标签: heat

Most electronic, mechanical, and thermal engineers think of how to keep the temperature of their IC or printed circuit board below some maximum allowable value. Others are more worried about the overall enclosure, which can be range from a self-contained package such as a DVR to a standard rack of boards and power supplies.   Basic techniques for getting heat from an IC, board, or enclosure involve one or more of heat sinks, heat spreaders (PC-board copper), head pipes, cold plates, and fans; it can sometimes move up to more-active cooling approaches including air conditioning or embedded pipes with liquid flow. That's all well and good, but obviously not good enough for the megawatts of a "hyperscale" data center. (If you are not sure what a hyperscale data center is, there's a good explanation here ). While there is no apparent formal standard on the minimum power dissipation to be considered hyperscale, you can be sure it's in the hundreds of kilowatt to megawatt range.   But where does all that heat go? Where is the "away" to which the heat is sent? If you’re cooling a large data center, that "away" to hard to get to, and doesn't necessarily want to take all that heat you are dissipating.   A recent market study from a BSRIA offered some insight in the hyperscale data-center cooling options and trends. I saw a story on the report in the November issue of Cabling Installation Maintenance , a publication which gives great real-world perspective into the nasty details of actually running all those network cables, building codes, cabling standards, and more. (After looking through this magazine you'll never casually say, it’s "no big deal, it’s just an RJ-45 connector" again.)   BSRIA summarized their report and used a four-quadrant graph (below) of techniques versus data-center temperatures to clarify what is feasible and what is coming on strong. Among the options are reducing dissipation via variable-speed drives and modular DC supplies, cooling techniques from liquid cooling to adiabatic evaporative cooling, or allowing a rise in server-inlet temperature. The graph also shows the growth potential versus investment level required for each approach; apparently, adiabatic/evaporative cooling is the "rising star."   Cooling approaches for hyperscale data centers encompass basic dissipation reduction, liquid cooling, and adiabatic/evaporative cooling, according to this analysis from BSRIA Ltd, with the latter a "rising star."   When you are worried about cooling your corner of a PC board, it's easy to forget that it's not enough to succeed in that goal; you have to think of the next person who will have to deal with the heat which you so nicely spirited away. That's why I am often wary of PC-board heat spreaders, unless the design has been thermally modeled "across the board", so to speak: they move the heat from your IC to the next one, and so make their thermal headache more difficult.   Although I know I won't be involved with design of such hyperscale cooling, I need to learn more about thermal principles, including adiabatic/evaporative cooling. It still hurts that a very long time ago, when I was told to take an engineering course on "thermal physics" to learn basics, I was misled. It turned out the course was about the personal thermal lives of atoms and molecules, and had nothing at all to do with "thermal" as engineers need to know it: heat, heat flow, thermal modeling, temperature rise, cooling techniques, and more. In contrast, "thermal physics" is what Einstein used in one his is five 1905 papers, " Über die von der molekularkinetischen Theorie der Wärme geforderte Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen " ("On the Motion of Small Particles Suspended in a Stationary Liquid, as Required by the Molecular Kinetic Theory of Heat"), but hey, he wasn't worried about cooling a hot component!

Heat dissipation is often tied with power usage; that's not news. But sometimes, in our focus on minimising power consumption and thus subsequent dissipation, we forget that there are situations where properly managed dissipation is the whole point of the design.   This became very clear after I read a fascinating article in a recent issue of IEEE Spectrum about the ineffectiveness of the standard home oven. The article, " Recipe for a Better Oven " by Nathan Myhrvold (yes, the former chief scientist at Microsoft) and W. Wayt Gibbs, discussed the many shortcomings of the standard home and even commercial baking oven.   It was both illuminating and somewhat discouraging for many reasons. It's not just that the oven is wasteful in terms of energy use, it's that it also does a poor job despite that inefficiency; whereas I had hoped that the inefficiency was a necessary part of the tradeoff for getting the baking done right.   I already knew that the simple on/off control of the gas line or electric power was a marginal approach. Anyone who has studied any control theory knows that a loop with on/off control, rather than true proportional control and a PID (proportional-integral-derivative) control strategy, is not good at keeping the error between setpoint and actual value small, unless you are willing to switch that on/off valve at a fairly high rate.   While it is true that thermal situations tend to have relatively long time constants and so can accept a somewhat lower cycling rate than some other situations, oven vendors like to keep the valve's cycle time to a few minutes. So it is common to see swings of ±50°F (about ±25°C) around the setpoint. That's just not good enough for many recipes or foods.   The problems of the standard oven design and subsequent performance go well beyond the basic temperature-control loop. Only after I read the article did I realize how much I hadn't known about the implications of oven design and the cooking process.   For example, the heat source doesn't cool the food directly; instead, it heats the oven walls (metal or brick), which then re-radiate the heat to the food. The article explained how this leads to very uneven heat zones in the oven for conventional metal-wall ovens, and numerous other problems. (The article did highlight some ways to improve an oven's performance, which was nice to see.)   These other problem are not trivial, either, as they have to do with how different foods are affected by the radiant heat of the oven walls versus direct heating; not surprisingly, it does make a difference, just as different types of electrical load impedances affect power supply performance. (There are times when I felt the whole discussion of oven idiosyncrasies was akin to the mysteries of black-body radiation , which were finally resolved with the development of quantum theory in the 1920s by Max Planck and others.)   The irony of the oven problem is that for most engineers, the primary concerns regarding heat are twofold: minimizing generation of it in the first place, through the use of low-power circuits and high-efficiency supplies; and maximizing its removal, using convection, conduction, and radiation via use of fans, heat sinks, cold plates, and even advanced active techniques. It takes a 90° or even 180° shift in thinking to begin to understand heat and its effective delivery as an outcome rather than as an obstacle.   Have you ever been faced with a design challenge that required that you reroute your established way of looking at something, to a very different course? Did you realize the differences right away, or did you have to really step back to understand the many inherent assumptions you were making and how they had to be changed as well?

