Method and Circuit for the Detection of Solder-Joint Failures in a Digital Electronic Package

CROSS-REFERENCE TO RELATED APPLICATIONS [0001] This application claims benefit of priority to U.S. Provisional Application No. 60/657,101 entitled "Non- Invasive Real Time Method for the Detection of Solder Joint Failures" filed on February 28, 2005 and U.S. Utility Application No. 11/325,076 entitled "Method and Circuit for the Detection of Solder-Joint Failures in a Digital Electronic Package" filed on January 4, 2006 the entire contents of which are incorporated by reference.

BACKGROUND OF THE INVENTION Field of the Invention

[0002] This invention relates to a method and circuit for detection of solder joint failures in digital electronic packages such as Field Programmable Gate Arrays (FPGAs) and Microcontrollers and between the packages' ball grid array (BGA) and printed wire board (PWB) .

Description of the Related Art

[0003] Solder-joint connections from digital electronic packages such as FPGAs or Microcontrollers to Printed Wire Boards (PWB) are a major reliability problem. Modern Ball Grid Array (BGA) packages have several thousand pins and the number of pins on these packages is likely to increase over the next several years. The increased number of pins on the packages is necessary to support the evolving complexity of circuits; however, one of the drawbacks of the increase is reduced reliability. For many applications it would be useful to detect failures or precursors to failures in the solder-joint connections from inside the digital logic on the die, through the multiple internal interconnects inside the package and through the external

interconnects to circuitry on the PWB; together the "solder joint network".

[0004] As shown in Figure 1, an exemplary FPGA 10 includes at least one flip-chip 12 consisting of a die mount 18 and die 16 mounted inside a cavity 14 of a BGA package 22. Electrical components such as transistors, diodes and capacitors that are configured via wire interconnects on a die 16 together constitute the FPGA. The FPGA includes write logic that is connected through an output buffer to a pad on the die. Similarly an input buffer is connected from a pad to read logic. The buffers may be either inverting or non-inverting. Flip-chip 12 is placed inside BGA package 22 so that the solder balls 20 (also called solder bumps) inside the BGA package 22 touch pads, lands or vias of flip-chip 12 and they are soldered to form solder-joint connections. Vias lead from contacts

(not shown) of solder bumps 20 to an outside ball limiting metallurgy (BLM) 24 and primary BGA solder balls 26 to complete the FPGA 10. The FPGA is placed so that primary BGA solder balls 26 contact solder paste on metal LANDs 28 on a PWB 30. The assembled PWB is heated and the solder balls 26 melt and reflow to attach themselves to the metal LAND. The PWB is configured so that the metal LANDs are connected by vias and/or wiring 32 to one or more I/O nodes 34 for external circuitry on the PWB.

[0005] As shown in Figure 2, the integrity of the external solder joint can be evaluated by measuring the bump connection resistance of a BGA package 50 flip-chip mounted to a PWB 52 during the test and judging defects by the degree of change in the connection resistance. Wire segments 54 connect vias 56 inside the package and PWB

wiring 58 connects pads 60 on the PWB to connect solder bumps 62 attached to ball-limiting metallurgy 66 from the primary BGA in a "daisy chain", A meter 64 directly measures the resistance for all bumps 62 at the same time by either applying a voltage and measuring a current or vice-versa. Alternately, the wire segments and PWB can be configured to measure the resistance between two bumps 62 at a time, so increases in resistance due to cracking can be monitored.

[0006] The techniques for direct measurement of the solder-joint resistance have a number of limitations. The BGA package and PWB are ^λ blanks' or Mummies' configured for the 4-wire or continuous measurements, and therefore they are not the same as the operational devices. Specifically, the package does not include the FPGA flip- chip containing the operational logic gates and buffers. It follows that these tests cannot be performed on actual operational devices, either in the lab or particularly in the field. Finally, there are several instruments available to perform these measurements, however, besides being bulky and suitable for lab testing, these instruments cannot be used to perform real-time, in-use testing of fielded, operational FPGA BCA solder joint networks.

SUMMARY OF THE INVENTION

[0007] The following is a summary of the invention in order to provide a basic understanding of some aspects of the invention. This summary is not intended to identify key or critical elements of the invention or to delineate the scope of the invention. Its sole purpose is to present some concepts of the invention in a simplified form as a prelude to the more detailed description and the defining claims that are presented later.

