KR19990032824A - Method and apparatus for controlling the temperature of the device during testing - Google Patents

Abstract

a) measuring a parameter relating to power consumption by the DUT such as current consumption during the test; And b) using the parameters related to the power consumption to operate the temperature control device with the parameters related to the power consumption to operate the temperature control device to compensate for the temperature change due to the change in power consumption by the DUT during the test. Comprising, the method of controlling the temperature of the DUT during the execution of the test. The control can be a closed loop or an open loop with control signals incorporated into the test program. a) a device for measuring parameters relating to power consumption by the DUT during the test; b) a temperature control device operative to control the temperature of the DUT during the test; And c) a device for controlling the operation of the temperature control device in accordance with the measured parameter relating to power consumption.

Description

Method and apparatus for controlling the temperature of the device during testing

The present invention relates to a method and apparatus that can be used to maintain a temperature of a semiconductor integrated circuit device during a test. Specifically, the present invention relates to a technique in which a device may be cooled and / or heated during testing to control the temperature of the device despite the varying levels of heat generated by the operation of the device.

Control of the DUT ("device under test") temperature at the time of test execution has been implemented for some time. For example, during burn-in, the DUT is placed in a temperature-enhanced environment (typically an oven) and applied signals for extended periods to cause device failures that may only occur after extended use. Burn-in processes use increased temperatures to accelerate very long processes in the DUT. In an oven burn-in operation, a large number of devices are placed on an oven and loaded onto a burn-in board that is tested together.

In other tests, it has been suggested to control the temperature of the DUT to simulate the possible ambient temperature during normal use. In this case, bulk temperature control is not suitable because the temperature of the individual DUT may change significantly over a short period of time. In addition, it is not possible to reliably or rapidly change the temperature of the DUT during the test process. Many proposals have been made for more precise control of the DUT temperature during testing. In high performance ICs, the significant effect of temperature changes during testing is to affect the evaluated maximum speed of the device in normal use, known as "speed binning". Inaccuracies in this assessment can lead to device failure in normal use.

U. S. Patent No. 5,297, 621 discloses a liquid container which is contained while the device is being tested. The liquid in this bin is inert and contains a mixture of two liquids with boiling points above and below the desired temperature. By varying the mixture of the two liquids, the liquid in the bin is arranged to have a boiling point at the desired operating temperature of the DUT ("set point temperature"). The heat generated by the DUT is dissipated by convection in the liquid reservoir and by nuclear boiling of the liquid on the DUT. Heat transfer from the DUT to the liquid is facilitated by placing a heat sink in contact with the DUT.

U. S. Patent No. 4,734, 872 discloses a system in which a flow of temperature controlled air is guided in a DUT. This air is introduced into the chiller where the dew point is low. The cooled air is then sent to a heater that raises the temperature of the air to the desired level to control the temperature of the DUT. Measurements of the DUT temperature and the air flow temperature are used to control the temperature of the air that eventually hits the DUT.

U.S. Patent Nos. 4,784,213 and 5,205,132 disclose variants of the system disclosed in the 872 patent described above. In both cases, the air stream is divided into a cooled stream and a heated stream. These two streams are then mixed in a reasonable proportion to produce a single stream that is guided in the DUT and has the desired temperature.

U. S. Patent No. 5,309, 090 discloses heating a DUT to apply signals / power to certain structures in the DUT to generate heat by their operation and evenly heat the DUT. This method does not allow the DUT to cool in the same way.

Recent developments in high performance microprocessor designs have resulted in increased power consumption and losses from about 10 watts to 60-70 watts. In addition, increasing the mounting density in microprocessor chips and the adoption of modern chip packaging structures have resulted in devices with extremely low thermal inertia, such as devices that heat and cool very rapidly. CMOS technology used in such chips is characterized by having power consumption and losses that vary with the activity of the device. In normal use, the chip dissipates heat generated by the operation of the device by having a cooling device, such as a fan mounted nearby or on a chip package. However, during the functional test, this cooling device does not exist and the power at loss during the very high speed functional test is sufficient to rapidly increase the temperature of the device to a level that can permanently damage the device.

