Viscowerkstatt Kern

History

As early as 1917, the American inventor Melvin L. Severy got a patent granted for a viscous fluid coupling. However at this time the only oils available were viscous oils of mineral origin, which are badly suited for transferring high torques. That’s because on the one hand, viscosity decreases when temperature rises, and on the other hand, higher temperatures cause disintegration.

Only the wonders of modern chemistry made the breakthrough of the idea finally possible. Now silicone oil could be produced synthetically. Therefore a fluid was available which is resistant to highest temperatures and also loses very little of its viscosity when warmed. Even so, using silicone oil is at best a compromise solution, since its pseudoplasticity is actually unwanted.

Thus higher difference revolution speeds (shear rates) cause an apparent decrease in viscosity, followed by a digressive increase in the transferrable drive torque. The opposite behavior, called “dilatant”, would be ideal for the use in a viscous-coupling. Sadly there are no known dilatant fluids which react similarly chemically stable to silicone oil when used under the special strains of a viscous-coupling.

The basic idea was taken up again by “Harry Ferguson Developments” in the 70s and the modern viscous-coupling, using silicone oil as a transfer medium, was born. In the automotive industry, the viscous-coupling was first used primarily as a torque converter, a vibration damper and as a viscous cooling fan. In the beginning of the 80s, new applications in 4WD vehicles were found after lots of development work:

The Viscous limited slip diff (VLSD):
Speed-sensitive differential lock

The Viscous-Transmission:
A middle differential and a differential lock between front and rear axle

By the way the first vehicle with a viscous-transmission was not the VW-T3 Syncro but the American AMC Eagle, which is said to be the precursor of nowadays’ SUVs. It was produced between the years 1979 and 1987.

Functionality

The viscous-coupling is very similar in its build to a multi-disc clutch (known from motorclycles). The only difference is that usually the drive torque is transferred through the shearing force of a fluid and not through mechanical friction.

The free inner volume of a viscous-coupling is filled to about 90% with silicone oil.

Actually, classifying silicone oil as an "oil" is rather misleading because the word makes you think of something greasy or lubricating. Due to weak intermolecular forces, the carrying capacity of a silicone-wall-film is rather low though, which is why you could actually call it an anti-lubricant, especially when looking at the material combination steel-steel.

The input variable for the transferrable drive torque in the viscous-coupling is exclusively the varying differential speed between the two axles. Concerning the transmission behavior we have to distinguish between two different modes. In principle, a Viscous-coupling first goes through the "Viscose-mode" and can switch to the so called "Hump-mode" in case of continuous stress.

Viscose-mode:

FLUID FRICTION

When the basket and the hollow shaft of the viscous-coupling turn at the same speed, the silicone oil is not exposed to viscous resistance. This is because the external plates in the basket are connected to the rear axle and the inner plates with the front axle. In practice this mode is never reached, because there are constantly small differences in rotation speed while driving (tire slipping, turning, slight differences in tire size and so on).

As soon as the external and inner plates rotate at different speeds, cohesion causes inner friction inside the molecules of the silicone oil, which tries to equalize the differential speed again. In this process the silicone oil is exposed to additional shearing forces at the holes and slots between the plates which are rotating against each other.

Silicone oils are pseudoplastic fluids, which means their viscosity decreases with increasing shear stress. This causes a digressive transmission behavior in the drive torque. The transferred torque depends mainly on the momentary viscosity of the silicon oil and the geometry of the set of plates. Then again the momentary viscosity depends on the base viscosity, the temperature and the shear stress.

Hump-mode:

SOLID STATE FRICTION

The inner friction that occurs in Viscose-mode causes the silicone oil to warm up. Since silicone oils have high thermal expansion (about fourty times that of aluminum), the inner pressure rises inside the hermetically sealed viscous-coupling. During this process it’s mostly the degree of filling (e.g. 90%) of the viscous-coupling which influences the speed at which the pressure increases. In case of a steady differential speed, the contained air therefore gets more and more compressed and creates a solution with the silicone oil, until an effective filling level of 100% is reached. In this state the inner pressure rises abruptly, so that further energy input would destroy the viscous-coupling, which is designed for a maximum inner pressure of approx. 100 bar. This causes the viscous-coupling to go into Hump-mode. In the past the Hump-effect was thought to be a dilatant behavior of the silicone oil. However the Hump-effect has nothing to do with sudden changes in viscosity since silicone oil is not dilatant but instead the opposite, namely pseudoplastic.

