4 Buffer dynamics

On approaching the bottom landing a descending lift should trip the

'terminal stopping switch' there and brake to a halt If the switch fails to
operate and the car continues to descend, it should then trip the W
limit switch' and be stopped. If all should fail and the lift were to go on
unchecked it would strike the bottom of the lift well and only come to
rest after damage to the lift and injury to the occupants. To avoid this
there are buffers fitted in the foot of the lift well, facing upwards so as to
cushion the impact of the lift if it overshoots. The buffers are not designed
to save a lift in h e fall. Other devias, such as the overspeed governor
co~~nected to the safety grips, do this. There are several types of buffer
systems. Our lift would use an hydraulic buffer. However, tirst I want
to examine a simpler arrangement. It comprises one or more suitably
proportioned helical compression springs set to face upwards and receive
the descending lift.
One feature of a spring is its stiffness. Another is its stroke, the amount
by which it can be compressed before the coils touch and the spring begins

7F
to behave as a tube (Figure 18). (A long spring may well buckle before
hit
the coils touch.) Of course, the stress should not exceed the appropriate cDmprerrrion enensiDn
limit for the material. The steel used for springs is chosen with great care
and is heat-treated to obtain optimum properties. The 'coil-to-coil' limit
for the stroke of springs in compression does not apply to springs in
tension, w h m only the stress limit i8 relevant. Figure 18 Spring characteristic
The idea of a buffer is to take energy from the descending lift, slowing it
down When a lift of mass m arrives at the b d e r head with velocity v it
has kinetic energy +mu2. This is to be absorbed by spring movement
(stroke) e. The energy stored in a linear compression spring, or the work (a)
done in compressing it, can be written as fke2, or fFe where F is the
magnitude of the force at the extension e (Unit 3). Using the work-energy
method, with the lift car as the system (Figure 19):
K, = f rnv"amva1 at buffer) V
l F

K, = 0 (zero speed, buffer compressed)

