SMG Gearboxes

Ok folks I've created this topic as a discussion point for the problems I've been having with SMG gearboxes. A brief summary of my problem:

I bought the car in August 2006, and everything was fine for the first while. Quite a few laps of the ring since then, but I'm not sure exactly when it first started happening, I think it was early 2008 because I first noticed it at the ring - crunching from 3rd to 4th. A good example on video was at Spa in May 2008: (http://www.vimeo.com/1093053).

The only relevant work to the car was the clutch (and flywheel?) were replaced in April 2007 by a German dealer, as the car was slipping and eventually died just outside the ring entrance, refusing to go into gears

I replaced the gearbox in June 2008, where we also put a race flywheel and clutch in as well It was all better for about a month or so, and then it started happening again

After much discussion with my local dealer, I eventually replaced the actuator in April and reinitialised all the software (including reflashing the SMG ECU), which on the road seemed better but as soon as I took it to knockhill it started again. So, annoyed by now, I replaced the whole gearbox AND hydraulic unit, thereby a process of elimination I'd replaced everything SMG gearbox wise with brand new parts. Only the pipes between the hydraulic unit, and the SMG ECU itself were the existing items. For information, I also think the gearbox temperature sensor wasn't replaced, but I'm going to try and find that out.

Anyway as most will know from my most recent videos, the problem started happening again (at my only trip to the ring this year, is that a coincidence?) and since then has got much worse. I've not driven as much on track this year, but I have done a few races. However I am not convinced the stress on the gearbox/SMG is any worse than the last few years, if anything its had a lighter workload.

I'm now in the process of working out what could be causing my problems. I think the gearbox needs fixed now, the syncro will be screwed, but I want to fix the source of the problem before repairing the box.

I've found a great resource online: http://www.leo.nutz.de/docs/bmw/smg_training_manual.pdf

From that I've worked out the following. I'd appreciate any comments or thoughts to help me!

----

The standard basic gearbox (S6S420G), as used in the M3, is transformed into a sequential gearbox with the addition of:
- a hydraulic unit (item 1 below - part number 21532229715)


- shift assembly/actuator (item 1 on the diagram below - 23412229789)

- a special clutch slave cylinder (item 8 on the diagram below - 1522229841)

- a new shift-lever module (item 1 on the diagram below - 25122283852)

The SMG II Control Unit, installed in the E-Box next to the DME, is a single board module with SKE (134 pin) gray colored connectors (2360783772)

Communication with the DME is via a dedicated CAN bus (SMG II CAN Bus) - could part of the problem be this communication wire(s)?

Based on instructions received from the DME the SMG II control unit manages the clutch solenoid valve (proportional directional valve), the selector shaft up and down solenoids (pressure control valves), and the selection angle solenoid (proportional directional valve). 4 solenoids in total, all on the hydraulic unit and therefore all brand new.

While the SMG II control unit receives many of the inputs and manages the control of the hydraulic system, the DME is responsible for and controls all gearshifts. Sensor inputs received by the SMG II control unit are relayed to the DME for processing and monitoring. I think this rules out the DME as it's evident with my problem that its trying to change gear, it's just failing whilst doing so.



Problem elimination / narrowing down:
- Score out the keylock, nothing to do with that.
- The Canbus between the SMG and DME - could that be faulty?

