BMW M5 Forum and M6 Forums banner

1 - 8 of 8 Posts

·
Registered
Joined
·
464 Posts
Discussion Starter #1 (Edited)
My M6 had an issue with clutch engagement. It was abrupt and clunky when cold, and threw an occasional 4FA0 (clutch position) code when driving before the transmission was fully warmed up. While still cold and starting from a dead-stop, the clutch control hunted around (surging vehicle speed and engine rpm) before full engagement. After warmup, it engaged better (still less than smooth) and shifted normally.

I started carrying my laptop in the car so I could run diagnostics when the problem occurred. Every time I was able to do the DIS tests, all the clutch parameters came back normal. My clutch, TOB, and guide sleeve are all fairly new, so those didn’t seem likely. I talked it over with my indie, he had seen this issue in an M5 once and it ended up being the clutch position valve. The manual he had from BMW SMG training showed this valve (item 16 in parts diagram) in the circuit to the clutch slave cylinder. Knowing this was the likeliest cause; I decided to take a gamble and had him order the part. I was a bit shocked to find out that this valve was over $600 (wholesale price), but went ahead with the job.

P/N ordered was 23017845630 (Directional Control Valve). Note: It really helps having two people do this job to help support the pump and realign things when reassembling.

We had to depressurized the SMG hydraulics (using DIS), then disconnect the battery, and lower the back of the transmission (after removing the engine skid plate and everything needed to get the driveshaft out) to get access to the allen-head fasteners on the side of the hydraulic pump. After removing the pump fasteners, the banjo bolt on the pump outlet pipe to the clutch slave had to be removed to allow the pump to be lowered enough to reach the solenoid valve. I supported the pump from hanging on the flex hoses, and wiring while Pierre removed the solenoid. The o-ring on the solenoid sits in the pump block and is tough to keep in position while pushing the new valve in. The last big issue is keeping the crush ring on the banjo bolt while reattaching the clutch slave pipe. The pump reservoir was full from bleeding down pressure before the job started, so the DIS bleed process was run first (about 15 minute process). The reservoir was topped off and everything checked for leaks before putting everything else back on the car and running the clutch adaptations.

After starting off I could immediately feel the difference in clutch action. It was very smooth and progressive (like never before). It felt like a skilled driver was operating the clutch! The next day after letting the car completely cool off, it was the same smooth clutch action. For me, this was money well-spent.
 

Attachments

·
Registered
Joined
·
5,673 Posts
I have long suspected this valve was key in shift smoothness. It's sort of the gateway for all fluid flow from the pump to the slave and 6 other control valves. I have a couple of used ones I want to test, but no time to build a test rig. I just can't see what fails on these.

The o-rings may leak by I suppose, but since those are external on the cartridge, they would only allow fluid to bypass the valve spool of old/worn.

The solenoid itself seems simple enough, the ones I have measure identical for both winding and insulation resistance.

That leaves the spool itself and since there's no software inside, the fluid boundary must be a tight tolerance fit. Perhaps the spool wears bit inside the cartridge? Maybe wear particles in the system build up over time since there's no fluid change/flush regimen? If you think about it, every single ounce of fluid that goes anywhere in the SMG travels through this valve, so perhaps particles are the cause.

Good to know it helps, do you think it would have been possible without removing the hydraulic unit if you just pivoted the transmission down?
 

·
Registered
Joined
·
464 Posts
Discussion Starter #3
On the M6 it's only possible to reach it by lowering the pump assembly.

I am wondering about the failure mode too. Especially since my problem was less noticeable after the system warmed up. I don't fully understand how this valve operates in the sense that it seems to be a two position valve. If that is true, it must have to rapidly cycle to vary the pressure to the clutch slave. That rapid cycling is how I suspect Vanos solenoids control positioning as well. The other possibility is that these aren't solenoids, but are stepper motors to continuously vary the balance of fluid flow with tapered profiles on the spool.

