signals is NO-GO.
Plan b: shorting the crank-sensor kills the engine, trips ECU errors.

Plan b2: maybe I can switch between crankshaft sensor and a 680ohm resistor, causing shut-off, without errors.
I'll do a trial run before I try this, because I'm suspecting the ECU will only do non-initiated firing sequence if the car is not stationary. This requires a bit of wiring setup to be able to drive and perform the external signal play... so It'll be while.

Plan d: get inside the N73 and emulate the key-positions, if detected via micro-switches (not in the least inclined to do this). I've seen some N73 pictures on the web, and getting my hands on a stacked-pcb electro-mechanical module with loads of thin wires would surely put make this mod "not-DIY". Besides, can anyone point a write-up on EIS extraction procedure?

 

[yiannossavva]

Eny news about this project?

 

[jpinto3912]

Unfortunately no. I quit developing a solution for the W211. Right after this post I had pressing problems (busted bolt on the pulley tensioner) and later on I battled never ending noises on front suspension, excessive oil consumption, etc... it took all my amateur-shop time to solve.
I sold the car mid 2015 due to my third child and wanting a station-wagon trunk.
I now own a auto-gearbox which makes start-stop retrofitting even more challenging (starter-lockout if not in P or N, for instance), and family time is all the time left I have.
Luckily this car (Audi A6) only needs a brief flicker of the key and it manages the starter-engage time, so it's pretty easy to manually start-stop it.

 

[CetiAlpha5]

Plus if you get a 2014+ model, I believe it comes standard with the eco mode which is hated by some people in the W212 forum as you can't permanently shut it off. Always comes back on once you shut the car off, you have to hit the off button each time you start the car to disable it.