Signals exchanged via SMG CAN Bus:
Engine Speed: The engine speed signal is transmitted twice to the SMG II control unit. One signal arrives via the SMG CAN, the other arrives via a hard wire from the DME.
Engine Coolant Temp, Engine Oil Temp and Intake Air Temp:The signals are inputs directly into the DME and are shared with the SMG II control unit as the information is needed over the SMG II CAN.
PWG: The PWG signal is input into the DME and forwarded to the SMG II control unit. The information is used to calculate engine load.
EDR Feedback Pots: These potentiometers provide throttle position information that is useful during slip intervention.
Wheel Speed: Conditioned digital wheel speed information is received from the DSC control unit. The SMG II control unit uses data from all 4 wheels to detect vehicle speed and wheel slip in the A-Mode and during down shifts in all modes to detect drag torque induced wheel slip. I think this can probably be scored out as it mainly relates to the A mode, but relates to down shifts in S modes as well. The original problem I had was during upshift from 3rd to 4th, so I'm presuming this can be scored out.
Transverse Acceleration: Transverse Acceleration data is transferred from the DSC control unit in the A-Mode so that up and down shifts may be prevented during high speed cornering. This sensor information is further evaluated for slip recognition purposes. I think this can be scored out as it relates to automatic gear changes i A mode, I use S (manual) mode all the time.
Steering Angle: This data from the steering angle sensor is sent via the CAN Bus for cornering and slip control information.
Cruise Status: The cruise mode is deactivated when a driver initiated shift is made in the S-Mode, during A-Mode operation, the cruise control setting is maintained. This is scored out as the CSL doesn't have cruise, and we're talking about S mode.
Parking Brake: Information concerning parking brake application affects gear engagement and vehicle operation. I suppose its possible that a faulty sensor is telling the system that the parking brake is engaged, stopping it engaging gear, but the chances of it only happening occassionally and only on 3rd to 4th I think are very unlikely so I'm scoring this out.
Door Contacts: The hall effect door position sensors integrated into the door latch mechanism signal the GM. This data is transferred via the K-Bus and CAN Bus for safety program initiation. Scoring this out as this is only for safety telling if the door is open or not. I have seen the gear display flash because the door is ajar, but it still changes gear OK regardless.
Brake Light Switch: Brake pedal status is transferred to the SMG II control unit via the DME. The signal is used for:
- Unlocking the shift lock
- Brake detection
- Engine Starting
- Downhill detection
- Clutch Disengagement while stopped.
Scoring this out as I'm not braking when changing from 3rd to 4th!!
- Score out the MFL shift up and down, don't think thats the problem (although I did have problems at Brands at Donnington where it wouldn't respond to the paddles and I had to use the gear lever - seperate problem or related?)
- 8 x shift lever hall sensors - what are these? Have they been replaced?
Located in the shift lever module, 8 hall sensors detect upshift and downshift requests and are also used for Selector Lever position indication (Forward, Reverse and Neutral). Signals from the hall sensors are also used to initiate mode changes from S-Mode to A-Mode and back.
- Score out drivelogic control
This is the switch located on the console just to the rear of the shift lever, scoring it out as its only a switch telling the computer which program to choose for the speed of gear change.
- Longitudinal acceleration sensor, not convinced its anything to do with this, but was one of my errors to do with this before?
The signals from the Longitudinal Acceleration Sensor which is mounted under the passenger seat are used for uphill driving programs.
Scoring this out as I think it's only relevant to the A mode, for uphill information.Very unlikely to be relevant unless the 3rd to 4th changes were only ever uphill, which they weren't!!
- General Module (I think this is the SMG module) - not been replaced, could this be the problem?
Load deactivation information (Sleep) is received by the SMG II Control Unit. Not sure what this means?
- Transmission input speed sensor - has this been replaced?
A hall sensor is used to measure transmission input shaft speed.
- Potentiometer Clutch - what is this? has it been replaced?
Input from the PLCD providing exact clutch position. The PLCD has been replaced so I'm scoring this out.
- 2 x Selection angle cylinder POTs - think these are part of the actuator and therefore replaced
The Position sensor provides two inputs (PS2 is the selection angle sensor, a dual pot) for position recognition of the main selector shaft in the transmission.
- 2 x Shift travel cylinder - think these are part of the actuator and therefore replaced
The Position sensor provides two inputs (PS3 is the shift travel sensor, a dual pot) for position recognition of the main selector shaft in the transmission.
- Transmission temperature - replaced?
An NTC sensor is used to measure Gearbox oil temperature.
- Hydraulic temperature - replaced? (number 7 on the diagram below)
An NTC sensor is used to measure hydraulic fluid temperature.
- Hydraulic pressure - replaced? (number 8 on the diagram below)
A Pressure transducer for monitoring hydraulic pressure sends pressure information to the SMG II Control Unit.
- Hood microswitches - score them out not the problem although they have caused problems in the past, slamming the bonnet down solves it
Two Micro switches located in the hood latch mechanism provide a ground signal to the SMG II Control Unit when the hood is closed. Richard from Thorney's had mentioned about getting a better ground signal, related?
- TD Signal Engine RPM comes from the DME, I presume this is not the problem.
Calculated engine speed received from DME via hardwire serves as redundant information with that received via the SMG CAN Bus. Could it be relevant?
- Score out the shift lock, its not the problem.
The Shift Lock Mechanism locks the shift lever in position for parking and prevents
unintended gear changes as part of the safety programming.
- What is the EWS module, could this be part of the problem?
During the start-up operation, the SMG II Control Unit confirms shift lever position (0) and brake application and signals the EWS module which then allows starter engagement. Scoring it out as the problem is not about starting the car!
- Hydraulic pump, which has been replaced. Could it be the salmon coloured relay that keeps the pressure in the hydraulic pump between 45 and 80bar? Has this relay been changed? I should carry spares of this relay.
The Hydraulic Unit is energized through a relay controlled by the SMG II Control Unit. The operation of four (4) solenoid valves located in the hydraulic unit that are used to actuate the clutch and shift the transmission are also controlled by the SMG II.
Detailed diagram of the hydraulic unit:



