FR-4 with a layer of copper

Again, this is the board that we are re-testing and was created by Balwant Lall.
The board is 6x6 inches and the FR-4 layer is 1.57 mm thick and the copper layer is 0.036 mm thick. There is a 1x1 inch heat source located in the middle of the board as well as a heat flux gage.( We will not use the heat flux gage for our testing)

* Jeremy Evans numerical study of conjugate natural convection from a discrete, heated plate. IBM- Tucson 2003
Our false or new bottom that we machined to make connecting the thermal couples easier seems to be working. This will also be the location of our ZONE BOX. The zone box is the location where the thermal couples temperature will be used to calculate the temperature difference of the opposite junction point of the thermal couple.
We attached the thermal couples to the terminal
 strips. The connector strips will also be attached to a ribbon cable that will connect the thermal couples to a 3479 system. This will read the thermal couples with a digital volt meter.
*A thermal couple works when you have two different conductors forming a closed loop. A difference in temperatures at the junction points will cause there to be a voltage difference in the conductors. By measuring the voltage difference we can determine the temperature difference.
There are 18 thermal couples attached to the board. They are made of alumel and cromel and are attached in pairs, so there are 36 wires that are attached to the terminal strips to make up the 18 thermal couples.
Here is a better look at the thermal couple wires:
To insulate the bottom of the cube We fill it with styrofoam beads. This will help to keep the heat transfer limited to the top of the cube.

I have attached the thermal couple to one side of the terminal strip and then continues their connection to a panel that will be inserted into a 3497 data acquisition machine.
 

The 3497 machine will read the voltage difference across each pair of thermal couples with a digital volt meter. Then we also run a thermal couple into an ice bath where the digital volt meter reads that voltage difference. At a temperature of zero degrees Celsius voltage is zero. Therefor, the temperature difference between the zero degree ice point and the zone box will be added to the temperature difference between the zone box and the board to get the total temperature of the board.
We use a water circulator that will run water through the cold plates at a constant temperature that we select to be about room temperature at approximately 20 degrees Celsius.

You can see the circulator to the bottom right of the apparatus, and the hoses that run from the circulator to the cold plates.

Results

Lall's Temperature surface profile


Evans's Prediction

* Jeremy Evans numerical study of conjugate natural convection from a discrete, heated plate. IBM- Tucson 2003
As you can see, Evans's mathematical prediction for 2.4W is almost identical to Lall's 4.8W reading.
The same pattern is apparent for all of the Lall vs Evans graph.

What we found

Unfortunately when we collected this data, only six of the eighteen thermal couples were operational. So we compared the data from each specific thermal couple with Lall's reading for that thermal couple.
For our 0.4W readings we got similar results as Lall's 0.8W readings. As well, for our 0.8 readings there were almost identical to Lall's 1.6W readings. As our wattage increased, our temperatures increased more than they should have for Evan's predictions. We can attribute some of this to breakdown in the adhesive material used on the surface of the board to hold down the heat source as well as other components, but we are not completely sure why our temperatures got as high as they did.
I believe however that we did prove that Lall's numbers for wattage on the heat source were twice what they should have been for this particular board, as Evans predicted.