In reality what is happening is a bit more complex: Due to a destabilizing flow, an inhomogeneous pressure distribution is created inside the viscous-coupling. This results in varying gaps between the plates. What is decisive here is the production type of the plates: A rounded leading edge on the top and a sharp burr on the bottom are created by punching. The rounded edge works like a hydrodynamic lubricating wedge on which the plate floats while the sharp burr scrapes the silicone off the neighboring plate’s surface.

If the gaps become too narrow, the silicone film tears, causing mechanical friction between the plates. Hence the transmitted torque suddenly increases, whereby the differential speed quickly declines. On the road this means that a vehicle that got stuck can now either be freed or that the motor dies. The temperature falls and the viscous-coupling changes back to Viscose-mode. The Hump-mode is meant for a momentary increase of traction in extreme situations, but it is also a constructive self protection mechanism of the coupling from overheating.

Problems

The same phenomenon occurs in 90% of old T3-Viscos in the test run before overhauling:
That of extreme hardening:

Gebrochene Antriebswelle im VA-Getriebe

Those extremely stiff viscous-couplings are practically always in Hump-mode while driving. That’s because the front differential usually reaches temperatures of around 60°C, which is why front and rear axle should be regarded as rigidly connected. The resulting tensions in the power train are responsible for excessive rubbing of the front wheels in tight curves on dry pavement. Not only does this lead to increased tire wear, but in the worst case to damages to the transaxle.

However this very common “extreme hardening” of the T3-Visco is no sign of aging but a construction problem. This specific extreme hardening of the T3-Visco only occurs when the viscous-coupling has sucked gear oil from the front differential. The problem of “oilsucking” was noticed at Steyr-Daimler-Puch (SDP) in the late 80s, which is why it was also investigated in a Dissertation.

Conclusion, Diploma Thesis, Thaller

The remarkable conclusion was the following: At wintery outdoor temperatures, a static vacuum forms in the viscous-coupling due to the high thermal expansion of silicone oil. This vacuum is enhanced during start-up. Because of the vacuum the viscous-coupling sucks in portions of gear oil in the cold-running phase and breaks over time. This problem does not depend on the mileage. In short-distance operation an extreme hardening of the viscous-coupling can already happen after 2000 km (approx. 1243 miles). However the results of this investigation came too late in the summer of 1990. Therefore SDP did not make any further effort in solving the problem.

A simple test setup with a manometer shows the virulence of the issue.

Test setup

In the first step, an excess pressure of 1 bar (14.5 PSI) was applied to a viscous-coupling filled at room temperature of 16°C (61°F). In order to test the absolute leakproofness of the setup, the viscous-coupling was immersed in a water bath. There were no signs of leaks.

In the second step, driving in winter was simulated by placing the viscous-coupling in a fridge over night, at arctic temperatures of -15°C (5°F).

As a result of this cooling the manometer showed a blatant fall in pressure. The remaining pressure was 0,15 bar (2.18 PSI). Therefore a temperature difference of 31°C (56°F) already causes a static pressure drop of 0,85 bar (12.3 PSI). After reheating, the pressure goes back to 1 bar (14.5 PSI).

The usual filling method at ambient pressure and room temperature thus means a latent danger of an irreparable extreme hardening of the viscous-coupling during the winter.

Our approach

In order to prevent the “oilsucking” caused by the vacuum, the viscous-coupling is filled through a special valve.

This makes it possible to insert a slight static excess pressure into the viscous-coupling. Moreover we can adjust the static pressure to the ambient temperature via the special valve.

Comparing it with the SDP factory adjustment back then, nowadays we need another filling quantity and an adjusted viscosity. Yet with help of a test bench, working out this modification is a solvable task.

Service life

Steyr-Daimler-Puch had the durability of the viscous-coupling tested in a Dissertation under specified operating conditions. The investigations showed that the service life of a viscous-coupling depends on the differential speed, the power loss and the height of the Hump-torque. On principle we have to distinguish between two separate service life- categories though:

VISCOUS CONTINOUS RUN

The service life of a viscous-coupling in Viscose-mode (fluid friction) depends mostly on the wear resistance of the silicone oil, since the aging of the silicone oil leads to a slow hardening of the viscous-coupling. Therefore the increase in toughness of the oil causes a rise in transmitted torque over time.

