Note: Descriptions are shown in the official language in which they were submitted.
2~16~113~
A TEMPERATURE CONTROL SYSTEM
FOR A CRYOGENIC REFRIGERATOR
Background
This invention relates to cryogenic
05 refrigerators, or cryocoolers utilizing linear drive
motors having pistons or displacers which
reciprocate in cylinders. Such refrigerators
include Gifford-McMahon or Stirling refrigerators
and expansion engines.
In various types of cryogenic refrigerators, a
working fluid such as helium is introduced into a
cylinder, and the fluid is expanded at one end of a
piston to cool the cylinder. For example, in
Gifford-McMahon type refrigerators, high pressure
working fluid may be valved into the warm end of the
cylinder. Then the fluid is passed through a
regenerator by movement of a displacer-type piston.
The fluid which has been cooled in the regenerator
is then expanded at the cold end of the displacer.
The displacer movement may be controlled by either
fluid pressure differentials or by a mechanical
drive.
A control system for a cryogenic refrigerator
is disclosed in U.S. Patent No. 4,5~3,793. One or
25 more parameters of a reciprocating-piston type
refrigerator are monitored to provide an electrical
feedback signal~ That signal is processed to
control the timing of the piston movement and or the
flow of refrigeration gas into the refrigerator.
--2--
Preferably, the feedback signal is an indication of
the position of the piston within its cylinder
and/or the temperature at the cold end of the
cylinder throughout a refrigeration cycle.
05 Continuous position indication may be provided
by a linear variable displacement transformer or by
a rotary encoder or by other means. Preferably, the
feedback signal is used to control valves which
introduce the refrigeration gas into the cylinder or
a piston drive motor. In a pneumatically driven
refrigerator, the feedback signal may be used to
control valves to and from both the refrigeration
cylinder and the drive cylinder. By controlling the
stroke, the temperature of the cold end of the
refrigerator can be controlled.
A split Stirling refrigerator is disclosed in
U.S. Patent No. 4,664,685 wherein a compressor
provides a nearly sinusoidal pressure variation to a
refrigerant gas in co ln;cation with a cold finger.
The compressor is comprised of a linear drive motor
having a drive coil which drives a reciprocating
armature. A detector circuit is coupled to the
drive coil for sensing an electrical parameter of
the coil which is a function of movement of the
armature. Motor drive circuitry which applies
current to the drive coil is responsive to the
sensed electrical parameter in controlling movement
o~ the piston element. The detector circuit can be
connected to sense back EMF in a displacer drive
motor within the cold finger.
.
:;
.: :
: : : .
'
--3--
Rotary driven compressors of cryogenic
refrigerators have been known t~ employ controlled
variations in speed to adjust the kemperature at the
cold end of the refrigerator.
05 Disclosure of the Invention
In a cryogenic refrigerator where the magnitude
of the displacement of the compressor armature is
adjusted to effect changes in the temperature of the
working fluid at the cold end of the refrigerator.
A reciprocating magnetic armature can be driven
by an electrically pulsed coil so that the armature
alternately compresses and expands the working
fluid. The working fluid is in communication with a
cold finger of the refrigerator. By controlling the
pulse width of the signal that drives the armature,
the armature displacement can be controlledO By
controlling armature displacement in this manner the
temperature at the cold end of a displacer within
the cold finger can be precisely adjusted for those
applications requiring variable temperatures, or
provide constant temperature in sonditions of
varying heat loads or ambient conditions.
To precisely control the displacement of the
armature, the position of the armature must be
accurately determined. A Hall-effect sensor can be
positioned along the wall or at the end of the
operating volume of the armature to sense the
location of a magnet attached to the armature. The
sensor generates an electrical signal whose voltage
~)q'~ $~
is correlated to a particular position of the
armature within the operating volume.
In cryogenic refrigerators using dynamic
absorbers or counter balance systems to reduce
05 vibration generated by the reciprocating armature,
where the operating frequency of the armature is
confined to a certain range, such a temperature
control system has many advantages. By maintaininy
a fixed frequency of operation for the armaturer a
dynamic absorber can be used to attenuate vibration
of the refrigerator housing.
The temperature control system of the present
invention provides for control of the amplitude of
motion of the armatur~ while it reciprocates at the
same frequency at which the dynamic absorber is
tuned to oscillate.
When the desired temperature of the system is
chosen, the temperature at the cold end of the
displacer is compared with the desired temperature
and an electrical signal is forwarded to the control
circuit of the linear drive motor to increase or
decrease the displacement of the reciprocating
armature.