All prior art methods of temperature control rely on direct measurement of the DUT temperature during testing to provide feedback control to the heat exchanger. There are many problems with this method. Due to the variability in the contact resistances encountered, it is difficult to reliably and consistently measure temperature at the surface of the device when testing in large manufacturing environments. Even in good temperature measurements, extrapolation of the device internal temperature in a high inertia package is problematic. In any feedback system, there is always a problem that the device must change temperature before the measurement can be made and the chip's thermal response time is as low as 30ms, while the response time of the heat exchanger is often in the range of 100-200ms. .

Thus, at best, such a device can only mitigate temperature changes and can result in undesirable temperature undershoots.

It is an object of the present invention to provide a method and apparatus which allows for temperature control of the device under test, in particular of the cooling under test.

1 is an IC tester incorporating the present invention,

2 is a block diagram of closed loop temperature control according to the present invention;

3 shows the operation of the temperature control device with respect to the current consumption (I _DD ) of the DUT, the temperature change and the time during the test;

4 shows one embodiment of a temperature control device,

5 shows another embodiment of a temperature control device, and

6 is a block diagram of an open loop control system according to the present invention.

One embodiment of the invention provides a method of controlling the temperature of a DUT during a test, comprising the steps of: a) measuring a parameter relating to power consumption by the DUT under test; And b) using the parameters related to power consumption to operate a temperature control device that compensates for temperature changes due to changes in power consumption by the DUT under test.

Another embodiment of the invention provides an apparatus for controlling the temperature of a DUT under test, comprising: a) means for measuring a parameter relating to power consumption by the DUT under test; b) a temperature control device operative to control the temperature of the DUT under test; And c) means for controlling the operation of the temperature control device in accordance with the measured parameter relating to power consumption.

A power supply (device power supply or "DPS") supplies current to the DUT, and measuring this current and, optionally, voltage can indicate the power consumption of the DUT almost instantaneously. Since all the actual power consumed by the device is represented by heat, and the relatively low thermal inertia of such a device, such measurements also tend to change temperature as a result of heat loss by the device. This method allows the temperature control signal to be well generated prior to the signal that may be possible using direct temperature measurements because there are parameters that are measured before the DUT changes the temperature. In high performance IC testers, the current consumption (I _DD ) of the DUT is routinely measured. It is this measurement that can be used to provide parameters related to the power consumption / loss of the DUT. The amount of temperature rise for a given current consumption depends on many factors, such as the IC package type, and forms a factor and sets the spot temperature. In many cases, the result is non-linear but can be determined by calibration for a given device type. If the device has relatively high thermal inertia, it is desirable to include a temperature sensor in the test device to calibrate the performance to correlate the I _DD with the device internal temperature. The use of I _DD measurements is particularly advantageous because it is a parameter that can be measured accurately and quickly, even during the use of solid test.

The particular form of temperature control device is a matter of choice. It is essential that the device can cool the DUT. In order to prevent temperature undershoot, it is necessary to include some form of heating to react the cooling effect at a suitable time. The choice of temperature control device affects how the control signal is used, but does not affect the principle of using power consumption measurements to control the operation of the temperature control device. Suitable devices may include the use of mixed heat conducting fluids (cold and hot water) to provide temperature control, or a combination of cooling fluid and electrical heating element.

It is also desirable to measure the actual temperature in the system while the main control of the temperature is provided by the power consumption measurement. The critical temperature is the temperature of the DUT, which depends on the effectiveness of the temperature control device. For example, if the temperature control device includes a heat exchange element that must be placed in contact with the DUT, relaxation of the contact surface and contact pressure can affect the effectiveness of the heat conduction. As a result, it may be desirable to include a DUT temperature sensor that can be used to validate the temperature control device and provide any offset required to ensure accurate control of the DUT temperature. In this way, measuring the junction thermal resistance prior to testing allows the offset to be determined. Other temperature measurements that may be made are measurements of the ambient or set point temperature, the temperature of the cooling fluid or the temperature of the heat exchanger.

