Announcement

Collapse
No announcement yet.

Viscous Coupler Testing (Rebuilt unit vs. 'Good')

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    #16
    Thank you all for the kind words! I have ordered a larger vacuum chamber for degassing the fluid and will start rebuilding a few more units soon. I'm hoping to have some more data to share in the upcoming weeks.

    Comment


      #17
      Awesome dude, this is cool, thanks for sharing this info
      https://www.facebook.com/Cry_77_IX-110483784967661
      https://www.instagram.com/cry_77_ix/

      Comment


        #18
        So cool man, nice work and keep it up! I have several extra TC’s with unknown VC’s that can be used for testing if need be.
        1989 325ix Alpinweiss sedan. Auto to manual conversion. Mostly stock.

        Comment


          #19
          AWD, I keep checking back to to how the VC project turned out. Any updates? Thanks!

          Comment


            #20

            Hey sorry guys that I've been quiet. I've been bouncing between projects and trying to track down additional good couplers to test.

            The latest testing I've done produced different results than my initial tests and I was hoping to sit down in the coming weeks to write a new coupler post with my findings. Essentially, 80,000 cSt fluid, when filled according to some BMW documentation (90% silicone/10% air from translated first photo below), produces very high torque (~220 ft-lbs @ 110 RPM). The brake test stand that is mentioned in the factory service manuals only yields about a minimum of 45 ft-lbs at the same speed when converted from roller torque to coupler torque (see second picture below). Thankfully, a test procedure is defined in this service manual that can be used to warm the coupler enough to evaluate but not overheat. I will be using these for some guidance but there is no definition of what a nominal torque should be. I am looking to test more good units soon but my 'good' coupler produces about 80 ft-lbs at that same speed. So, why the higher number this time do you ask? Well, my first test was a guess at how much I should fill the coupler and I did not have any accurate fill method. This time I measured the weight of a low viscosity fluid to fill the coupler, converted that to a volume, then converted that back to weight of silicone. I suspect my first test had a lot more air and lower torque as a result.

            When I get time in the next few weeks, I plan to test with more and more air in the coupler in order to reproduce the torque from my first test. Then I am going to switch to 30,000 cSt fluid to understand where that lands. This will give me some understanding of the viscosity vs. torque and fill volume vs. torque relationships so I can aim for the 80 ft-lbs. Hopefully I will also get more good couplers soon for testing so that I can confirm around 80 ft-lbs at 110 RPM as being 'good'.



            Click image for larger version

Name:	Screenshot 2021-06-15 150506.png
Views:	307
Size:	177.9 KB
ID:	10021529


            Click image for larger version

Name:	Screenshot 2021-06-15 145257.png
Views:	316
Size:	129.3 KB
ID:	10021528

            Comment


              #21
              If we were local, I'd let you borrow my brand new transfercase. But unfortunately, there's no way I'm letting that out of my hands now, lol.

              I also had a new coupler, but sold it to buy the brand new transfercase instead.
              Build thread

              Bimmerlabs

              Comment


                #22
                lol thanks nando and I don't blame you. There's a guy on Facebook who has two brand new VCs and he's in the same boat. I've had several people who are interested in helping out but haven't come through yet. The most promising was a local guy with 4 couplers and at least one confirmed good. Otherwise, I'm waiting on others to pull them from transfer cases and cars.

                In the meantime I can continue to play with my experiments and learn more about the sensitivity to air volume/viscosity.

                Comment


                  #23
                  I did some testing to understand why I had lower torque in my initial testing and much higher torque in later testing with the same viscosity fluid. It appears that my initial testing that produced around 80 ft-lbs must have been 70% filled instead of 90% filled. I know I didn't have a good way to measure the volume or weight when I did my initial testing so this seems feasible. I will double check my internal volume again to make sure these numbers are accurate next time I clean out the coupler. Next step will be to drop down to 30,000 cSt fluid at 90% fill and see what I get. Stay tuned!

                  Click image for larger version

Name:	Volume_Fill.png
Views:	281
Size:	34.0 KB
ID:	10022500
                  Attached Files

                  Comment


                    #24
                    Tuned in!

                    Comment


                      #25
                      I was finally able to capture hump-mode on hump-day but ran out of energy to post here. This data was collected from one of the units that I made rebuildable. Hump-mode is essentially lock-up and is the protection mechanism for viscous coupler during extended periods of slip between the front and rear axle. The internal temperature and pressure spike causing the metal discs to contact. This drives the torque up and reducing the slip between axles. When the slip decreases the temperature of the coupler drops back down and the coupler unlocks. Tuning this hump-mode is critical to ensure that the seals and silicone do not overheat causing premature failure.