As the temperature chanyes within the cold
finger assembly~ the temperature sensor adjacent the
cold finger adjusts the armature displacement to
slow the rate of temperature change. As the
temperature of the refrigerator approaches the
desired level, the armature displacement is
i~.
3S~
gradually altered so that the temperature does not
progress beyond the desired level.
The above, and other features of the invention
including various novel details of conr,truction and
05 combination of parts, will now be more particularly
described with reference to the accompanying
drawings and pointed out in the claims. It will be
understood that the particular temperature control
system for cryogenic refrigerators embod,ving the
invention is shown by way of illustration only and
not as a limitation of the invention. The principle
features of this invention may be employed in
various embodiments without departing from the scope
of the invention.
Brief ~escription of the Drawings
Figure 1 is a cross-sectional view of a linear
drive assembly of a cryogenic refrigerator of the
present invention;
Figure 2 is a cross-sectional view of a cold
finger of a cryogenic refrigerator in fluid contact
with the compressed fluid of the linear drive
assembly of Figure 1; and
Figure 3 is a schematic diagram of the
temperature control mechanism of the present
invention; and
Figure 4 is a cross-sectional view of another
preferred embodiment of a linear drive assembly of a
cryogenic refrigerator; and
.
~. .
--6--
Figures 5A-5F schematically illustrate
waveforms that occur at certain positions within the
circuit of Figure 3.
Detailed Description of the Invention
05 A linear drive assembly of a helium cryogenic
refrigerator utilizing a temperature control system
of the present invention is illustrated in Figure 1.
A linear motor is used to control the movement of an
armature 10 in the compressor 5. The linear motor
utilizes an involute laminated stator 20 first
disclosed in U.S. Patent No. 4,761,960, of G. Higham
et al. filed July 14, 1986 entitled "Cryogenic
Refrigeration System Having an Involuted T-~ml nated
Stator for its Linear Drive Motor."
As shown in Figure 1, this compressor 5 com-
prises a reciprocating armature lO which compresses
helium gas in a compression space 24. From the
compression space 24 the gas passes through a port
14 in the stationary piston 11 to pre-formed bores
through the piston 11 and plate 31 to form conduit
13. Conduit 13 runs along the core of stationary
piston 11, then curves at a right angle in insert 98
to a~gas fitting assembly 15. From the gas f.itting
assembly 15, gas is delivered to a cold finger of a
cryogenic refrigerator such as a split Stirling
refrigerator in which a displacer is housed as
disclosed in U.S. Patent 4,545,209. The stationary
~. .
: :~ . :
~no~s~
piston 11, mounted at one end onto plate 31, is the
sole support for armature 10.
The compressor is charged with helium gas
through the port 17. The gas is allowed to communi-
05 cate with an armature volume 12 through port 16which is in communication with a second pre-formed
conduit 18.
The armature 10 comprises an iron mass 38 Eixed
to a liner core 82. Iron is used because of its
high magnetic permeability and high magnetic induc-
tion; however, other materials having the same
characteristics may be used. A tungsten alloy ring
or other high density non-magnetic material 25 may
be incorporated at one end of the armature to give
more mass to adjust the resonant frequency of
operation and to help keep the armature's center of
gravity within the confines of the clearance seal of
the piston.
Preferably, the armature lO is fitted with a
ceramic cylinder 83 to provide a clearance seal with
the stationary piston. It is preferred that a
sleeve 82 made of non~magnetic stainless steel or
aluminum line the cylinder 83 to provide structural
support to the ceramic cylinder. A cermet liner 84
is mounted on the piston 11 to form part of the
clearance seal.
Surrounding the armature 10 just described is a
pressuxe housing 26. The size of the pressure
housing is constructed to allow helium gas in the
armature volume 12 to flow freely between the
.
;,' ..
- ~,: .
3S~
pressure housing 26 and the iron mass 3~ as the
armature 10 shuttles back and forth.
A stator 20 is located around the perimeter of
the pressure housing ~6. The stator 20 comprises
05 two coils 21 positioned between involuted
laminations 23 and separated by a magnet 22. This
stator assembly is further described in U.S. Patent
No. 4,761,960, by G. Higham et al. recited above,
which is incorporated herein. The splitting of the
involute stator contributes to the amount of stray
flux generated about the coils. Two shields 90 have
been concentrically disposed about the involute
lamination 23 to convey the magnetic flux lines along
the inside wall 51 of the housing 50.