When used in a closed loop control system, the measurement of power consumption provides for faster generation of control signals in place of intensive temperature measurements.

However, this method is still reactive and the slower thermal response time of the temperature control device is faster than the response time of the thermal control device. Another embodiment of the present invention is to use an open loop control system. This is possible because the operation of the DUT is known to advance somewhat ahead of the test pattern applied by the test program. Thus, the operation of the temperature control device can occur simultaneously with the test program. During the development of the test program, the measurement of I _DD is made and from this the power consumption as a result the temperature rise of the DUT can be determined. Another method is to measure the temperature of the device during each test segment while using this measured temperature profile to control the temperature of the device during development and testing. By determining the temperature profile during the test segment, the temperature control profile can be derived during the segment to maintain the DUT at or near the set point, which can be encapsulated in the test program. In this case, the operation of the temperature control device is controlled by the tester in the same way as the functional test applied to the DUT. When using this method, a change in the DUT temperature can be expected and the signal applied to the temperature control device can be controlled to avoid delays due to thermal response time prior to the change in the thermal change actually received by the DUT.

While active control of the temperature control device described above does not rely on measurements of DUT temperature or power consumption during testing, it is still desirable to make such measurements to account for the change in effectiveness of temperature control between tests.

In CMOS technology, which is commonly used in VLSI integrated circuits, the IC's current consumption and consequently, residual power consumption varies with device activity.

Typically, about 99% of this power is in heat, and the exact percentage depends on factors such as the specific device, geometry and packaging. In normal use, this heat loss is achieved with the use of heatsinks and fans attached to the chip so that the temperature is kept within defined boundaries. However, during testing, this heat loss device does not exist. As the power consumption of the device increases to nearly 100 W and the low thermal response time of the package typically accounts for about 30 ms, the likelihood that the temperature of the device temperature rises dramatically during testing is increased. The device is often tested at one or more setpoint temperatures, such as 0 ° C, 25 ° C and 100 ° C. The performance of the device at this temperature, especially at higher temperatures, is used to determine the speed bin to which some are allocated. The rise in temperature during the test itself changes the device in this regard and the yield of the device at a given rate. It is estimated that the effect on yield can be on the order of 0.2% / ° C above the set point encountered during the test.

A typical test environment is shown in FIG. An environmental chamber 10 is provided having an internal temperature that is controlled to be maintained at the desired set point temperature Tsp for the test. A tray or other carrier 12 containing several IC devices to be tested is loaded into chamber 10. The individual devices 13 are removed from the tray by a pick and place robot 14 and are in contact with a load board 18 to connect with the test head 20. Loaded onto a bed, the test head of Schlumberger Technologies, Inc., San Jose, CA. It forms part of a high speed tester such as the ITS 9000 GX available from. The robot 14 includes a temperature control device, which will be described in more detail below in connection with FIGS. 4 and 5 and is in contact with the DUT 13 until the test is completed. It is then returned to tray 12 of the DUT and another DUT is selected and tested.

FIG. 2 is a block diagram of a closed loop temperature control system for use with the apparatus of FIG. 1. Tester 20 includes a power supply (DPS) that provides a power supply of a predetermined voltage VI _{DD to} the DUT, and the drawn current I _DD is determined by the activity of the DUT. The DPS includes a sensor that outputs the measurement of I _DD during the test. The temperature control device includes a heat exchanger HX disposed in contact with the DUT, the heat exchanger HX including a temperature sensor for measuring the temperature Tj of the contact between the heat exchanger and the DUT.

