GEF Thermal Tech Paper Final 040501

Thermally Conductive Plastic for PLC I/O Housings and Power Supply Housings

BYMartin PeltzJeffry Tillery

GE Fanuc AutomationRoute 29N and Route 606

Charlottesville, Virginia, 22911

1. Abstract (Define)

The use of thermoplastic housings is a proven cost-effective means of packaging industrial control electronics. Thermally conductive plastic housings are an important capability enhancement over conventional thermoplastics for those electronics. By increasing the output rating of the electronics, or by decreasing the electronic component temperatures, the product capabilities are more highly valued by the Customer. Thermally coupling the electronics to the more capable housing will enable the heat generated by the electronics to be dissipated over a larger surface area. The use of 6 Sigma DFSS analyses and actual test data will be used to predict and verify the efficacy of the thermally conductive plastic housings in an industrial control application.

The sections of this paper will follow a Six-Sigma DFSS outline (DIDOV): Define Identify, Design, Optimize, and Verify.

Key terms

6-Sigma DFSS Analysis, Thermally Conductive Plastic, Thermal Conduction, Convection, Thermal Impedance, Thermocouples, SDRC I-DEAS, ESC (Electronic Systems Cooling), PLC (Programmable Logic Controller), VersaMax, Genius, FEMs (finite element models), GEP EXC P0018

2. Identify

To successfully use thermal plastics in PLC housings, the following CTQs are to be satisfied. The thermal plastic material must: 1. Have the highest achievable thermal conductivity while maintaining non-electrically-

conductive 2. Be cost effective 3. Be capable of being injection molded with a short cycle time equal to or better than

existing cycle times with Cycolac/Cycoloy 4. Be laser markable with contrasts equal to or better than existing contrasts with

Cycolac/Cycoloy5. Be color matched to the existing in-use colors with Cycolac/Cycoloy6. Be capable of surviving a drop from a height of five feet7. Have a UL 94-V0 flammability rating

Utilization of this plastic will theoretically enable increased wattage capacity per module. To take advantage of the plastics’ thermal capabilities, the module will have to use a thermally-conductive interstitial pad at the interface between the heat generating components and the housing.

GE Fanuc Automation Confidential and Proprietary Information—For Internal Distribution Only

3. Design

The conductive heat transfer equations used for these analyses are credited to “Cooling Techniques for Electronic Equipment” by Dave S. Steinberg (Figure 1).

Equations:

A = specimen area normal to heat flux, m²

= experimental thermal conductivity, W/(m · K)

L = length in the direction parallel to heat flux, mQ = time rate of one-dimensional heat flow through the specimen, W

T = temperature difference, K

where: Taverage = (1 + Laverage = (L1 + L2) / 2 2A = occurs because the power flows through two plastic specimens

Figure 1

Using SDRC-ESC, and SDRC-IDEAS, finite element models where created (figures 2 & 3) comparing GEPs’ thermally conductive prototype plastic with other suppliers’ thermally conductive plastics. Although these FEMs used a Genius Module, rather than VersaMax configuration, assessing the thermal performance of the plastic is still possible.

Figure 2: Printed Circuit Board Thermal Analysis


Figure 3: Housing Thermal Analysis

In order to have the FEMs reasonably simulate and predict the thermal performance of the various plastics, numerous characterization tests were performed on materials from three suppliers: LNP, Cool Poly and GEP. The test setup (Figures 4 thru 10) was fashioned after the ASTM specifications. It provided the opportunity to compare the manufacturers published material specifications to the acquired test data.

Figure 4: ASTM Test


Figure 5: Layers of Testing Apparatus

Figure 6: Thermocouple Attachment Method


Figure 7: Insulation Layers and Test Sample

Figure 8: Magnified View of Insulation Layer with Heat Source


Figure 9: Orientation and Alignment of Insulation and Test Sample

Figure 10: Assembled Testing Apparatus


The results of the characterization tests were then inserted into SDRC-ESC for the finite element thermal analyses. Table 1 represents the analyses’ predictions vs. material:

Material Thermal Analyses’ ResultsComparisons of Zn, Al, Cool Poly, Konduit, and GEP # EXC P0018

Run Materials Boundary Conditions Results ( °C )Module Walls PCB Smart

SwitchesThermal

PadHeat Load Radiation

View FactorFlow Surfaces Module

WallsPC

BoardsSmart

SwitchesMaterial Emmisivity Thermal Conductivity (W/mK)

VIII N0. 7 Die Cast Zinc

0.8 113 FR4 4oz Copper

Ceramic Furon 15 W Note A Note B 82.1 82.8 83.6

IX 380 Die Cast Aluminum

0.8 96.2 FR4 4oz Copper

Ceramic Furon 15 W Note A Note B 82.7 83.2 84.2

X Cool Poly 0.8 20 FR4 4oz Copper

Ceramic Furon 15W Note A Note B 91.9 90.4 93.2

XI Konduit 0.8 1 FR4 4oz Copper

Ceramic Furon 15 W Note A Note B 116 113 117

XII 380 Die Cast Aluminum


Ceramic Furon 30 W Note A Note B 101 102 104

XIII GEP # EXC P0018


Ceramic Furon 15 W Note A Note B 115.5* 112.5* 116.5*

Notes:

A.) .5 on front face; 0 on base; 1 everywhere elseB.) Rough PCB's and all Module Walls except the front face, and the baseC.) * = extrapolated values based upon transfer function of thermal analysis model as derived from Minitab regression analysis of data

Table 1

4. Optimize

Given the results of the predictive thermal analyses and the temperature testing of samples of the various thermally-conductive materials, GEP forwarded enough raw material of #EXC P0018 to GE Fanuc’s molding supplier to manufacture sample VersaMax I/O housings. With these housings, a more realistic evaluation of the plastics’ relative performance could be achieved.