As a consequence of the armature 10
reciprocating back and forth, mechanical vibrations
are produced by the comprsssor 5. To eliminate the
vibrations, a passive vibration absorber or dynamic
absorber 39 is attached to one ~n~ of the compressor
and is tuned to resonate at the same frequency as
the compressor's operating frequency. Preferably,
the dynamic absorber 39 comprises a counterbalance
mass 40 mounted with flange 45 between two springs
41 and 42 having small damping characteristics. As
a result, the axial motion of the compressor is
countered by the axial vibration from the
counterbalance mass 40 of the absorber 39. A
further description of dynamic absvrber operation is
found in U.S. Patent No. 4,783,968, of G. Higham
et al., filed August 8, 1986, entitled "A Vibration
,: ". , ~ :
:, -
2~
Isolation System fvr a Linear Reciprocating
Machine." The present system has bumpers 48 on the
front 98 and rear 47 spring supports to absorb any
impact of the absorber against the mounting frame of
05 the compressor. The absorber system is mounted onto
the housing extension 86 by ring nut 43. A spacer
44 is used to properly adjust the distance between
the front 98 and rear 47 spring supports. The screw
flange 46 is used to attach the flat spring 61 to
the end of the compressor,
The compressor system utilizes isolators
mounted on opposite ends of the compressor. The two
isolators have flat spiral springs 61 and 71 which
are soft in the axial direction while being very
stiff in the radial direction. The outer diameter
of the two springs 61 and 71 are attached to the
housing end plates 60 and 70 respectively. The
inner diameters are mounted onto flanges 64 and 72
and in turn attached to a screw flange 46 and
housing plate 31, respectively, using bolts 62 and
73. The inner and outer diameter of the two springs
are connected by a plurality of spiral arms. The
springs are mounted on elastomeric material 95 and
96 located at both ends of compressor 5 providing a
25 substantial level of damping to the isolator system.
A soft metallic gasket 30 is configured between
the plate 31 and flange 32 to seal the armature
volume 12 of the linear drive unit from the external
atmosphere. In order to detect the position of the
30 armature a sensor 80 is used to detect a target
--10--
magnet 81 fitted at one end of the armature 10. The
magnet 81 is mounted on an extended cy]inder 85 that
oscillates within an extention ~6 of the armature
housing 26 during motor operation. This extension
05 permits the utilization of an otherwise unused
volume within a countermass system 39 concentrically
disposed about the extension 86. By isolating the
magnet 81 and sensor 80 away from the stator 20, the
magnetic field of magnet 81 is decoupled from the
magnetic field of the stator magnet 220
A schematic illustration of the cold Einger 100
for a cryogenic refrigerator of the present
invention is depicted in Figure 2.
A nearly sinusoidal pressure variation in a
pressurized refrigeration gas such as helium is
provided through a supply line 102 from the yas
fitting assembly 15 on the compressor 5 of Figure 1.
Within the housing of the cold finger 100 a
cylindrical displacer 114 is free to move in
reciprocating motion to change the volumes of a warm
space 109 and a cold space 116 within the cold
finger. The displacer 114 contains a regenerative
heat exchanger 28 comprised of several hundred
fine-mesh metal screen discs stacked to form a
25 cylindrical matrix. Helium is free to flow through
the regenerator between the warm space 108 and the
cold space 24. A piston element 103 extends
upwardly from the main body of the displacer 114
into a gas spring volume 104 at the warm end of the
30 cold finger. The operation of the cold finger is
:
,
3~
more fully described in U.S. Patent No. 4,545,209
referenced above.
A semiconductor temperature sensor 120 is
positioned at the cold end of the displacer 114 for
05 measuring the temperature of the cold space 116.
The sensor 120 is used to monitor the cryogenic
temperature. The position sensor 80 is used to
measure armature displacement.
Another embodiment of a linear drive assembly
utilized in cryogenic refrigeration is illustrated
in Figure 4. This embodiment employs a stationary
piston 160 mounted to the housing on a flexi~le
tubular stem 162 having an axial bore 164 to provide
fluid communication between the cold finger and the
- 15 compression space 190. The tuhular stem 162 is
secured to the housing with a ferrule 168 brazed to
the outer surface of one end o~ the stem 162, with
nut 166 and seal 165.
Front and rear flexure supports 180 and 182 are
used to support the armature 170 relative to the
piston 160. The supports 180, 182 assist in
maintaining good axial alignment between the piston
160 and the armature 170. The inner cylindrical
element 172 of the armature 170 forms a clearance
seal 174 with the piston 160 that experiences
reduced wear due to the use of the flexible stem 162
and the flexure bearing supports 180 and 182.