The temperature Tj is the effective temperature of the DUT. The heat exchanger also includes a temperature sensor for indicating its own temperature Thx. The temperature measurements Tj and Thx and the current measurements I _DD are provided to the temperature control unit TCU together with the fixed temperature Tsp, the fixed temperature Tsp being derived from the sensor or inputted by the operator. Can be. The TCU outputs a control signal ΔThx to the heat exchanger HX to change the temperature of the heat exchanger, thereby affecting the DUT temperature by ΔT. Measurement of I _DD Provides an instantaneous measurement of power consumption to the DUT, which can be translated as a temperature increase (ΔTj) that can be seen in the DUT after a short time (thermal response time), for example after 30 minutes. As I _DD increases, the temperature of the DUT rises. As a result, the TCU outputs a signal ΔThx to the heat exchanger HX to lower the temperature ΔT and derive heat from the DUT. The amount of temperature change in the heat exchanger (HX) will depend on the current heat exchanger temperature (Thx), current DUT temperature (Tj) and expected temperature rise (ΔTj). When the current consumption is so low that there is no substantial rise in the DUT temperature, the heat exchanger temperature (Thx) is compared to the fixed temperature, and at lower temperatures the signal (ΔThx) causes the heat exchanger (HX) to under shoot. Will increase the temperature to prevent. The difference between the DUT temperature (Tj) and the heat exchanger temperature (Thx) can be used to indicate the thermal resistance of the contact between the heat exchanger and the DUT, thus providing a residual bias to compensate.

3 shows the change in I _DD , Tj and Thx during testing in the system as described above. The spikes in Tj depend on the slow response time of the temperature control compared to the temperature control of the DUT. These are much less than what is achieved with a simple thermometric feedback closed loop system (shown in dashed lines).

One form of temperature control device for use in the present invention is shown in FIG. 4. This includes a twin rope fluid system having heat exchange pads 30 in contact with the fluid chamber 32. When picked up by the pick and place robot 14, the pad 30 is in contact with the DUT and remains in contact until repositioned in the carrier 12. The fluid chamber 32 has inlets for the low temperature fluid ($ 3) and the high temperature fluid 36 and a common outlet 38. The fluid supply (not shown) has heaters and coolers to derive the temperature of the fluid to a predetermined level in the group and is used by the pump and valve system to adjust the mixing of the two fluids to achieve the desired temperature. do.

Systems for mixing temperature controlled fluids in this manner are known, see eg US Pat. No. 4,784,213. Another type of temperature control device is shown in FIG. 5. In this case, the chamber 40 has a fluid inlet 42 and a fluid outlet 44. Cooled fluid is sent out through this chamber. A heat sink pad 46 is connected to the chamber 40. Heat sink 46 has a heater pad 48 that includes a resistive heater in contact with the DUT. In this case the cooling fluid acts to back the heat from the pad 46 and a heater is provided to prevent undershoot when the DUT stops drawing high current. Such systems are described in US Pat. No. 5,420,521. It will be appreciated that the precise operation of the temperature control device is not very important in the present invention and that other types of devices can also be used effectively. That's how the temperature control unit is self-regulated in response to the important DUT power dissipation.

These systems are all closed loop systems and, therefore, inevitably have some temperature overshoot or undershoot as shown in FIG. 3. Another aspect of the invention seeks to overcome the problem of using an open loop method. 6 shows a block diagram of an open loop system according to the present invention. During the production test, the test program to be applied to the DUT is partially developed. At each step, the I _DD can be determined by direct measurement of the device under test or by a calculation that recognizes the power consumption characteristics of the device. As a result, for each test step, the current consumption profile can be determined. The profile of the temperature control signal needed to compensate for the temperature increase due to current consumption can be similarly determined using the same approach as used in closed loop systems. The temperature control profile is then summarized as part of the test program, and the temperature control unit is operated under the control of the test program rather than merely responding to the measurements obtained during the test. The profile can be time-advanced to compensate for the slow thermal response time of the temperature control device, which reduces the possibility of temperature overshoot or undershoot. The measurements of I _DD , Tj and Thx are still obtained during the test, to ensure correct operation of the system by determining the difference in variation between the tests.

