Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
2000-10-30
2002-02-26
Le, Dieu-Minh (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C714S029000
Reexamination Certificate
active
06351827
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to electronic test systems, and more particularly to margin testing of memory modules including SIMMs and DIMMs.
BACKGROUND OF THE INVENTION
Testing of electronic circuits and systems is of critical importance. Electronic systems are usually mass-produced, and a small percentage of the systems produced are expected to fail. Testing ensures that those failing systems do not reach customers.
Electronic systems are described by various specifications that detail voltages to be applied to inputs, timing of signals, and temperatures of operation. Gross failures are quickly detected by a large power consumption or inability to generate expected outputs when a sequence of inputs is applied. While such major failures are easily detected, more subtle failures can also occur. For example, the system can meet all specifications at a nominal temperature, but at the maximum operating temperature it fails some of the timing requirements. A higher than normal resistance in an internal signal path could cause such a failure. A higher than normal resistance causes greater signal delay. At an elevated temperature, the resistance becomes even higher causing an even greater signal delay. This could result in a violation of timing requirements such as setup and hold times. For example, a greater than normal delay for a specific signal that is part of a data bus will have more skew relative to the other bits of the data bus. The system could operate within specifications when the typical Vcc power-supply voltage (Vcc) is applied, but fails some timing specifications when the minimum-specified Vcc is applied.
Electronic systems or parts that have such subtle failures are known as marginal parts, since they fail only at the margins or extremes of the specified operating conditions. Detecting such marginal parts is desirable, since such parts, if undetected, could be used in larger systems and cause these to fail. Automatic test equipment has been used to detect such failures, by applying varying voltages to the parts being tested. The temperature of the parts under test can also be adjusted by heating or cooling devices.
One of the most important of electronic parts is the dynamic-random-access memory (DRAM). DRAM memory chips are often mounted on small, removable memory modules. The original single-inline memory modules (SIMMs) have been replaced with dual-inline memory modules (DIMMs), and 184-pin RIMMs (Rambus inline memory modules) and 184-pin DDR (double data rate) DIMMs.
The memory-module industry is very cost sensitive. Testing costs are significant, especially for higher-density modules. Specialized, high-speed electronic test equipment is expensive, and the greater number of memory cells on high-speed memory modules increases the time spent on the tester, increasing costs.
Handlers for integrated circuits (ICs) have been used for many years in the semiconductor industry. Handlers accept a stack of IC chips that are fed, one at a time, to the tester. The tested IC is then sorted into a “bin” for IC chips that have passed or failed the test. More recently, handlers have been made for memory modules. U.S. Pat. No. 5,704,489 by Smith, describes in detail a “SIMM/DIMM Board Handler” such as those in use today.
FIG. 1
shows a SIMM handler connected to a high-speed electronic tester. Memory modules
18
to be tested are loaded into the top of handler
10
in the input stack. Memory modules
18
drop down, one-by-one, into testing area. Module-under test MUT
20
is next to be tested. Arm
26
pushes MUT
20
laterally until it makes contact with contactor pins
16
that clamp down on “leadless” connector pads formed on the substrate of MUT
20
.
Contactor pins
16
are also connected to test head
14
, which makes connection to tester
12
. Tester
12
executes parametric and functional test programs that determine when MUT
20
falls within specified A.C. and D.C. parameters, and whether all memory bit locations can have both a zero and a one written and read back. Margin testing can be performed on some testers by varying voltages applied to different pins of the device being tested.
Tester
12
can cost from ten-thousand to millions of dollars. Cost can be reduced if a less-expensive tester replaces tester
12
. Since most memory modules are intended for installation on personal computers (PCs), some manufacturers test memory modules simply by plugging them into SIMM or DIMM sockets on PC motherboards. A test program is then executed on the PC, testing the inserted module. Since PCs cost only about a thousand dollars, tester
12
and handler
10
of
FIG. 1
are replaced by a lowcost PC. Equipment costs are thus reduced by a factor of a hundred.
FIG. 2
shows a PC motherboard being used to manually test memory modules. Substrate
30
is a motherboard. Components
42
,
44
, mounted on the top side of substrate
30
, include ICs such as a microprocessor, logic chips, buffers, and peripheral controllers. Sockets for expansion cards
46
are also mounted onto the top or component side of substrate
30
.
Memory modules
36
are SIMM or DIMM modules that fit into SIMM/DIMM sockets
38
. SIMM/DIMM sockets
38
(hereinafter SIMM sockets
38
) have metal pins that fit through holes in substrate
30
. These pins are soldered to solder-side
34
of substrate
30
to rigidly attach SIMM sockets to the PC motherboard. Both electrical connection and mechanical support are provided by SIMM sockets
38
.
Margin Conditions Would Cause PC Motherboard to Fail First
While using PC motherboards for testing memory modules greatly reduces equipment costs, margin testing is not performed. The SIMM sockets are integral with the substrate
30
of the PC motherboard, preventing variation of voltages applied to a memory module being tested in one of the sockets
38
. The power-supply voltage (Vcc) to the entire PC motherboard could be varied, causing the Vcc to the memory module under test in socket
38
to also be varied. However, since the PC motherboard has so many components, increasing the power-supply voltage to the PC motherboard would likely cause failures in the motherboard components before failures occurred in the memory module being tested.
Likewise, hot air could be blown on the memory module being tested in socket
38
. While this hot air would heat the module under test, it would also heat the PC motherboard and its components near socket
38
, perhaps heating all of the motherboard to some extent. This heating is likely to cause failures of components
42
,
44
, or of solder and wiring connections, before the memory module fails. Thus margin testing of a memory module being tested in socket
38
is problematic.
The parent application teaches a small daughter card known as a test adapter board that is attached to the reverse side of the PC motherboard. The reverse-side attachment of the test adapter board facilitates attachment of the SIMM/DIMM handler, since the front side of the PC motherboard is too crowded for attaching the handler. The inventors realized that the back or solder-side of the PC motherboard is less crowded and provides unobstructed access.
The PC motherboard is modified to provide reverse attachment of the handler to the solder-side of the PC motherboard using the handler adapter board. The SIMM socket on the component side of the PC motherboard is removed, and the handler adapter board is plugged from the backside into the holes on the PC motherboard for the SIMM socket.
Handler Mounted Close to PC Motherboard—
FIG. 3
FIG. 3
shows a SIMM/DIMM handler mounted close to the backside of the PC motherboard using the handler adaptor board. Handler
10
is not drawn to scale since it is several times larger than a PC motherboard. However,
FIG. 3
does highlight how handler
10
can fit close to the removed SIMM socket. Such close mounting reduces loading and facilitates high-speed testing.
Contactor pins
16
within handler
10
clamp down onto leadless pads on the edge of module-under-test MUT
20
when arm
26
pushes MUT
20
into place for
Co Ramon S.
Lai Tat Leung
Nguyen Thang
Auvinen Stuart T.
Kingston Technology Co.
Le Dieu-Minh
LandOfFree
Voltage and clock margin testing of memory-modules using an... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Voltage and clock margin testing of memory-modules using an..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Voltage and clock margin testing of memory-modules using an... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2953709