HUMP CYCLE STABILITY

During Hump-mode the plates experience wear-intensive mechanical and thermal stress. From a certain abrasion onwards, the plates lose their ability to initiate Hump. The Hump-moment which is important for self- protection cannot be achieved anymore. This can lead to an overheating of the coupling, causing the inner pressure to rise unstoppably.

FURTHER INFORMATION CONCERNING SERVICE LIFE

VISCOUS CONTINOUS RUN - INCREASE IN VISCOSITY

In principle the increase in viscosity of the silicone oil can be expressed with the following function:

Silikonölverhärtung

80-120% Zähigkeitszunahme

During its service life, the viscous-coupling hardens after a third-order function, meaning that in the first third the viscosity increases only slightly. Then you can see a sudden, nearly linear, jump. In the last third of the function it levels off again at about 180-220% of the original base-viscosity.

VISCOUS CONTINOUS RUN – SERVICE LIFE FUNCTION

In the Diploma thesis it was assumed that the viscous-coupling would be replaced when it had hardened by 10%. The curve for service life is therefore exclusively dependant on the differential speed level and is defined by the following function:

In real life, differential speed can occur in the viscous-coupling at up to about 300rpm (full-load acceleration on gravel), however only for a few fractions of a second. The actual differential speed and its commonness while driving were determined with the help of test drives. It was found that it falls within a range of 0-80rpm. However it’s important to note the logarithmic division of the percentage scale of the speed duty cycle-measurements: During the 2168 hour long test run which covered 150.000 test kilometers (about 93 206 miles), a differential speed higher than 50rpm was only observed in 0,1% of the time (equaling 130 minutes). Translated to the function for silicone oil hardening written above (pic. 31), this data shows us that during a 20 hour test run with the viscous-coupling at a differential speed of less than 50rpm, the expected increase in viscosity would be 20%.

Even though this formula can only be applied roughly to real life driving (e.g. in kilometers or service hours), there is a clear tendency showing: Permanently increased differential speed leads to a faster rise in viscosity of the silicone oil, due to higher shear stress. Using a decoupler as a switch-off 4WD at speeds of over 90km/h (about 56 miles/hour) on highways in the summer therefore would be the most sensible thing to do.

Choosing the base viscosity

In order to maximize the service life in Viscose-mode, choosing a base viscosity which is as low as possible is crucial. On the one hand, highly viscous oils have a strong tendency to harden, and on the other hand we can cover a broader bandwidth of still acceptable viscosity. Also the lower limit in the test bench diagrams should be drawn nearer to the tolerance curve, because the silicone oil becomes more and more viscous over time of use anyway.

Hump - Service life

The ability to go into Hump-mode depends on the wear condition of the inner and outer plates. If the viscous-coupling can’t change neatly into Hump anymore, this can lead to an overheating of the coupling, while the inner pressure rises uncontrollably. This high pressure usually leads to a heavy damage to the seal rings of the viscous-coupling. In the differently built viscous-coupling of the T4-Synchro, a blasting of the aluminum casing can also occur.

Back then in the tests done by Steyr-Daimler-Puch, a viscous-coupling was driven to Hump-mode so often that the pattern of a Hump-abort showed. In this, the occurring Hump-moment was limited to intervals of 200 to 800 Nm. The function for the Hump-cycle-stability is therefore the following:

The corresponding service life curve for a specific number of Hump-cycles shows an inverted exponential dependency of the Hump-moment on the service life:

In the experimental setup of back then, the maximum number of possible complete Hump-cycles before Hump-abort at a Hump-torque of 600 Nm was 243. At half a Hump-torque of 300 Nm it was already 3688 Hump-cycles and at a torque of 200 Nm the trials were prematurely ended, because even after 3300 cycles the coupling showed no signs of a Hump-abort.

For maximizing the Hump- life span of the T3-Visco this means the following: The viscous-coupling should get into Hump-mode as fast as possible in order to keep the necessary Hump-moment as small as possible. However should initiating the Hump-mode take unnecessarily long, the rear wheels has probably already dug itself deep into the ground so that the necessary Hump-torque at the front axle has to be correspondingly large.

Viscowerkstatt Kern
DE | EN