Heat dissipation is typically linked with power usage; that's not news. But sometimes, in our focus on minimising power consumption and thus subsequent dissipation, we forget that there are situations where properly managed dissipation is the whole point of the design.   This became very clear after I read a fascinating article in a recent issue of IEEE Spectrum about the ineffectiveness of the standard home oven. The article, " Recipe for a Better Oven " by Nathan Myhrvold (yes, the former chief scientist at Microsoft) and W. Wayt Gibbs, discussed the many shortcomings of the standard home and even commercial baking oven.   It was both illuminating and somewhat discouraging for many reasons. It's not just that the oven is wasteful in terms of energy use, it's that it also does a poor job despite that inefficiency; whereas I had hoped that the inefficiency was a necessary part of the tradeoff for getting the baking done right.   I already knew that the simple on/off control of the gas line or electric power was a marginal approach. Anyone who has studied any control theory knows that a loop with on/off control, rather than true proportional control and a PID (proportional-integral-derivative) control strategy, is not good at keeping the error between setpoint and actual value small, unless you are willing to switch that on/off valve at a fairly high rate.   While it is true that thermal situations tend to have relatively long time constants and so can accept a somewhat lower cycling rate than some other situations, oven vendors like to keep the valve's cycle time to a few minutes. So it is common to see swings of ±50°F (about ±25°C) around the setpoint. That's just not good enough for many recipes or foods.   The problems of the standard oven design and subsequent performance go well beyond the basic temperature-control loop. Only after I read the article did I realize how much I hadn't known about the implications of oven design and the cooking process.   For example, the heat source doesn't cool the food directly; instead, it heats the oven walls (metal or brick), which then re-radiate the heat to the food. The article explained how this leads to very uneven heat zones in the oven for conventional metal-wall ovens, and numerous other problems. (The article did highlight some ways to improve an oven's performance, which was nice to see.)   These other problem are not trivial, either, as they have to do with how different foods are affected by the radiant heat of the oven walls versus direct heating; not surprisingly, it does make a difference, just as different types of electrical load impedances affect power supply performance. (There are times when I felt the whole discussion of oven idiosyncrasies was akin to the mysteries of black-body radiation , which were finally resolved with the development of quantum theory in the 1920s by Max Planck and others.)   The irony of the oven problem is that for most engineers, the primary concerns regarding heat are twofold: minimizing generation of it in the first place, through the use of low-power circuits and high-efficiency supplies; and maximizing its removal, using convection, conduction, and radiation via use of fans, heat sinks, cold plates, and even advanced active techniques. It takes a 90° or even 180° shift in thinking to begin to understand heat and its effective delivery as an outcome rather than as an obstacle.   Have you ever been faced with a design challenge that required that you reroute your established way of looking at something, to a very different course? Did you realize the differences right away, or did you have to really step back to understand the many inherent assumptions you were making and how they had to be changed as well?

