Built-In Self-Test for 3 -Generation Mobile Users

advertisement
Built-In Self-Test for
rd
3 -Generation Mobile
Users
John Sunwoo
Electrical and Computer Engineering
Department
Auburn University, AL
Smart Phones
► Download
mp3 files and listen.
► Online shopping
► Online gaming
► GPS
► Mobile pay
► Wireless Key
► Start the car
2
Paying School Tuition
Are you sure you are paying $5,000, not $50,000?
3
Near future
4
Objective
► Extensive
usage of data transmissions via
mobile devices
 Tx/Rx of critical data
► Make
sure your mobile phone has no fault
 Users need to have access to high-level
functional test
 The testing operation should be easy
► BIST
for users
5
History of mobile communications
- 2G has limited data capability
6
Cellular network
- Base station is located each cell
- Base station has physical connection to
phone/data line
- One user connect to other users via base
stations
7
TDMA vs CDMA
► TDMA:
Time Division Multiple Access
 Allows a number of users to access RF
channel without interference by allocating
unique time slots to each user within each
channel
► CDMA:
Code Division Multiple Access
 Every communicator will be allocated the
entire channel all the time by having different
code than the others
8
2G to 3G… Test challenges?
► 3G
testing are related to the fact that it is
fundamentally different than testing 2G
networks
► Adjustment with just a power meter. (2G)
► Scan multiple neighboring base stations
for interference that may affect network
performance. (3G)
9
Case Study
10
Model
► Samsung
SPH-I500
PDA Phone
► $500 ~ $600
► CDMA
800Mhz/1900Mhz
► Built-in memory
32MB
11
USB Interface
12
Inside
►
►
►
►
CDMA Processor
Base band-to-radio frequency transmit processor
IF-to-base band processor
RF-to-IF processor
13
My BIST Approach
► Test
control: Via USB
► TPG: CDMA processor
► ORA: CDMA processor
► DUT: Transceiver circuitry
 RFT3100 -> Power amplifier -> RFR3300
-> IFR3300
14
Plan
BIST
START
TPG &
ORA
-BIST start
-No need of additional
hardware within certain
assumptions. (Making
assumptions means the
design is could be very
vague)
-Is it an effective test?
(Diagnostic resolution)
15
How did others tested RF device?
► Dr.
Chatterjee
 Test point insertion
algorithm for
determining the best
nodes for sensor
insertion
 Sensors outputs can
predict system and
module specifications
 Area overhead < 15%
16
Dr. Dabrowski (Sweden)
17
Drawbacks
► John
have never took RF classes. (Major)
► Qualcomm never responds my email for
asking the actual data sheet of the
MSM5100 modem.
18
Conclusions
► The
applicability of the presented BIST
has only for the higher-level model
 Mainly useful for hard faults such as spot
defects rather than parametric faults.
 Only applicable in a stable production
process or after the production.
► Exactly
what consumer want
19
Good and Bad
► Avoids
affecting the internal RF parts to
noise or external disturbances.
► Fault diagnosis is not possible.
20
Future work
► Bluetooth
Testing
 2.4GHz
 No published paper on Bluetooth BIST
21
Download