[0008] The present invention provides a method for evaluating solder-joint integrity of digital electronic packages, such as FPGAs or microcontrollers, that have internally connected input/output buffers, and particularly for evaluating the integrity of the solder-joint network during normal operation of the digital electronic packages. The invention is particularly applicable to high density ball grid array (BGA) packages.

[0009] The invention is accomplished by applying a time- varying voltage through one or more solder-joint networks to charge a charge-storage component. Each network includes at least one I/O buffer on the die in the package and a solder-joint connection, typically one or more such connections inside the package and between the package and a board. The time constant for charging the component is proportional to the resistance of the solder-joint network; hence the voltage across the charge-storage component is a measurement of the integrity of the solder-joint network. In different embodiments the time-varying voltage may be generated on- or off-chip and measured on- or off-chip.

[0010] In a first exemplary embodiment, a digital electronic package, which includes a die mounted within a Ball Grid Array (BGA) package, is soldered to a printed wire board (PWB) . Write logic on the die is configured to apply a time-varying voltage through a first solder-joint network including an output buffer, any internal solder- joint connections and an external solder-joint connection between the package and PWB to charge a charge-storage component coupled thereto. Read logic on the die measures a voltage caused by the charging of the charge-storage component through a second solder-joint network including

an input buffer, any internal solder-joint connections and an external solder-joint connection, as a measure of the resistance of the first solder-joint network, hence the integrity of the first solder-joint network.

[0011] . These and other features as well as advantages of the invention will be apparent to those skilled in the art from the following detailed description of preferred embodiments, taken together with the accompanying drawings, in which:

BRIEF DESCRIPTION OF THE DRAWIlSfGS

FIG. 1, as described above, is a diagram illustrating the solder-joint network in a typical digital electronic package;

FIG. 2, as described above, is a diagram of a conventional direct resistance measurement to evaluate the integrity of the solder joints formed between the primary BGA of a blank BGA package and the PWB; FIG. 3 is a flow diagram for the indirect measurement of the resistance of the solder-joint network in an operational digital electronic package;

FIG. 4 is a voltage plot illustrating the charging of a charge-storage component for good and bad solder-joint networks;

FIG. 5 is a schematic diagram of an embodiment in which the write and read logic that constitute the test circuit are provided on the die inside the package;

FIG. 6 is a flow diagram of an embodiment for writing a test voltage through a solder-joint network and reading a voltage caused by charging the charge-storage component to determine a fault;

FIG. 7 is a voltage plot illustrating the charging and discharging of the charge-storage component for different

network resistances;

FIG. 8 is a schematic diagram of an embodiment in which the write and read logic that constitute the test circuit are provided off-chip outside the package; FIG. 9 is a plan view of a die illustrating a preferred location of the solder bumps to be tested;

FIG. 10 is a diagram of the digital logic on the die illustrating that the test logic may be multiplexed with the operational logic to use the same solder-joint network for test and operational purposes; and

FIG. 11 is a flow diagram for embedding a test core in a user application to program an FPGA.

DETAILED DESCRIPTION OF THE INVENTION [0012] The present invention provides a method for evaluating solder-joint integrity of operational digital electronic packages, such as FPGAs or Microcontrollers that have internally connected input/output (I/O) buffers and high density ball grid array packages, and particularly for evaluating the integrity of the solder-joint network during normal operation of the digital electronic packages. The invention will be described for a FPGA including a flip- chip mounted in a BGA package that is reflow soldered to a PWB, but it will be understood that the method and circuit for evaluating solder-joint integrity is generally applicable to any digital electronic package that has internally connected input/output buffers in the solder- joint network regardless of whether the die is configured as an FPGA, microcontroller or otherwise, or whether a specific mount method, such as a flip-chip is used, or whether a specific bonding method, such as wire bonding is used, or regardless of a specific BGA package type such as BGA or a Fine-pitch BGA, or whether a socket is used prior to assembly on the PWB.

[0013] As illustrated in Figures 3 and 4, a time-varying voltage is applied to and through a solder-joint network that includes at least one logic gate on a die inside the FPGA' s BGA package and at least one solder-joint connection

(step 100) . Typically, the network will include an I/O buffer, multiple internal solder-joint connections including the "buffer connection", "die connection" and

"flip-chip" connections, and the external "package connection" formed between the BGA package and PWB. In certain configurations, multiple networks will be evaluated at once. The time-varying voltage is suitably a pulse of the type in digital logic circuits but could be a ramp, saw tooth or other variable voltage.

