Background
In state-of-the-art high density electronics assemblies, slice-form electronics with attached conduction cards are fastened into chassis using card retainers, or wedgelock type clamps. The card retainer provides a mechanical connection that fixes the card in the chassis and provides a thermal path for heat rejection from the card. The ease of integration and maintenance of these electronics packages has resulted in industry adoption of these assemblies and development of standardized chassis and card dimensions. These standard assemblies, often referred to as embedded computing systems, require thermal performance improvements to remain viable solutions for growing industry requirements. Improving the thermal performance of the card retainer is one key to improving the thermal performance of these standardized systems for demanding applications.
Conventional card retainers provide an excellent means of clamping conduction cards into embedded computing chassis. With this attachment approach, the electronics cards can be inserted, operated and removed multiple times for testing or field replacement. Most designs consist of multiple wedges that are held together with a bolt. In the relaxed state, the wedges separate from each other and allow the clamp height to be minimized (Figure 1A). They operate by exerting an outward force as a result of wedges sliding against one-another as they are drawn together by torquing the bolt (Figure 1B). As the wedges expand outward, some wedges contact the conduction card while others contact the chassis. As shown in the right view of Figure 1B, this card retainer is expanded vertically. These simple assemblies provide a reliable means of mechanically attaching the conduction electronics card into the chassis.
Figure 1 A conventional card retainer in a relaxed state (A) and expanded (B). Here the primary direction of expansion is vertically.
The primary function of the card retainer is to secure the conduction card to the chassis. However, another function of the retainer is to provide a thermal connection between the conduction card and chassis. Conventional designs of unilaterally expanding card retainers rely primarily on heat flowing directly from the card to the chassis. The retainer itself provides a weak thermal path between the card and chassis due to 1) multiple metal-to-metal interfaces in the conduction path and 2) limited contact area.
The ICE-Lok®
The patented ICE-Lok® card retainer is designed to enhance card-to-chassis conduction by enhancing the heat flow through the card retainer by making additional contact between the card and the chassis. The ICE-Lok®, shown in Figure 2, is a wedge-type card retainer, but with the wedge faces cut at a compounded angle. When actuated, this allows the wedges to move with two degrees of freedom, expanding the card retainer in two, rather than one direction. The ICE-Lok also uses “L” shaped wedges to 1) make a direct thermal connection between the card and chassis through a single wedge and 2) significantly increase the contact area between the card, clamp, and chassis. This approach bypasses the key wedge-to-wedge thermal resistance of a conventional card retainer and contacts four faces rather than two as with conventional card retainers. Figure 3A shows a side view of the ICE-Lok in a relaxed state. Here is has a minimum height similar to that of conventional card retainers. Figure 3B shows the ICE-Lok actuated, which is achieved by turning the screw and expanding the wedges both vertically and horizontally. It should be noted, a wall perpendicular to the ICE-Lok mounting surface is required on the electronics card to provide the fourth surface for ICE-Lok contact. Accurate dimensioning of these walls is required for proper actuation of the ICE-Lok. Figure 4 shows the additional wall geometry required for ICE-Lok® actuation.
Figure 2 The ICE-Lok® Card Retainer by Advanced Cooling Technologies, Inc.
Figure 3 A) ACT’s ICE-Lok® relaxed and B) expanded. Here expansion happens horizontally and vertically to increase contact area between the chassis and conduction card.
Figure 4: Required electronics card geometry for proper ICE-Lok® actuation
In order to get the thermal benefits of the ICE-Lok®, both the Card Guides and Conductance Card need to make physical contact with ICE-Lok®, and by extension with each other. See the below graphic for the dimensional specifications and tolerances for both Card Guides and Conductance Cards and for both the 250 and 375 Series.
Thermal Performance
The ICE-Lok® is designed to be a low thermal resistance card retainer. To demonstrate this, ACT fabricated a thermal test apparatus to meet the dimensional requirements of the VITA 48.2 specification. A mock conduction card is inserted into the chassis and clamped in place. Heaters are sued to simulate waste heat from electronics. The heat is rejected from the system using liquid cold plates attached to the chassis walls. The thermal test apparatus is shown in Figure 5. Temperatures are recorded at the edge of the card and on the chassis wall as shown in Figure 6. Calorimetry was performed on the liquid cooling loops to determine the actual power transferred through the card-to-chassis interface rather than lost via convection. The thermal resistance is calculated by dividing the temperature gradient from the card to the chassis by the transferred power.
Figure 5: Thermal Test Apparatus
Figure 6: Thermal test thermocouple locations
The ICE-Lok® and two similar commercial off-the-shelf card retainers were tested in the thermal testing fixture for a side-by-side comparison. Each card retainer was tested at four different installation torques to determine the effect of installation toque on card retainer thermal performance. The thermal results are summarized in Figure 7. It is evident from the test results the ICE-Lok® performed 29% and 40% better than the standard card retainers A and B respectively. Care was taken to maintain consistency in each test setup to limit errors due to setup differences.