Other variations of this approach can be used to substantially measure the device temperature when deploying test portions and using the measured temperature profile to control the operation of the temperature control unit. The measurements will be influenced by the package type of the device. Thus, it is necessary to make measurements for each particular package type. The relationship of the heat exchanger, the I _DD measurement and the temperature measurement is optimized using the appropriate software control for each part.

Claims (20)

CLAIMS 1. A method of controlling the temperature of a DUT during a test, comprising: a) measuring a parameter related to power consumption by the DUT during a test; And b) using a parameter related to power consumption to operate a temperature control device that compensates for a change in temperature of the DUT due to a change in power consumption by the DUT during the test. The method of claim 1, wherein step a) comprises measuring current consumption during testing. 2. The method of claim 1, wherein step b) comprises operating a temperature control device to maintain a temperature of the DUT near a predetermined set point during testing. 4. The method of claim 3, wherein the temperature control device is operated to reduce the temperature when the power consumption by the DUT increases and to return the temperature to a predetermined set point when the DUT power consumption decreases. 2. The method of claim 1, further comprising using the measurement to measure the DUT temperature and the temperature control device temperature to operate the temperature control device. 2. The method of claim 1, wherein step b) includes generating a control signal indicative of a change in temperature to control the temperature of the DUT under test by the temperature control device. 2. The method of claim 1, comprising test program instructions for performing step a) and operating a temperature control device during the formation of a test program, such that power consumption by the DUT during testing when testing the DUT using the test program. Operating the temperature control device to compensate for a change in temperature of the DUT due to a change in the. a) producing a test program comprising a series of control signals that cause the temperature control device to control the DUT temperature and a series of stimuli to be applied to the DUT; b) applying a stimulus to the DUT to measure its response; And c) applying a control signal to a temperature control device simultaneously applying a stimulus to the DUT to compensate for a change in DUT temperature due to a change in power consumption by the DUT due to the stimulus. 9. The method of claim 8, wherein step a) comprises using the current consumption to determine a current consumption by the DUT in response to the stimulus and form a series of control signals. 9. The method of claim 8, wherein step a) comprises producing a test program at a number of segments that are subsequently connected to each other to produce a program applied to the DUT. 11. The method of claim 10 including using the current consumption to determine current consumption by the DUT for each segment and form a control signal related to the segment. 12. The method of claim 11, wherein the determination of current consumption for each segment comprises applying a stimulus of the segment to a DUT to measure the resulting current consumption. 11. The method of claim 10 including measuring a device temperature profile for each segment and using the device temperature profile to form a control signal relating to the segment. 9. The method of claim 8, further comprising using the measurement of the DUT temperature to measure the DUT temperature and to control a temperature control device while applying the stimulus. a) a tester providing a series of test signals to the DUT; b) a temperature control device operative to control the DUT temperature during the test; And c) means for controlling the temperature of the temperature control device in accordance with a predetermined parameter related to power consumption by the DUT while the test signal is applied to the DUT. 16. The apparatus of claim 15, further comprising means for measuring a parameter relating to power consumption of the DUT during testing. 17. The apparatus of claim 16, wherein the means for measuring a parameter related to power consumption comprises means for measuring current consumption by the DUT during testing. The apparatus of claim 15, wherein the means for controlling the operation of the temperature control device comprises a sensor to determine the DUT temperature and the temperature control device temperature. 16. The apparatus of claim 15, wherein the tester comprises a memory storing test signals to be applied to the DUT and control signals to be applied to the temperature control device. 20. The apparatus of claim 19, wherein the tester coincides the application of the control signal simultaneously with the application of the test signal.