                      The sensitivity of silicone fill volume is pretty crazy. My previous attempt at hump-mode (no temperature plot yet) was at 87% fill volume and the temperature was around 300°F. If I can get that test to repeat, that would mean 2% more silicone drops this hump-mode 75°F. I have not decided what temperature to target for hump-mode but perhaps somewhere around 240°F (or less) might be ideal. I suspect that around 89% or 90% fill, the hump-mode would engage around this temperature which lines up with the BMW literature of around 90:10 fill ratio of silicone-to-air. This is the upper limit of what ATF should be used at according to the internet. Quick research has multiple people reporting 200°F on transfer cases for various vehicles under normal operating conditions.

                      Also, fingers crossed I might have multiple more couplers to test later this weekend. If all goes well, one or more will be good and provide more information on OE performance.

                      Click image for larger version

Name:	Hump-Mode_30k_85percentFill.png
Views:	258
Size:	46.5 KB
ID:	10023967

                      Comment


                        #26
                        FYI, you can also fill the transfercase with motor oil, if you're worried about temperature. I think the primary failure on these seems to be the seal in the VC itself, which I assume over heats like you're talking about.

                        When the plates touch doesn't that cause damage to them? I thought they weren't supposed to touch.
                        Build thread

                        Bimmerlabs

                        Comment


                          #27
                          Agreed Nando on the seals. I was thinking about high performance ATF but I hadn't considered engine oil. I'm trying to think of what the benefit would be of running high temp hump mode? Fewer instances of it occurring maybe? So far every ruined coupler I've seen has had the seals blown out. Buna N is only good to 250°F where Viton is good >400°F. I started with Buna N but quickly realized that limited my operational range for temp.

                          I do have one coupler now from a guy with slight heat discoloration but no leaks... But it has quite a bit lower torque transfer than I expect and a very early hump mode (~100°F). Not sure if I'm ready to call it bad yet.
                          Well it's sort of tricky on the metal plates. Under normal usage they don't contact but they do during hump mode. This is by design and is documented in every technical paper I've read so far. I actually had a discussion with coworkers today on why friction plates aren't used. My suspicion is that hump mode is not expected to be a regular phenomenon and so disc durability isn't a concern. I'd like to hear other people's thoughts on this. Additionally, the plates that rub against each other appears to be somewhat random based on the three I've torn down. In other words only a few plates actually rub during hump mode. So installing all friction plates would be overkill since only a few would get used.

                          This also points to the need to have hump mode be above normal operating temperature of the transfer case and oil but below a temp that damages seals and fluid. A low temp hump mode would likely prematurely ruin discs and cause additional binding when turning. I'm open to other ideas but this is where all my testing and research has led me.

                          I appreciate everyone's interest and feedback. I hopefully will have more test results soon. Cheers!

                          Comment


                            #28
                            I mean, 100f is probably bad. you could get that in ambient operation without any slipping at all.

                            Maybe it would be worth putting a temp probe in a stock ix transfercase and just see what the normal operating temp is. It should definitely be above that, but below the limit of the seals. I don't think you'll damage the fluid or the mechanicals, it's mostly going to be that VC seal that gets roasted.
                            Build thread

                            Bimmerlabs

                            Comment


                              #29
                              Good idea - a temp sensor and gauge would actually be a good permanent arrangement.

                              True story as follows:

                              About eight years ago I was immensely proud of my lady as she hammered through a rain storm (I repeat rain, not snow) at 80-85 MPH. Her automatic 325iX felt perfectly secure at speed hitting puddles running on Nokian WR GR2 treads. I could hear/feel the constant torque transfer from the passenger's seat and was enjoying the lady's latent mechanical aggression. Well everything was going all rad and such until I smelled the pungent smell of car doom. At that moment, I had no idea what was going on and had her pull over. So then, as far as looking over a car in the rain goes, I found nothing that matched the nasty smell that had saturated the interior of the car. We pressed on only to find that the smell dissipated - we continued pressing on for another 550 miles and hit snow. My turn at the wheel was a little less rad! As I started to turn into my intended direction at speed there was no VC to guide me. No worries - no damage - just a fried VC. All of that hydroplaning had fried her VC within a half hour earlier in the trip. Yes, a temp gauge might have given a warning about a weak VC.

                              Comment


                                #30
                                Sorry for the radio silence everyone, I've been pretty busy. I found the rabbit hole is pretty deep and after rebuilding 5 or 6 of these I have learned quite a bit. I have a few previous comments I'd like to correct based on what I've learned over the last two years. For example, the improvements in torque from vacuum testing were actually due to adding more fluid (devil is in the detail!). Should I start a new thread that shows all the improved knowledge or edit previous posts...or tag onto this one? I sort of want a new clean thread with all the knowledge in one place but there's also good content here already.

                                -Devon

                                Comment

                                Working...
                                X