To evaluate the thermal performance of the GEP # EXC P0018-molded VersaMaxhousing relative to the Cycolac plastic housing that is currently in production, two IC200MDL331 “AC High Output” I/O modules were obtained. For each module, one of the Poron vibration dampeners was replaced with a Furon thermally-conductive pad, and four (4) thermocouples affixed as illustrated in Figures 11 and 12. The test assemblies were then placed in a temperature chamber and made operational. A Genius block was used as a data-logger for the thermocouples.

The first series of tests intended to evaluate the thermal transient capabilities of the housings. The chamber was programmed to cycle from 24 C to 60C with temperature measurements recorded every thirty (30) minutes for 48 hours.

The second series of tests evaluated the steady state characteristics of the housings. The test assemblies were placed into the temperature chamber, which was programmed to soak at the assemblies at 60 C until they and the chamber reached the same temperature. The test assemblies were then made operational, and data collection started. This test was run for four (4) hours. The test assemblies were then removed and examined for any separation of the thermocouples, as well as any damage caused from the effects of the heat. No detrimental effects were observed.


Figure 11: VerasMax I/O Housing

Figure 12: I/O Module Electronics Assembly

Housing Steady State Temperature Test Results :


Thermocouple

Thermocouple

Thermocouple location

Furon Thermal Pad

Heatsink

Power FET’s 4 on each side of heatsink

Thermal Pad

Thermal Plastic Housing

Poron Pad

A B

F

C

D

E

GEP EXC P0018 Cycolac X37Thermocouple #

1 2 3 4 5 651.8 40 38.9 53.3 40.5 40.656.3 49.3 48.3 56.9 50.4 50.758.5 54.9 54.3 58.7 56.1 56.259.6 57.8 57 59.3 58.5 58.959.9 59.3 59 59.7 59.8 60.260.2 60 59.7 60 60.5 60.960.5 60.6 60 60.1 60.8 61.260.5 60.9 60.3 60.1 61 61.460.5 61 60.6 60.1 61.2 61.560.6 61.1 60.7 60.2 61.2 61.660.7 61.3 60.8 60.3 61.4 61.860.6 61.2 60.9 60.2 61.4 61.760.6 61.2 60.5 60 61.3 61.760.6 61.2 60.9 60.1 61.3 61.660.7 61.2 61 60.2 61.3 61.860.7 61.3 61.1 60.3 61.4 61.8

Mean 59.5 58.3 57.8 59.3 58.6 59.0

% change

1 & 4 0.3%

2 & 5 -0.6%

3 & 6 -2.1%

Table 2

Table 2 Summary:

Given the fact that the test unit’s thermal design was originally and primarily for free-convection heat transfer, and once the test units were augmented to permit conductive heat transfer directly into the housings, the results indicate a measurable decrease in component temperatures. This confirms that the thermally-conductive plastic indeed conducts heat more effectively than the Cycolac X37. Transient thermal test results (not included herein) indicated no significant improvement in component temperatures, but the numerical improvements appear to be directly related to the increase in housing mass (see Table 4) associated with the P0018 material relative to the Cycolac.

DIMENSIONAL COMPARISONSDrawing CTQ Dimension Description Cycolac GEP EXC P0018

4.330 Overall length (A) 4.334 4.3542.610 Overall width (B) 2.610 2.635


1.994 Bottom Step width (C) 1.994 2.0031.970 Height (D) 1.959 1.9680.062 Color bar hole (E) 0.061 0.0614.101 Lens opening (length) (F) 4.105 4.110

Nominal dimensions shown See Figure 11

Table 3

Table 3 Summary:

The tooling for this particular housing was designed with Cycolac and Cycoloy as the two materials of intent. The P0018 material exhibits in the as-molded state slight to significant dimensional variations as compared to the Cycolac material. Changing materials will require changing the tooling to match the needed process capabilities of the product.

WeightsGEP Thermal Plastic Cycolac X37 Cycoloy C2800

73.6 grams 38.4 grams 44.3 grams

Table 4

5. Verify

Final verification of GEP EXC P0018 will not be achieved until the production plastic is injection molded into an I/O housing specifically designed for its usage (Genius X). The tests described herein were performed on a VersaMax Module Assembly that was designed primarily for natural convection heat transfer. The module heatsink has a thin edge and by default a smaller surface area. Consequently, this results in increasing that interface’s thermal impedance. To verify the effectiveness of the plastic, our new tests will increase the area of the conduction interface by increasing the contact surface area, and a Furon pad of comparable interface surface area will be included. Reducing the thickness of the thermal pad and increasing its surface area will decrease the thermal impedance--thereby increasing the heat transfer to the thermal plastic housing. The thermal tests previously performed in the Optimize phase will be repeated on new test samples. Verification and acceptance of the GEP EXC P0018 material will be based on the thermal performance verses final cost assessment, and in conjunction with the other required CTQ’s.


GEF Thermal Tech Paper Final 040501

Documents