0
Research Papers

# A Convection/Radiation Temperature Control System for High Power Density Electronic Device Testing

[+] Author and Article Information
Matthew Sweetland

Department of Mechanical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139-4307sweetlan@alum.mit.edu

John H. Lienhard V

Department of Mechanical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139-4307lienhard@mit.edu

Alexander H. Slocum

Department of Mechanical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139-4307slocum@mit.edu

The value of the device minimum test temperature and tolerance vary among manufacturers and across device types.

Solid thermal interface materials generally have liquid silicon or some other liquid imbedded in them, which leaves a residue when removed.

With component sizes shrinking, devices that may have had open areas on the back side are becoming covered with external interconnects.

Calculation based on fully turbulent flow over the entire width of the device.

The number of nozzles for each TTV is different, but the spacing, nozzle diameter, and offsets are identical, so for the same Reynolds number (or manifold pressure) the convection coefficient is the same.

A fast IR bulb typically has a 1–3 s settling time to reach 90% of the steady state.

The actual high temperature point varied over the range of devices due to external losses. The range varied from $70°C$ to $75°C$ with an air forcing temperature of $85°C$.

Supply voltage was known to an accuracy of $±50 mV$(13).

Teradyne, Inc. has a patent pending on this system.

J. Electron. Packag 130(3), 031012 (Aug 12, 2008) (10 pages) doi:10.1115/1.2966437 History: Received October 02, 2007; Revised February 10, 2008; Published August 12, 2008

## Abstract

Active control of die-level temperature is required during production testing of high power microprocessors in order to ensure accurate performance classification, but control is becoming more difficult as the device power densities increase. With power densities approaching $100 W/cm2$, the current passive control systems are no longer able to maintain the required temperature tolerance for production testing. This paper describes the design and testing of a temperature control system that combines high performance impingement cooling with higher power laser heating with application to packaged integrated circuit devices under dynamic testing conditions. Also presented are system design concepts and experimental results for typical microprocessor thermal test vehicles.

<>

## Figures

Figure 1

Typical cross section of a high power microprocessor device. Not all devices will contain all shown components.

Figure 2

TTV images with superimposed die and measured RTD positions

Figure 3

Base line test power profile. Peak power is 46.6 W. Test sequence consists of 0.1 s square waves.

Figure 4

Single and stacked nozzle modules. Number of nozzles in the array could be varied by changing the number of modules and by blocking off specific nozzle mounting holes.

Figure 5

Manifold pressure versus h¯c for the prototype system. A 5×5 nozzle array was used with the HPLD1 TTV and a 3×3 nozzle array was used with the HPD2 TTV.

Figure 6

Volumetric flow rate versus the manifold pressure under standard conditions (101.36 kPa and 21.4°C) for the 5×5 nozzle array used with the HPLD1 TTV and the 3×3 nozzle array used with the HPD2 TTV

Figure 7

Side view and cross sectional view of assembled laser/convection system. Manifold system and base support structures are not shown.

Figure 8

Schematic diagram of the control and data acquisition systems for laser/convection prototype system

Figure 9

Uncontrolled die temperature for a HDP2 TTV subject to a 23.3 W peak die power test sequence with a 41.37 kPa manifold pressure for the nozzle cooling system. The die temperature data are from all ten RTD sensors.

Figure 10

Uncontrolled die temperature for a HDP2 TTV subject to various peak die power scaled test sequences with a constant manifold supply pressure of 41.37 kPa. Data are from a single die centered RTD channel for peak die powers of 14 W, 23 W, 28 W, 46 W, and 56 W.

Figure 11

Uncontrolled die temperature for a HDP2 TTV subject to the same 23.3 W peak die power test sequence with nozzle cooling system manifold pressures of 20.69 kPa, 41.37 kPa, 62.05 kPa, 82.74 kPa, and 103.4 kPa. Data are from a single die centered RTD channel for each pressure.

Figure 12

Uncontrolled and controlled die temperatures for a HDP2 TTV subject to a 46.7 W peak die power test sequence with a 41.37 kPa nozzle cooling system manifold pressure. Data represent a single die centered RTD channel.

Figure 13

Controlled die temperatures for a PDP2 TTV subject to a 23.3 W peak die power test sequence at cooling nozzle manifold pressures of 20.68 kPa, 41.37 kPa, and 62.06 kPa. Data represent a single die centered RTD channel.

Figure 14

Controlled die temperature for HDP2 TTV subject to a 23.3 W peak die power test sequence with a 41.37 kPa manifold pressure for the nozzle cooling system. The die temperature data are from all ten RTD sensors.

Figure 15

Uncontrolled and controlled die temperatures for a HPLD1 TTV to a 46.7 W peak die power test sequence with 41.37 kPa cooling nozzle manifold pressure. Data represent a single die centered RTD channel.

Figure 16

Detailed view of controlled die temperature for a HPLD1 TTV to a 46.7 W peak die power test sequence with 41.37 kPa cooling nozzle manifold pressure. Data represent a single die centered RTD channel.

Figure 17

Control power sequence for HPLD1 TTV controlled temperature response shown in Fig. 1

## Discussions

Some tools below are only available to our subscribers or users with an online account.

### Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related Proceedings Articles
Related eBook Content
Topic Collections