1. Hydraulic Line for Shift Travel 1 Gears 1,3,5.
2. Hydraulic Line Slave Cylinder Clutch Control.
3. Hydraulic Line for Selector Angle Control Cylinder.
4. Hydraulic Line for Shift Travel 2 Gears 2,4,6.
5. Pressure Accumulator.
6. Supply Line from Hydraulic Tank.
7. Hydraulic Fluid Temp Sensor
8. Hydraulic Fluid Pressure Sensor.
9. X 5330 18 Pin connector.


Just noticed on the detailed diagram that there is a seperate hydraulic line for gears 2, 4 and 6. Tom from Thorneys said he couldn't get the autologic machine to go into those gears. Coincidence? According to the diagram below



one line moves it into 1st, 3rd or 5th, and the other releases pressure(?) to move it the opposite way into 2nd, 4th or 6th.

Were the pipes from the hydraulic unit to the gearbox replaced? I need to check but I don't think so. Could they need cleared out?

page 27: "The forces required for shifting are strongly dependent on gearbox-fluid temperature. The fluid pressures acting on the pistons of the two shift cylinders are adapted according to fluid temperature."
I dont think the gearbox temperature sensor was changed? Perhaps the gearbox oil is getting too hot? Why are other CSLs not having this problem though?

[DuncanR]

Steve, me being an expert on CSL's having owned one since Friday just gone ! Im probably of little use !, however,I do mess about with hydraulics from time to time, reading through your problems you dont say if you replaced the oil cooler (have assumed they have one) or if you flushed the matrix out, it could be that its bunged up with sludge, and not working as efficiently, giving a rise to increase temperature. The double wammy could be if the temp sensor is also U/S. Just trying to assist with a bit of logic if i can !

Originally Posted by DuncanR

Steve, me being an expert on CSL's having owned one since Friday just gone ! Im probably of little use !, however,I do mess about with hydraulics from time to time, reading through your problems you dont say if you replaced the oil cooler (have assumed they have one) or if you flushed the matrix out, it could be that its bunged up with sludge, and not working as efficiently, giving a rise to increase temperature. The double wammy could be if the temp sensor is also U/S. Just trying to assist with a bit of logic if i can !

Exactly what I'm looking for Duncan, so thanks for the reply. I'm too close to the situation, can't see the wood from the trees, and my techy knowledge is worse than my driving (not saying much ).

There is no oil cooler that I'm aware of - either for the SMG oil or the gearbox oil. Certainly not as standard.

Not sure what matrix you're talking about?

I'll look into the sensor, it might not be helping, perhaps reporting temperature as OK when its actually warmer and requires more SMG pressure to the actuator?

[shimmy]

Steve

i have a video of my CSL gearbox doing exactly the same thing in 2007 at Donington. Caught it on film 3 times in one trackday missing 4th gear from 3rd only

BMW Murketts saw the film, checked the gearbox and said it was just starting to wear out the synchro and gave me a new SMG gearbox on goodwill

no problems ever since.

