Torque converters have been an important part of torque transmission systems since the first truly automatic transmissions were created in the late 1930’s. Their past, present, and future paint the picture of a wonderfully useful technology whose role in the transmission system has been gradually either completely replaced by more fuel-efficient solutions like the dual-clutch system (like those notoriously in the Ford Focus), or has been reduced by transmission features like the lock up clutch. I think back to my first car, a 1996 Chevy Blazer, and how silky smooth the power got to the wheels – the only way to tell if that car had shifted gears was to watch for sharp changes in engine speed. You could drive across the country and never create even so much as a wave in your coffee mug. Then, when I got a new car – 2010 Ford Fusion – I noticed how much I could feel some of the gear shifts. I would later learn why this ride in my new car was noticeably less smooth than the Blazer. A major part of the story is about the reduce role of the torque converter in the Fusion. What was lost in the smoothness of the ride was gained in fuel efficiency, and this has been the main theme of the story of torque converters in the last 20 years. Here we will dig into the behavior of flow in the torque converter; like every other fluid physics analysis, you don’t need to care much about cars or the trade-offs that go into deciding what the end behavior of a product is. The morals of the fluids story, as always, are conservation of mass, momentum, and energy.
Some Background
Torque converters are typically the component in an automatic transmission system which are nearest to the engine. Before discussing the details of the physics, let’s just get an overview of the behavior in the system and some general rules that apply.
Figure 1 shows an example of a engine curves (torque vs engine speed at three throttle levels), alongside curves which show an example of the torque converter’s ability to absorb torque from the engine across the range of engine speeds. This is the input side of the converter – what it amounts to is, the converter acts as a first filter on the engine torque and speed. Only conditions which reconcile with both the engine output and converter input get into the converter – that is to say, only points where the engine output curve and the converter input curve intersect are achievable. You can see in the figure above that there are some different engine curves corresponding to various levels of throttle applied. There are also two different torque converter curves shown. SR in this case is “speed ratio” which is the turbine speed divided by the pump speed, which most often ranges from 0 (no turbine speed) to 1.0 but can go higher, as described further below. This ability to de-couple the input speed from the output speed is an important feature of the torque converter (why you can sit at a light with the engine still running).
Figure 2 is what you would call a “good enough for government work” sketch showing a cross section of the torque converter – a view called the meridional view. It is roughly what you would see if you sliced a converter in half. This view shows kind of the footprint that the blades make – in terms of the inner and outer radii of each blade, and the amount of the cross section that the blade occupies. You can’t see anything here about the blade shape, just the general space claims. And keep in mind, this is a generic shape for demonstration. Over time, the axial space claim of the converter has been shrinking, and they take on more of an elongated oval profile in the view shown (example here, where they call them “squashed” converters).
Here’s an interesting bit, which also teaches a lesson: early versions of the torque converter were not torque converters, but a simpler device called a fluid coupling, and consisted of just two components. There was the pump, which was connected to the engine, and the turbine, which was connected to the gears and ultimately the wheels. When the engine turns, the pump turns, and when the wheels turn, the turbine turns. Here’s the lesson: the fluid coupling system is essentially a closed system, so that torque that the pump exerted on the fluid got translated into angular momentum in the flow, and then the turbine, operating at a lower speed than the pump, would reduce this angular momentum in the flow and absorb 100% of the torque that the pump put into the fluid. That’s a key property of a fluid coupling, a 1:1 torque transfer between pump and turbine.
Somewhere along the line, somebody understood what was going on enough to introduce a third component to the system and the torque converter was born. The third component, as it is in place today, doesn’t rotate during most of the converter operating range and is thus called a “stator.” With this third component, a new world of possibilities opens up – it’s the ability to transfer more torque to the turbine than what is on the pump.
If we take the “torque converter is a closed system” assumption at face value (very, very close to true – there is only a leakage level of flow entering and leaving the converter normally), then we again understand that any torque put into the fluid by the pump gets absorbed by these new components, and we can write the equation below.
Tpump + Tturbine + Tstator = 0
So, there it is. We now have a torque relation with more than two terms! You can re-arrange this equation to show that the torque on the turbine (Tturbine) can be different from the pump torque (Tpump) by an amount equal to the torque absorbed by the stator (Tstator). Let’s introduce another quantity, called the torque ratio (TR), which is TR = Tturbine / Tpump (I’m playing a bit fast and loose with the signs here, since strictly speaking the pump puts a torque on the fluid which would can say is a positive torque, and the turbine gets a torque put on it from the fluid, which would then be a negative). With a little bit of algebra, we can show that the equation below is true, and it’s just a re-statement of the equation above, written in a way that highlights some important behavior.
Tstator = (TR – 1) Tpump
So, the torque on the stator component, we note, can be non-trivial within the context of the system. Some three-component torque converters have torque ratios in excess of 2, so that the stator torque is higher than the pump torque itself!
The Nitty Gritty
Ok, so let’s look at the details of converter behavior. I have seen some weird descriptions of torque converter behavior over the years (they are “like a waterfall”? what?), which can be less than enlightening, or even misleading. As noted above, we are just playing with the same rules of mass, momentum, and energy conservation as we always are in classical physics, we just have to think about how to apply them in this context. It’s really pretty simple, but to apply the rules of physics and think about what is going on is informative here.
At the heart of torque converter operation is angular momentum. If you remember back to your physics class, angular momentum is the rotational analog of linear momentum, just that instead of force, mass, and linear acceleration we have torque (T), moment of inertia (I), and angular acceleration (/alpha), and these are related to angular momentum (L) through equations shown below.
I = r2 m T =dL/dT = d(I * omega)/dt
So the idea is, if we follow a small point mass of fluid, it can gain in moment of inertia by moving radially outward, or by gaining mass (eh, we can ignore this). And, any rate of change in moment of inertia or angular speed means that a torque is getting applied to something somewhere. We come to the crux of the behavior then: in the pump (at a constant angular speed, for example), a particle of fluid moves radially outward and thus experiences a change in angular momentum. As it happens, the corresponding torque is coming from the pump blades through a primarily pressure-based interaction, as described in more detail in this post. So we understand where the pump torque comes from. Correspondingly, in the turbine, the fluid particles are moving radially inward. So then, even at a constant turbine angular velocity, with fluid particles moving radially inward and thus experiencing a time rate of change to their angular momentum, and a corresponding torque is conveyed to the turbine blades.
Other Considerations and Observations
The currency for torque converter performance is really the circulating mass flow rate (i.e. the rate of mass flow passing across a plane between the stator and pump, for example). All else being equal, higher mass flow converter designs will be able to achieve a higher level of performance for a given metric. One of the trade-offs that you find when doing design work here is that design changes which increase the torque ratio will “cost” you with some reduction in flow rate.
What is colloquially referred to as “cavitation,” but what is really the effe “cavitation,” but which is more likely to be air which is dissolved into the converter fluid (it’s an oil, could even be something like 10W40) coming out of solution. There’s a description of this distinction here, but the effect in converters will be basically the same – the pressure field will be truncated and converter performance reduced. This phenomena also has the ability to substantially erode and even destroy converter blades over time.