[0014] The time-varying voltage charges a charge-storage component that terminates the network (step 102) . The charge-storage component is preferably a capacitor but could be an RC circuit, an inductor, LC, RLC or a memory cell and can be positioned either inside or outside of the package under test. In a microcontroller, analog components such as a capacitor are available and could be connected internally through MCU programming. Alternately, a special purpose FPGA including internal capacitors could be designed. The time constant for charging the component is proportional to the resistance of the solder-joint network. If each of the solder-joint connections in the network is good, the resistance will be low and a component voltage 104 will charge to a voltage level 106 relatively quickly. However, if any of the solder-joint connections are defective, ranging from complete failure to various stages of degradation, the resistance will be higher and a component voltage 108 will charge up to the voltage level more slowly.

[0015] The component voltage is measured (step 110) and used to evaluate the integrity of the solder-joint network

(step 112) . The voltage can be measured across the component, at a different point in the test circuit or by inductively coupling the voltage to a measurement device.

The voltage is typically measured at a time relative to the application of the time-varying voltage that would allow a good network to charge the component at or close to the voltage level but would inhibit a defective or degraded network from charging the component substantially. For example, if the time-varying voltage is applied as pulses, the component voltage may be read out near the falling edge of the pulse. The measured voltage can be either analog or digital (binary or multilevel) .

[0016] Because the measured voltage is proportional to the network resistance, the voltage provides a measure, albeit indirect, of the resistance. The presence of logic gates in the network prevents a direct resistance measurement. As described earlier, the network resistance is an indicator of the health and integrity of the network.

This information can be used in a deterministic, prognostic or diagnostic manner. For example, if the measured voltage is low, the resistance is high, and the network and the associated solder-joint network (or "pin") will be determined to have failed. As will be discussed below, if a test pin is determined to fail, the FPGA can be declared to fail or it can be used to predict the remaining useful life of the FPGA. Alternately, the low voltage and the number of occurrences of low voltage may be used to predict when the pin, hence FPGA, will fail. The resistance levels or failure modes of one or more test pins may provide some diagnostic information as well.

δ

[0017] This approach has particular value for FPGAs (or other packaged devices) that are soldered to the PWB and in the field and operational as part of a larger system. For example, let us assume the FPGA is a critical component of the guidance system for a missile. Clearly it would be undesirable to launch a missile or launch an aircraft with a missile that has a solder-joint connection failure or a substantial risk of such a failure that could affect the guidance capability. The described test procedure could be initiated automatically, say on-power up of the guidance system, or by a crewman prior to loading or launching the missile to verify solder-joint integrity. Now the FPGA would not ordinarily be configured to test every pin but would typically be configured to test those pins that are the most likely to fail. Therefore degradation or failure of any of these pins is a good predictor that the operational pins may fail. The capability to test the entire solder-joint network in an operational and fielded device is critical for many military applications and highly desirable for consumer applications, and is a capability that is not supported by known test procedures, namely the direct resistance measurement. The relevant question for the end user is whether the fielded and operating device is reliable, not whether the materials and manufacturing processes used to make the package are statistically reliable.

[0018] The described method of evaluating the solder- joint network can be implemented in the FPGA and a test circuit in a number of different configurations. The circuitry for applying ("write") the time-varying voltage and measuring ("read") of the component voltage can be configured in the FPGA (fig. 5) or external to the FPGA

under test (fig. 8) . Alternately, the voltage could be written on the die and read externally or vice-versa. For purposes of clarity and illustration, the internal chip mount and solder-joint connections are not shown but it is understood that those connections are part of the network and can be provided in many different configurations. Furthermore, only a few representative solder ball joints are depicted. It is understood that typical devices will have hundreds to thousands of solder balls connected through I/O buffers to write and read logic that implement the FPGA.

[0019] As shown in Figure 5, a FPGA assembly 120 configured for solder-joint network testing includes an FPGA 122 having solder balls 124 and 126 positioned underneath a BGA package 127 and reflow solder to LANDS 128 and 130, respectively, on a printed wire board (PWB) 132. A flip-chip 136 with an FPGA die 138 is mounted in package 127 in electrical communication with vias 140a and 140b. In a typical FPGA the die is configured to provide read/write capability at each of the output pins (solder balls) . The die includes write logic 144a, 144b that is configured to write a voltage pulse to and through an output buffer 146a, 146b, through the internal solder-joint connections (not shown) , vias 140a, 140b, Ball-Limiting Metallurgical contacts 148a, 148b and solder balls 124,126. Die 138 also includes read logic 150a, 150b that is configured to read a voltage at solder balls 126, 124 through input buffer 152a, 152b as a digital, typically binary, value. The sequencing of writes and reads, beginning and ending of tests, and recording and sending of fault information is controlled by Control Logic 142.

