Bringing Together Mechanical and Electrical Engineers
By Jim Wilson
Electronics cooling design and analysis problems inherently bring together engineers from different backgrounds, especially mechanical and electrical engineers. Engineers with a heat transfer or fluid mechanics background usually studied these subjects in the mechanical engineering department. Understanding of semiconductors and related circuits usually involves engineers that have electrical backgrounds. My personal background is mechanical but I have worked with digital and analog electrical engineers for many years and the fairy tale for this column is that we all immediately understand each other. The related fact is that there is much to be learned by realizing that while there is much in common, communication is easier when we recognize that our backgrounds influence our terminology and perspective. A risk in writing a column about this topic is that it most certainly will not be comprehensive, but will only reflect some of the perspectives I have noticed that potentially create communication barriers.
Some terminology is immediately relatable between disciplines. Starting with the electrical-thermal analogy, the electronics cooling community has a long history solving conduction heat transfer problems by relating voltage and temperature, current and heat flow, thermal and electrical resistance, and capacitance for transient effects. Prior to the ease of performing computational simulations, circuit analysis was found to be an effective method of temperature prediction [1,2]. However, even a relatively simple concept like resistance can have a different perspective between thermal and electrical disciplines. From the electrical point of view, the scale range of electrical resistivity is very large. Not counting superconductors, electrical resistivity in ohm-m ranges from about 10-8 for metals to 1016 for insulators. This means that an electrical insulator as part of an electrical circuit can truly be treated as not having any current flow. From a thermal point of view, resistance to heat conduction is a function of material thermal conductivity and the comparable scale range of thermal conductivity is much smaller, from about 10-3 to about 103 in W/m-K. This smaller scale range means that thermal engineers do not have true insulating materials and typically must consider more of the physical domain for simulations. For example, an electrical voltage analysis of a circuit card would typically only consider the conducting traces and ignore the dielectric, while a comparable thermal simulation would most likely need to include both the metal and dielectric layers.
Electrical engineers that deal with power levels or signal strength, like Radio Frequency (RF) engineers, deal with the large scale range by using a log scale, most often using decibels. Expressing gain of a system in this manner is convenient, especially for systems that have a very large range of power levels. For example, an antenna system may transmit 1000s of Watts but only receive a few mW. Power levels are often expressed in units of dBm, or dBW, where the m refers to 1 milliwatt and the W refers to 1 W (for example 0 dBm is 1 milliwatt of power). Thermal engineers who interact with RF engineers should learn to communicate in dBs but the RF engineers should learn that thermal engineers have a preference for Watts. Sometimes power levels are measured to tolerances of dB and this can be challenging for thermal engineers. If a signal is measured to a tolerance of +/- 0.5 dB the percentage tolerance is about +/- 12%. Depending on the magnitude of the signal, this can cause grief to the thermal engineer trying to perform an energy balance. The RF engineer might be happy the measurement is within 0.5 dB of expectations but the thermal engineer would like a smaller uncertainty.
Aside from the thermal engineer who remembers solving 2D potential flow problems in graduate school, electrical engineers are usually more comfortable with using complex numbers. The use of the term impedance has transient implications, as in the impedance of an alternating current circuit expressed as the complex ratio of voltage to current. Some thermal engineers correctly use the term impedance in a thermal sense when describing the transient behavior of devices, especially power devices. However, sometimes thermal contact resistance is described using the term thermal impedance and this is incorrect. Thermal engineers would be better served to use terms like contact conductance or contact resistance for describing thermal interfaces to avoid confusion.
In keeping with the thermal facts and fairy tales theme, regardless of our backgrounds, communication with our coworkers is always easier when we take the time to understand the perspectives of others.
References
[1] Robertson, A.F. and Gross, G., “An Electrical-Analog Method for Transient Heat-Flow Analysis”, Journal of Research of the National Bureau of Standards, Vol. 61, No. 2, August 1958.
[2] Ellison, G.N., Thermal Computations for Electronic Equipment, Van Nostrand Reinhold Co., New York, 1984.