A magnet 186 is attached to the armature 170
with a non-magnetic element 187. The magnet 186
30 reciprocates with the armature so that a sensor
.
,
3~
12-
positioned within the fluctuating magnetic field of
the magnet detects the position of the armature.
The sensor can be secured to the housing anywhere
within the vicinity of the magnet 186.
05 The interaction of these system components is
now more fully described with reference to the
schematic illustration of the control system in
Figure 3.
The signal at 136 can be, but is not limite~d to
a sinusoid of certain amplitude and at a frequency
optimal to a specific embodiment of the compressor.
It may, in principal be any frequency required, but
in practice may range between 10 and 200 Hz. The
frequency is singular to a specific design and is
maintained in close tolerance over the operating
range. The amplitude of the wave form is chosen to
represent 100% of stroke of the piston.
Consider, for example, how the system responds
when the measured cold end temperature is
considerably higher than the desired target
temperature 130. A number of waveforms are shown in
Figures 5A-5F to illustrate operation of the
circuit. The amplitude multiplier 134 will provide
no amplitude changes to the reference waveform. A
typical output signal for the amplitude multiplier
is seen in Figure 5A. The comparator 138 will
produce an error signal output when the position
feedback 80 as normalized, is not precisely the same
as through the amplitude multiplier. Figure 5B
illustrates the output of the normalized position
~: '
-13
feedback circuit 137. The nature of the error
siynal output of comparator 138 is not the typically
defined null condition. Rather, it is designed to
produce a w~veform of substantially the~ same
05 characteristics as the inputs. Figure 5C shows the
output of comparator 138 that provides one input to
the current amplifier 140. In the event of a true
equality between the compared signals, the
comparator 138 continues to provide output due to
the natural quadrature relationship existing between
command and feedback signals~
The net error signal provides a current command
input to a closed current loop at the current
ampli~ier 140. The feedback portion of the current
loop is a normalized measurement of motor current in
the linear motor 144. Various techniques can be -~
employed ko measure motor current, ranging from
simple voltage analogs of the current through a low
value resistor to isolating current transformers.
The normalization of the current feedback 148
establishes the measured current in a form
compatible with the current command input. Such a
normalized current feedback signal is illustrated in
Figure 5D. A conveniently expressed relationship
may be volts per ampere at the motor. Line 150
represents an output which relates measured current
considered detrimental to the system and arises due
to some malfunction. The action usually performed
in such systems is to shutdown, pending correction
2~ 4.'~
-14-
of the malfunction. The signal level is generally a
logic level change.
The optional line from 146 provides, if and as
necessary, a current proportional term to offset an
05 equivalent value of coupled energy reaching the
position feedback under influence of the motor.
Such conditions may exist in compact embodiments.
The signal at this point is amplitude proportional
to the coupled energy but opposite in phase for
cancellation.
The output of the current amplifier 140, as
illustrated in Figure 5E, i5 the net difference
between the current demand imposed by the position
loop and actual motor current. The difference
signal now forms the input to the Pulse Width
Modulated (PWM) Motor Drive. The signal level from
140 is generally of the same waveform as the
operating frequency and at a low power level.
Frequency terms may be added as a function of
20 various motor and resonant dynamics.
The composite current difference signal is
applied to a Pulse Width Modulation ~PWM) circuit to
convert the dynamic amplitude of the signal to one
capable of providing turn-on and turn-off inputs of
switching transistors in a circuit arrangement
commonly referred to as an H-Bridge. Other
arrangements can also be utili2ed. In the H-Bridge
configuration, signals from the PWM are used to
selectively turn on two switching transistors at one
time to effect curren-t flow from the positive
. .
:: :
.. . ~ : ~ . .
:
-15-
voltage, through the motor load, and to the negative
voltage. Following normal convention, the positive
going current flow through the motor establishes a
positive motion. The value of current flowing
05 through the motor is a function of the time the
transistor pair is turned on relative i:o the time
the transistor pair is turned off. Figure 5F
graphically illustrat s the output of the PWM motor
drive unit.
It is convenient to consider the repetitive
time period established by the switching frequency
as a single event equal to one (1~. If the time on
of the transistor pair is defined as d, the time off
is established as (1-d). The ratio of d/~l-d)
establishes, within the context of all circuit
parameters involved, the instantaneous current flow
to the motor. Since the value of d may be varied in
fine increments of microseconds, and may be changed
selectively at each new time period of the switching
frequency, a high level of control is achievable.