[DuncanR]

If it had a cooler, the matrix is the sort of element inside similar to a rad, and sometimes they need flushing out. Im surprised that this gearbox doesnt have one given the stick it takes !, if it had one, it would at least help to keep the oil temp in some sort of limited band of temperatrures, good luck mate !

[AlexGTT]

Steve, a faulty gearbox oil temp sender sounds possible, as does the pump relay you mention. Is there any way a diagnostic reader could isolate faults with those individual components? Would be interesting to see what codes have been thrown up when the diagnostics are checked. It's worth a swop of those components, individually to accurately diagnose. Probably the cheapest items to change on the whole SMG system!

[NZ_M3]

Pump relay's related to the start up procedure and keeping the hydraulic unit at operating PSI (i.e. the noise you hear when you first open the door if it's not). Just open your fuse box and check that you have the updated 'salmon' colour relay.

What clutch are you running? I know a lot of the US guys have problems running aftermarket clutch paks in SMG cars and even more so when they go force induction. This could be your problem (as the SMG control unit is programmed for a certain set parameter for the OEM clutch pak, so if your pressure plate has double the clamping load it could mean that it's not disengaging the clutch properly before it shifts gear, so over time you'd get premature synchro wear and eventually the gear selection issues)

Check that all the hydraulic hardlines are kink free - a tech may have inadvertently bent a hydraulic line when doing a pump install or even from pulling the gearbox out (I've seen it happen before).

Definitely also sounds to me like the gearbox is getting too hot - these SMG gearboxes really need a cooler unit of some sort in my opinion.

Another possibility - broken compression spring (part #6) - see post number #32 http://www.m3forum.net/m3forum/showt...44#post3497344



Also check that item 10 below has been changed - this is unique to the SMG gearbox.




One more thing .. no doubt you've read this post on m3forums already? Quite useful for trouble shooting

http://www.m3forum.net/m3forum/showthread.php?t=231113

[The Gorilla]

Hi,

I have read many views and opinions on the SMG Gearbox
and reasons as to what causes the issues.

Lighter flywheels do not work well with SMG as the inertia
is altered and it can cause puluse issues with the software.

No such issues on a manual car, same gearbox internally.

My own view is that the SAC Clutch and being of poor design
is at the start of the SMG woes.

The SAC was designed so that clutch take up should be about
the same as the clutch wears.
This should be ideal for an SMG car as it should produce the same
clutch shift pattern time and time again.
Something that actuators and the hydraulics function best with.

However, once the SAC incurrs wear, its clutch travel can alter
then you have accuators and even more importantly hydraulics
which can be a fraction out of sequence which can bring on
issues such as clutch 'slur' and delays before engaging either up
or down in the shift pattern.

The Software can only react to the input signals it receives, rpm, etc.

So once you have hydraulics that are a fraction out of sequence
then you will incur increased wear in the gearbox syncro rings, and
this will then lead to pump fatigue, as the selection process
is being forced as opposed engaged.

On a normal road car, its progressive over a period depending on use etc,
but on a Track / Race car, given the extra shock and load that is pushed
through the drivetrain, the attrition rate and wear is much faster.

I do not believe that there is a way to ensure that the Clutch Travel
is perfected time and time again, on a SMG car,
and thus these problems on Cars that see hard use
will be a by product of that use.

Harder Clucth disc, like Cerametallic as opposed Organic helps to slow
down the wear rate, but the shift changes become more agressive
for a Car that is driven on the raod as well.

Just my thoughts.

Regards,

The Gorilla.

[The Gorilla]

Hi,

I should have added that gearboxes, SMG included
suffer most wear and at most
risk of damage etc when the clutch is in, not out.

Regards,

The Gorilla.

[shimmy]

Originally Posted by The Gorilla

Hi,

I should have added that gearboxes, SMG included
suffer most wear and at most
risk of damage etc when the clutch is in, not out.

Regards,



The Gorilla.

would getting the gearbox reprogrammed/retaught in a regular basis help