![](http://datasheet.mmic.net.cn/300000/RF-PABASESTATION_datasheet_16204667/RF-PABASESTATION_4.png)
using two tones RF signal technique. At the same time RF-electrical parameters (linearity, gain, and
efficiency) were monitored and combined with infrared (IR) thermographs. The loading measurements
were performed at temperatures below and above the typical recommended values. The RF-PA with
impedance matching circuitry (Device Under Test - DUT) were assembled on water-cooled plates with
adjustable flow rate. This feature provided possibility to adjust for several working conditions starting
from low to high and harsh temperatures. LabView based Data Acquisition System (DAQ) was used to
monitor stimulation parameters and perform thermal mapping by means of a set of thermocouples. The
DUTs were separated in 4 groups where each group was stimulated to reach predefined working
conditions. The experiment was adjusted to covered junction temperatures (at surface of the die) in range
of
T
jmax
=160
225
°
C. At the same time, the case temperatures (
T
c
- temperature on the surface of the
flange at the center of symmetry of component, see Fig 3b) were adjusted and varied in range of
T
c
=70
110
°
C. Several series of IR measurements were done to correlate the device power, maximum
junction temperatures and RF-electrical parameters. The DUTs were decapped and black painted for IR
imaging as shown in Fig 3a. Figure 3b shows cross section of the DUT and position of thermocouples for
thermal mapping and characterisation.
a)
b)
Figure 3. a) The RF-PA mounted into impedance matching circuitry (DUT) ready for stimulation and IR imaging.
b) Cross-section view on the DUT for temperature mapping by several thermocouples.
IR measurements with both types of the components (WCu and Cu-laminate) were performed with the
same stimulation and assembly conditions. Results of the measurements are presented in Fig 4. As can be
seen, the Cu-laminate component has nearly 29% reduction in
R
jc
and a more uniform temperature
distribution across the die compared to the standard WCu based device. Such advanced components were
further examined by long-term loadings at elevated temperatures. Results of the IR measurements for the
Cu-laminate component are presented in Fig 5.
Analyzing the thermal resistance in the heat-flow path of Cu-laminate components, the major contributors
can be proportioned as Si die (~35%), die bond (~10%), metal flange (~40%) and the thermal interface
between the component and next level heat-sink (~15%). In addition to a reduction of about 18
35% in
R
jc
, a more uniform temperature distribution across the die and the package was observed. Furthermore,
the Cu-laminate structure with active RF-PA component showed high out-of-plane stability resulting in
remarkable flatness of the flange. Such flatness allowed reduction of interfacial thermal resistances,
sufficient heat transfer and further lowering of
T
jmax
.
For a set of working conditions at elevated temperatures (
T
jmax
= 160-225
o
C
T
c
= 70-110
o
C) the RF
electrical parameters were monitored over long period (1.5 months) to observe stability at severe
loadings. Results of the RF stability experiments with Cu-laminate components are presented in Fig 6. At
elevated junction temperatures, operational window is typically limited by RF electrical properties.