[0020] To implement the invention and evaluate the integrity of the solder-joint network 153a (Joint 1) or 153b (Joint 2) between the write logic 144a or 144b and a node 154, a capacitor 156 is connected between node 154 and ground potential 158 to terminate both networks. The write logic is configured to write a voltage pulse in response to a CNTRL signal, which can be generated on-chip or off-chip, to charge the capacitor. The read logic is configured to read the voltage across the capacitor at an appropriate delay with respect to the write pulse, evaluate the measured voltage, e.g. compare it to the write pulse, and, if appropriate, declare a fault. In this configuration, the voltage pulse is written through one pin (solder ball 124) and read through the other pin (solder ball 126) to measure the resistance of the written network Joint 1 153a. The test can be repeated in the opposite direction to test Joint 2 153b.

[0021] An exemplary test procedure for the FPGA 122 is illustrated in Figures 6 and 7. The FPGA is powered on

(step 160) and a determination is made whether to execute the test (step 162) . This determination could be done automatically on every power on or some other criteria or could be based on an external command signal. If no, the test exits (step 164) . Otherwise write logic 144a generates a voltage pulse 166 off the internal clock signal and applies it through Joint 1 to charge a voltage 168 on the capacitor 156 (step 170). Read logic 150b measures voltage

168 through Joint 2 off the same clock signal at some prescribed delay, which is determined by the amount of capacitance, expected network resistance and applied voltage pulse (step 172) . Once the voltage pulse is removed, the capacitor will discharge.

[0022] Input buffer 152a is suitably configured so that if the network resistance is low (good) the voltage will have charged sufficiently to exceed the buffer' s high-on trigger point and drive the buffer high. Conversely, if the network resistance is high (bad) , the voltage will be less than the buffer's low-off trigger point and drive the buffer 152a low. The read logic checks to see whether the applied voltage (high) and the measured voltage are equal, e.g. same logic state (step 174) . If not, the algorithm skips forward to step 186 to record the fault.

[0023] In this example, Joint 1 passes so the process is repeated to test Joint 2, which for this example is bad (high resistance) . Write logic 144b generates a voltage pulse 176 off the internal clock signal and applies it through Joint 2 153b to charge a voltage 178 on the capacitor (step 180) . Read logic 150b measures voltage 178 through Joint 1 (step 182) and determines whether the applied and measured logic states are equal (step 184) . The inequality of the applied voltage 176 and measured voltage 178 means that the resistance of Joint 2 is high, hence defective. The read logic records a fault and writes a ^λλ 0" through a different pin, e.g. pin 1, to force a hard reset of the capacitor voltage (step 186) . The hard reset is needed because of the long time-constant for discharging the capacitor through a defective network. If the voltage fails to reset (step 188) than at least a second pin has failed and a permanent fault is recorded (step 190) and the test is exited (step 192) . Otherwise the test can be repeated over a period of time for the same pair of pins to determine whether a failure happened only once, repeats sporadically, is a hard failure, etc. or conducted for

different pairs of pins. It is understood that the test procedure shown in Figure 6 is only exemplary and that many other sequences of steps can be envisioned to test the solder-joint networks in digital electronic packages.

[0024] As shown in Figure 8, the write logic 200 and read logic 202 (not necessarily integrated or discrete logic circuits) can be provided external to the BGA package 204. The FPGA die 205 is configured to connect an input buffer 206 for a first solder ball 208 to an output buffer 210 for a second solder ball 212 to define a solder-joint network 214 between the write logic and an external node 216 that includes a first solder-joint network (SJNl) between the write logic and the output of input buffer 206 and a second solder-joint network (SJN2) between the input to output buffer 210 and node 216. A capacitor 218 is connected between the node and ground potential to terminate the network.

[0025] Write logic 200 generates a voltage pulse that is applied to SJNl. Assuming that SJNl is good, the output of buffers 206 and then 210 will switch to a logic state high. This high voltage is then applied, much like as described in Fig. 5, to and through SJN2 to charge capacitor 218. Read logic 202 measures the voltage and determines whether there is a failure within SJN2. If SJNl is defective the read logic may or may not detect a failure. For example, if there is a "hard open" in SJNl, the input buffer will remain low and a low voltage will be applied to the capacitor, in which case the read logic will determine a failure, although it will not know whether the failure occurred in SJNl or SJN2. However, if the resistance of SJNl is merely high the applied voltage pulse may or may

not switch the input buffer high. If the resistance combined with any parasitic capacitance (SJWQ does not see the capacitor) causes the voltage seen by the input buffer to charge slowly, the buffer may not switch high in time to charge the capacitor before the read cycle, and thus a fault will be declared. Conversely, if the parasitic capacitance is low the buffers will switch high quickly and the read logic will not detect the failure in SJNl.