Power efficiency is achieved by the fact that
current flows through the turned-on switching
transistors in saturated, low impedance conditions.
When the transistors are turnsd off, little or no
current flows. Losses are inherently reducad,
relative to linear power control techniques.
It can be seen that positive current and
subsequent force may be applied in any manner
responsive to th~ dictates of the current amplifier
140 in Figure 3. When motion in the opposite
. ' ', ~ ' ~, ~
: i
-16-
direction is required, the actions described are
performed by an additional transistor pair which
during the positive cycle have been maintained in a
turned-off state. These transistors, when turned
05 on, connect the negative voltage to the motor load
and thence to the positive voltage, efEecting a
negative current flow and negative reactionO All
actions described beforehand are equally true. The
physical result may be considered equal but
opposite. In this manner, the piston motion is
effectively controlled to the desired instantaneous
values required for proper compressor operation.
Turning to the section of Figure 3 dealing with
control of cryogenic temperatures, the system
components involved are the sensor conditioning 120,
the temperature feedback 121, the temperature
feedback normalize 131, the comparator 132, and the
target temperature reference 130.
~ usual practice is the use of a silicon
semiconductor junction, properly conditioned, as the
sensing element of cryogenic temperatures. The
junction is located in close thermal proximity to
the cryogenic temperature source.
While in no way limited to a specific device, a
temperature sensor may be comprised of a silicon
transistor that is forward biased through the base -
emitter junction with a regulated DC current flow
of, for example/ 1 milliampere. The voltage from
base to emitter will be in the range of 1.06 volts
at approximately 77K At higher temperatures, the
.
':- ".,, :~ ~
5~
voltage will be lower. Conversely, at lower
temperatures, the voltage will be greater. In a
range above and below the 77K value, the increments
of voltage change versus temperature are reasonably
05 linear and monotonic. In close proximity of the
stated 77K, the changes are uncler 2 millivolts per
degree ~elvin.
In khe interest of minimizing the time required
to achieve cool-down, the strategy employed is to
maintain full stroke conditions from ambient to a
cryogenic temperature which is several degrees above
the target temperature. At that point, which may be
selected in design, the effect of the aforementioned
elements i9 to reduce the amplitude of the reference
waveform at the output of the amplitude multiplier
134 by voltage action of the signal line from the
comparator 132. In good practice, the range of
linear changes of amplitude as a function of sensor
defined cryogenic temperature will extend somewhat
below the target temperature to maintain loop
linearity before any saturation effects may occur.
Note that the temperature chosen to begin
amplitude reduction will effect the gain of the
temperature loop. A convenient expression for the
gain may be given as the percentage change of
amplitude per degree Kelvin. Stability of the
temperature loop may be established by proper
selection of the gain figure. Additional means of
stable operation may be typical in-circuit values of
reactive components in passive or active filter
'
,
. ~ ;
. .
3~'3
-18-
combinations, serving as a low-frequency
compensation network. The consequence of the
temperature loop is to properly adjust the stroke
amplitude while maintaining the defined operatiny
05 stroke frequency at a constant value.
In a preferred embodiment, the current,
position and temperature feedback signals are all
normalized to correct for noise and variations in
operation temperature. For example, a switch mode
amplifier within the PWM control generates noise
within the circuit which can be substantially
corrected by filter techniques. The linear drive
motor 144, switch mode amplifier and other elements
of the control circuit may generate sufficient heat
to perturb efficient operation of the control
circuit. Corrections for noise and temperature are
made at points in the system indicated by Tc and Nc
from circuits 122 and 124 respectively.
The control circuitry as described is analog in
20 nakure. All the same actions may be performed by
digital circuitry, provided that operational speeds
of the control circuitry is sufficiently high to
maintain required resolution of the system.
A digital version would preferably use a
25 microprocessor of sufficient data word capacity and
computing fre~uency. An eight bit processor in
double precision or a sixteen bit processor in
single precision can be used. Operating clock
frequency may be established in the 4 megahertz
30 range. Algorithms may be developed for the
.. . .
Pi.~5~
functi.onal requirements, for example, the reference
waveform 136 is modeled by move instructions that
incrementally describe the waveform as point by
point instructions. The position feedback may
05 remain as described, followed by an analog to
digital circuit of 21~ resolution with a settling
time of better than lO microseconds. Alternately, a
linear incremental encoder may be substituted.
'; : '' ' ~ ~ .
~.
;