Figure 19
W, = mge
W,= -fke2
W,=O
Actually the tension Twill do some negative work, and so in taking it as
zero, I am erring on the side of safety.
K2=KI+W
0 = )mu2 + mge - f ke2
In this equation only k and e are unknown, so once one is chosen we can
calculate the other, and also find the maximum spring force F = ke.
An example of a lift that would be permitted to use spring buffers is a
goods lift of capacity 300 kg and rated speed l m S-'. Simple spring
buffers are not recommended on faster Lifts than this, where something
more sophisticated is needed. The mass of the W, sling, etc., for this lift
would be perhaps 700 kg, so the total loaded mass would be 1000 k p
Notice how in a small lift like this the mass of the empty lift is higher in
proportion to the useful load than was the case for the l200 kg rated lift.
In the Standards documents on l i there are recommendations dealing
with buffer design. These are that they have a stroke of at least e = 0.135v2
(where e is in metres and v is the rated speed of the lift in m S-'), and that
they be designed to cover this stroke under a static load of between 2.5
and 4.0 times the weight of the fully-loaded car (F = 25mg to 4mg). Note
that this stroke, and the corresponding loads, define the stiffness of the
19
 springs supplied as buffers. However, the definition is not directly in terms
of what happens at the foot of the lift well. Rather, it is couched in terms
of what can be demonstrated in a laboratory test heforehand. The buffer
intended for our goods lilt will require an extension e = 0.135 m and a
stiffness less than k = 4mg/e= 291 kN m-'and greater than k = 2.5mg/e =
182 kN m-'.
We have just seen from the work-energy procedure that k and e cannot
be chosen independently:
f kef = +mu2+ mge
10 mu2 2mg
k=-+-
e2 e
Actually we should allow for the lilt arriving a little faster than its rated
speed, so I shall take an arrival speed of 1.15u= 1.15m S-'. Say I now
choose a maximum extension e = 0.135 m, just complying with Standards.
I shall require
k = 72.6 + 145.3 = 218 kN m-'
which falls within the range permitted by Standards, so all is well. The
maximum force will be F = ke = 29.4 kN, a piece of information that will
need to be considered in the design of the buffer beam for this li.
After full compression the spring expands, restoring its strain energy to
the lift car. Neglecting losses, the car would leave the spring at its amval
speed. It rises into the air beyond the uncomprcased wn6guration of the
spring, only to1 fall again. The lilt will wntinue to bounce; the system
vibrates. Actually all the actions just described incur losses of energy in
the metal, the supports, and the Lift. Air resistance in the enclosed well
plays a part in damping the motion and eventually the lilt wmes to rest.
To get a further impression of the effect of the buffer action, we can
estimate the acceleration undergone by the lift and its contents. Figun 19
shows the appropriate FBD. We have already calculated that the max-
imum spring force is F = 29.4 kN, the weight is 9.81 kN, so neglecting T
the resultant R is 19.6 kN upwards.
a = R/m= 19.6 X 103/103= 19.6 m S-' upwards
The acalerations in the buffered lilt are rather less than those encountered
during flight in a wmmercial airliner.
For lihs having a design speed not e x d i g 0.4 m S-' it was once the
practice for the buffers to be of timber or rubber, the rubber being cm-
ployed as simple columns. One of the features of rubber is its particular
type of 'elastic' behaviour. In Engineering Mechanics the word 'elastic'
means a return to the original form when a load has been applied and
removed. Rubber does this under some circumstances, but the
load-extension deflection curve is non-linear (Figure 20). The extension
or contraction is not proportional to the load. Also the return path is
not the same as the outward path. For a buffering application these
load-deformation charactcristia have an advantage and a disadvantage.
On the credit side the loop formed by the loading and unloading line
means that the rubber buffer gives back less energy than it receives, the
diference bkng retained mainly as thermal energy in the rubber. Thus
wmpmsion the lift cage will not rebound so much from a rubber bu6e.1 as it would
Figure 20 W - e x t e n s i o n curtmfor from a metal wil spring. On the debit side the increasing slope of the
rubber characteristic means that the deceleration rate builds up to a serious peak
value at the end of the stroke, jarring the lift and the occupants.
If the rated speed of the lift exceeds 1.0 m S-', oil buffers are used. This
is the sort of buffer used in our passenger lift. Basically these wmprise a
vertical cylinder in which moves a piston. The desctndig lift contacts the
piston rod, forcing the piston down in the cylinder (Figure 2la). This
generates a pressure in the oil below the piston and forces the oil through
holes in the piston. The buffer is elaborated by the addition of a rubber
pad and a spring which cushion the initial impact between the lift and the
piston. Finally, when the buffer has completed its vital task and the Sit is
being raised off the buffer a light spring returns the piston to the raised
position. Using a hole or annulus in the piston means that a higher speed
of oil transfer corresponds to a larger pressure and hence force on the
piston. So, when the piston begins its downward travel forces are high;
towards the end they are low. The size of the annulus must be a comprom-
ise between reduction of the initial forces, and reduction of stroke. Figure 21 Hydraulic buffers

A better idea is to form the annulus in the cylinder wall, connected by a

series of holes of different siz+s: large at the top, small at the bottom
(Figure 2lb). There is now the prospect of levelling off the force peak.
Figure 22 compares the force-position curves for different buffering units form
to give the same amount of energy absorption (area under the graph). hydraul~c
-fixed hole size
Notice how the simple oil buffer behaves in the very early part of its stroke
like a stiacoil spring. This is because the fluid hss not yet started to move, spring
and the only iduence on the contact force is the small rubber buffer at hydraulic
the piston rod tip, and the spring below it. If this force (and therefore rarlable
hole size
acceleration) 'spike' can be shown under test to occupy less than 0.04
to the requirements
seconds it is excludedoffrom
Standards.
the calculation of buffer suitability, according L 5 9 stroke

Figure 22
To determine the buffer performance, we must first analyse this impact
between the lift cage (taken as fully loaded here) and the piston rod and
attachments totalling about 200 kg (Figure 23). I estimate a coefficient of before ZMkg after
restitution of zero, so immediately after impact the speed of both parts is
v. The arrival speed u is taken as 1.5 m S-'.
Along y: G1 = G1

The speed change of the S i car in this impact is not very great, as you
might expect, because it has a much greater mass than the piston unit,
After this impact the standard hydraulic buffer aims at an average decel-
eration of g = 9.81 m S-'. Newton's Second Law can be applied to the car
FBD shown in Figure 24.
Along y: mg - f = ma (T neglected)
3MOb

c
0 m0

Note that m is now the total 2880 kg + 200 kg= 3080 kg, so the force that Figure 24
we want the buffer to exert is 60.4 kN. You could now select a suitable
buffer from manufacturers' performance data. Alternatively, if you wanted
to manufacture the b d e r yourself you now have the necessary perform-
ance specification, and could go on to design it appropriately.