[0026] Other configurations of the FPGA and test circuit are available to implement the invention. For example, the FPGA could be configured to write the voltage pulse through an output buffer and solder ball to charge the capacitor with the read logic being placed outside the package. Conversely, the write logic could be external to the package to drive the solder ball and capacitor to present a voltage to the input buffer and read logic in the FPGA.

[0027] As mentioned previously, the FPGA would not ordinarily be configured to test every pin (solder ball) , to do so would consume excessive space on the die, be expensive and is unnecessary. In a typical FPGA a considerable number of the pins and portion of the available logic is not configured for the user application, as much as 40% or more. One approach would be to place test pins and the R/W logic throughout the FPGA. In this case, the test pins would see approximately the same mechanical, thermal and general operational stress as the operational pins and should fail at about the same time and at the same rate. Another approach, as shown in Figure 9, is to place the test pins 230 on the FPGA 232 where they will experience the highest stress and thus have the highest probability of failing first. These test pins then act as ^M a canary in the mine", when one or more of the test

pins fail it is a good indicator that operational pins will soon fail. Depending upon the application, the FPGA assembly and the FPGA mounted on it may be taken out of service immediately or scheduled for replacement prior to expected failure. In most applications, the greatest stress load will occur at one of the corners of the FPGA. Therefore, at least a couple test pins are positioned at each of the corners.

[0028] Although typical FPGAs have many unused pins available for configuration as test pins, as shown in Figure 10, an FPGA 240 may be configured so that a I/O pins 242 and 244 can support both the test and user applications. For simplicity, only the logic required to write one pin and read the other pin is shown. This is done by inserting a 2:1 Mux 246 between the application's write logic 248, the test logic 250 and the output buffer 252. If a test mode (TMODE) flag 254 is enabled the mux 246 passes the voltage pulse from the test logic 248 through the output buffer 252 to pin 242. Otherwise the mux passes the write pulses from the application write logic. Similarly, a 1:2 Mux 256 is positioned between the input buffer 258 and the application read logic 260 and test logic 250. If TMODE is enabled, the mux 256 passes the logic state at pin 244 from the input buffer to the test logic, otherwise it is routed to the application read logic.

[0029] In a FPGA nothing is initially connected. To configure an FPGA for a particular user application and solder-joint test application, the connections between the transistor circuits, input and output ports, gates, flip- flops, switches, power, ground etc. must be fused closed or

blown open to implement a particular design. An exemplary embodiment for programming an FPGA to include both a user application and the test circuit is shown in Figure 11. The first step is to determine application requirements for solder-joint testing (step 300) . These requirements may include the maximum and minimum number of pins to be monitored; location of those pins; the input clock frequency; the number of output fault signals; the fault- resolution (1 per FPGA, 1 per cluster of ports connected to each capacitor, 1 per port, 16 counts instead of 1 or maybe 32 or 64 ....nnnn); the mode selection - normal or test), etc.. Next, a test core of, for example, the type shown in figure 5, is designed to monitor specified pins at the specified frequency to produce the specified fault signal (s) and count (s) when faults (detected loss of integrity of one or more solder-joint networks) per the specified selection mode (step 302) . A Hardware Description Language (HDL) representation of the test core is written that maps the hardware design into a programming language (step 304) . The HDL representation of the design is compiled and simulated as though it were an FPGA (step 306) . Step 304 is repeated if any compile errors are detected. Steps 302 and 304 are repeated if any logic errors in the design are revealed by the simulation. At this point, the compiled test core software is embedded in user application software (step 308) . Presumably the user application software was designed and debugged off-line. Thereafter, user application software (including the embedded test cores) is synthesized to produce an executable hardware program (step 310) . The executable program is loaded into a hardware tool that programs the FPGA (step 312) . The FPGA is tested (step 314) and steps are repeated as desired and necessary to satisfy the requirement that the FPGA works as desired and designed.

[0030] While several illustrative embodiments of the invention have been shown and described, numerous variations and alternate embodiments will occur to those skilled in the art. Such variations and alternate embodiments are contemplated, and can be made without departing from the spirit and scope of the invention as defined in the appended claims.