Figure 7: Card Retainer thermal performance testing results
Clamping Force
The forces generated by the ICE-Lok® are a function of multiple design parameters, including the number of wedges, the wedge geometry (angles), the bolt diameter, the installation torque, and surface conditions which impact surface friction factors. Figure 8 shows a free-body diagram for the half-wedge of the ICE-Lok and the associated forces for the individual wedge, which are given mathematically in Equations 1 through 4.
Eq.1
Where (τ) is the applied torque to the bolt, (μ) is the bolted-joint friction factor, and (D) is the bolt major diameter.
Eq.2
Where (θ) is the apparent angle of the wedge when viewed normal to the Y-Z plane of Figure 8.
Eq.3
Where (φ) is the apparent angle of the wedge when viewed normal to the X-Z plane of Figure 8.
Eq.4
Where (θ) is the apparent angle of the wedge when viewed normal to the Y-Z plane of Figure 8, and (φ) is the apparent angle of the wedge when viewed normal to the X-Z plane of Figure 8.
The total forces exerted by the assembled ICE-Lok®, outward onto the card and chassis are all in the “horizontal” and “vertical” directions. These depend on the total number of wedges and are given in mathematical form in Equations 5 and 6, where the number of wedges (Nwedges) counts both half and full wedges equally. This free-body diagram analysis neglects the effect of wedge-to-wedge, and wedge-to-card/chassis friction due to the fact that the direction of resulting friction forces is indeterminate due to the sliding contact between components.
Eq.5
Where (Fvertical) is the force shown in Figure 8.
Eq.6
Where (Fhorizontal) is the force shown in Figure 8.
Figure 8. Free-body diagram for a half-wedge within the ICE-Lok® card retainer.
Friction Locking Feature
The forces produced by the ICE-Lok® place mechanical loading on the walls of a chassis, which may be intentionally thin-walled for reduced weight and reduced thermal resistance to external cooling. Uncontrolled outward forces can cause chassis deflection that could misalign geometric features or damage the chassis by deformation. However, the ICE-Lok has been designed to achieve a friction-locking effect in which the conduction card tab will be locked in place, relative to the chassis guide.
The ICE-Lok® can be designed such that the outward forces developed in both the horizontal (toward side-wall) and vertical direction (toward chassis guide rails) are different. This is achieved by varying the apparent angle of the wedges, (θ) and (φ) in Figure 8. Viewing Figure 9, if the angles are chosen such that the ratio of the force developed in the vertical direction (blue arrows) and the force developed in the horizontal direction (red arrows) exceed the criteria shown in Equation 7, a “locking” effect can be achieved which uses friction (yellow arrows) to resist separation of the card tab from the card guide.
Figure 9. Schematic showing friction locking concept
Eq.7
Where (Fvertical) and (Fhorizontal) correspond to the forces shown in Figure 8, and (μ) corresponds to the friction factor between the ICE-Lok® L-brackets and the card/chassis.
Locking the card tab to the chassis increases the stiffness of the assembly. During operation, the card tab will be loaded in tension, making it resist the expansion of the chassis side-walls – minimizing deflection to acceptable values. Operation of this concept has been verified through testing and shown that the ICE-Lok® deflects thin chassis walls minimally.
Retention Testing
Although the ICE-Lok® is designed for thermal performance, the mechanical performance of the card retainer must be maintained. To demonstrate this, the parts were tested for retention. A dedicated test apparatus, Figure 10, has been developed which utilizes a linear actuator and a custom control box to forcefully attempt the removal of a card that has been clamped into a chassis by the card retainer. A mock chassis is bolted to a fixed testing bed, and a conduction card with ICE-Lok® card retainers is installed into the chassis. A linear actuator is attached to the chassis in series with a load cell and a spring, which is used to control the response of the chassis to applied load. The linear actuator retracts to pull forcefully on the conduction card while a fixed displacement sensor measures the position of the conduction card during the testing. The position sensor reading is recorded before and after 300 pounds of force have been applied, with the net deflection recorded. A successful clamping test will show a zero-deflection response (<0.0005”).
Figure 10. Retention testing apparatus
This test was performed on a freshly assembled ICE-Lok® and again after the ICE-Lok® was inserted, torqued, loosened, and removed into a chassis/card assembly 100 times. This was done to simulate the beginning and end-of-life of the ICE-Lok. Retention testing results for the ICE-Lok card retainer are shown in Table 1 and verify that the clamping ability of the retainer is sufficient and is maintained over the life of the device.
Table 1. Retention testing results for ICE-Lok® card retainer
Device / Trial | Prior Cycles | Torque (in-lbf) | Force Applied (lbf) | Net Deflection (lbf) |
ICE-Lok/ 1 | 1 | 20 | 285 | 0.000 |
ICE-Lok / 2 | 1 | 20 | 338 | 0.000 |
ICE-Lok / 3 | 1 | 20 | 308 | 0.000 |
ICE-Lok / 1 | 100 | 20 | 283 | 0.000 |
ICE-Lok / 2 | 100 | 20 | 293 | 0.000 |
ICE-Lok / 3 | 100 | 20 | 302 | 0.00 |