The above diagram shows two sets of pressure volume relations.
The descending and ascending arms have been pulled apart slightly for clarity though, of course, they both lie along the same xy = a constant curve
The top set is the familiar Carnot isothermal expansion and compression.
The bottom set shows the pendulum arm expansion (aka extension) and compression.
The Carnot is shown without the adiabatic legs which means that no
area is enclosed by the cycle and no area is generated. Without any losses the best one could hope for is an eternal oscillation.
Is that perpetual motion of the first kind?
Likewise with the pendulum there is no equivalent to the adiabatic legs and so in the limi all we have is the pendulum rising to its zenith at 12 o'clock and falling to its nadir at 6.
So the problem is to insert the mechanical equivalent of adiabatic legs into the, erm...., let's call it the Bessler cycle.
We can get one helpful clue from the Carnot cycle.
We know that the essential difference between isothermal and adiabatic change in the Carnot is one of speed.
If we change the pressure very slowly in the Carnot then the PV relation will be isothermal. However, if we change it very fast it will be adiabatic because heat won't have time to enter the cylinder and all energy exchange takes place within the cylinder.
Equivalently, of course, is the use of insulation to slow down the transmission of heat. This is the conventional way the Carnot cycle is viewed.
A classic example where speed of change is important is the speed of sound. This even tripped up the sainted Newton. He assumed compressions and rarefactions followed the isothermal equation of
PV = a constant
instead of the correct adiabatic equation of
PV^(5/3) = a constant
and wound up with the wrong answer.
So how does knowing we need something fast to add to something slow help us?
Well, the Keenie shows one way forward. In the Keenie the weight transferred to the low inertia wheel drops faster under gravity than the unbalanced weight on the high inertia wheel.
No doubt members can think of other examples.