Both of these possible modes of operation would explain my car's symptoms. Maybe I'll dissect my old valve and see what's what.

I'm sure there's a group of engineers in some cubicles over in Germany who could tell us how this system works, but they probably signed nondisclosure agreements!
 

·
Registered
Joined
·
5,673 Posts
I've studied these far...far too much.

The clutch valve is proportional, so it can advance, hold, and relax in any position.

If you take yours apart, you'll find three main components (aside from software):
- solenoid body
- cartridge body
- spool piece

The solenoid body in black has a lip at it's business end. The cartridge body is staked in 3 (or 4) places to engage that lip on the solenoid body. When you separate them, that's it! There's no connection inside, just a slug from the solenoid core which pushes it's pin up inside the cartridge body and pushes against the spool piece.

The spool piece slides freely inside the cartridge housing and has a spring at it's far end which returns the spool to it's resting (shut, I believe) position.

The spring seems to be progressive rate, so by varying the current to the solenoid coil, you generate just enough linear force to move the spool slightly. More current = further travel.

I believe, looking at the schematics, that the valve is also hydraulic assist, so testing it at varying positions is likely impossible without system pressure providing the proper feedback.

So, the signal from the SMG module to the solenoid valves is pulse width modulated. By increasing the duty cycle of the signal, greater (average) current is applied to the solenoid coil and more force generated.

If you look at the adaptation process, there is a "curve" generated when teaching in the clutch valve. I forget how many data points it stores, but I always wondered why the points in the middle were higher than the ends. Finally one day it dawned on me (unverified) that the current values must indicate progressive positioning speed values, then regressive let-off values.

The valve has to be able to let pressure off the slave cylinder ever so slow and slightly to allow for smooth clutch engagement. At the same time, when you are in the higher shift modes, it has to move much faster to allow for fast clutch motion.

If you look at the schematic (my handy animated gif below, really bored at work one night), it also allows fluid to the shift proportional valves (total of 2, one for incoming gear, one outgoing) and the four (non-proportional) shift travel valves. This provides a hydraulic interlock of sorts to prevent pressure from being applied to the shift valves unless the clutch slave is pressurized.

Animated SMG clutch valve.gif
 

·
Registered
Joined
·
464 Posts
Discussion Starter #5 (Edited)
I suppose that's what makes them so expensive. Variable position with consistent performance in such a small package. When Pierre was researching the part number there were two of the gear position valves in the US and something like 14 of the clutch position valves. For a part that BMW has no service procedure on, that's a lot of valves (relatively speaking) and may indicate a significant failure rate.
 
  • Like
Reactions: MOTRSPRT

·
Registered
Joined
·
1,211 Posts
I love threads like this. This is how we all learn, great job. I'm surprised the hydraulic block had to be lowered. Must be a difference with the M6. When I did my pump motor I had room to access all those parts by just lowering the transmission. I too have suspected this part as being one of the culprits to jerky shifts. When I changed the smg pump motor out, I added a few ounces of Archoil 9100 oil along with the CHF-11 fluid as it helps lubricate components. It helps big time on diesel injectors with stiction. I'm running in my motor oil too to see what affects it has on the vanos solenoids periodically running vanos solenoid tests to see where the numbers are. So far my shifts have never been better. Here is a picture of my access with just the transmission lowered on an M5.
 

Attachments

·
Registered
Joined
·
464 Posts
Discussion Starter #7
Looks easy to get at with the pump motor removed! Thanks for the tip on Archoil. I'll look into it.
 

·
Registered
Joined
·
1,211 Posts
It's suppose to help clean varnish and carbon build up so I'm hoping it will help the rings. Powerstroke injectors get immediate help from this as they are oil operated. Some here say they have high oil consumption and might be blow by. This may help it but my oil level stays the same during my 5K oil change interval, so I won't be able to tell. Plus all my compression numbers were good to begin with. It's fun trying stuff as I don't think it will hurt anything.
 
1 - 8 of 8 Posts
Top