Jim Smith, president of Electronics Manufacturing Sciences Inc., sent me an email in response to my previous post on engineers who cannot solder. He said the people at his company focus on soldering, because it is both the heart of electronics manufacturing and the process that causes the most problems. For these reasons, his company specialises in soldering training/education, certification, and process development. I reproduce Jim's email in its entirety as follows: Hello Max. I came upon your column by fortuitous accident. You wrote with great insight into a critical problem in the US electronics industry. And, as I wrote hurriedly in my posted response, the problem is much greater than most people realise. Almost no one—not just engineers—knows how to solder. Most of those who take pride in their soldering skills are unaware that their technique is faulty. I've been developing soldering processes, troubleshooting soldering process problems, and teaching all forms of soldering for close to 50 years. I've worked with hundreds of companies—from start-ups to the biggest corporations—all over the world, and I can tell you without fear of contradiction that lack of soldering knowledge is more prevalent now than ever. The prevailing belief holds that a cosmetically attractive solder connection must be a good solder connection. But this is not necessarily true, especially if a soldering iron was involved in making that connection. At soldering iron temperatures, solder will stick to oxides and give the false appearance of a proper connection (you can read more in my document The Metallurgy of Heat-Induced Soldering ). The compromised integrity of the solder bond itself affects reliability under conditions of vibration or thermal cycling, but the far more serious reliability consequence consists of degradation inside components, especially ICs. Applying such high temperature long enough to achieve adherence of solder to the oxidized surface causes the wire bonds inside ICs to degrade to an extent that would require decades of product use. The phenomenon is known as the "purple plague" ( click here to see some discussions on this phenomenon). The damage is invisible, so everyone blames the component manufacturer when devices fail prematurely, but the root cause is improper soldering technique. Today's hand soldering procedure was developed in the vacuum tube era for attaching wires to sockets. Those materials could not be damaged by overheating. Rather, the challenge was how to get enough heat into big metal objects using irons that were not very efficient at turning electricity into heat. The focus was entirely on keeping the parts hot enough, long enough so the solder would flow adequately without freezing. When solid-state devices entered the picture, we began soldering the components themselves rather than their sockets. To prevent heat damage to the component, metal clips known as heat sinks were placed on leads next to the component body. This allowed the excess soldering heat to flow into the sink rather than stressing the component. As parts got smaller, however, there wasn't room for heat sinks, so they disappeared from the work instructions, but every trainee continued to be told to use the same technique developed for wiring vacuum tube assemblies. This is insane. Complicating matters is the fact that electronics "soldering" has mostly been welding. Surfaces to be soldered have mostly been tin or tin/lead and those surfaces melted at or below 450°F (232.22°C), which is much lower than the temperature reached by those surfaces during "soldering." When a surface melts during application of solder, the heavy liquid solder easily pushes oxides and even contaminants aside; the liquid metals (solder and component surface) can then flow together. It's hard to imagine a less challenging application. But the lead-free movement and fear of tin whiskers have caused the use of new component surface metals that have much higher melting temperatures. Those surfaces don't melt during soldering, which means that the industry—for the first time in its history—must actually solder. But they (including most of the people who set industry standards) don't know how to solder; they only know how to weld. In short, no one understands wetting forces and solderability. Nor do they have meaningful understanding of flux properties ranging from ionic contamination (acid residue) hazards to hygroscopic solids. Soldering is the heart of electronics manufacturing, and lack of process knowledge is killing industry. Touchup (most of it unrecognised by management; what we used to call "the hidden factory") and rework are rampant. Engineers get no coherent education about soldering. I don't know of any course aside from my company's Science of Soldering that teaches the chemistry, metallurgy, and physics of soldering. Operators and technicians get "certified" in the idiotic ritual of memorising A-610 or J-STD-001 acceptance rules so they can answer open-book multiple-choice questions. The training tells them the appearance of acceptable solder connections but provides no knowledge at all of how to meet those requirements (aside from pushing the solder around with the iron until it finally achieves a shape everyone can live with). Rather than rewarding operators who bring material and process problems to management's attention so the problems can be corrected, the industry places highest value on assemblers who can produce visually acceptable connections with un-solderable materials. The whole system is like teaching pilots how to fly simply by giving them route maps without any instructions about how to operate the plane itself. (My 2011 Assembly Magazine column about the damage done by A-610 and J-STD-001 training can be found by clicking here .) To sum up, you've opened the journalistic door to the single most important challenge facing electronics manufacturing today. Thanks for your time and the outstanding column. Best wishes, Jim. I don't know about you, but I think he is passionate about soldering. Also, since he says such nice things about me and my writings, I think it's fair to assume that he is a very clever and discerning person. However, I fear he has thrown down the gauntlet to some parties with an interest in certain industry standards. I await everyone's comments and feedback with bated breath.  

As a follow up to my previous post on engineers who cannot solder, I received an email from Jim Smith, president of Electronics Manufacturing Sciences Inc. He said the people at his company focus on soldering, because it is both the heart of electronics manufacturing and the process that causes the most problems. For these reasons, his company specialises in soldering training/education, certification, and process development. I reproduce Jim's email in its entirety as follows: Hello Max. I came upon your column by fortuitous accident. You wrote with great insight into a critical problem in the US electronics industry. And, as I wrote hurriedly in my posted response, the problem is much greater than most people realise. Almost no one—not just engineers—knows how to solder. Most of those who take pride in their soldering skills are unaware that their technique is faulty. I've been developing soldering processes, troubleshooting soldering process problems, and teaching all forms of soldering for close to 50 years. I've worked with hundreds of companies—from start-ups to the biggest corporations—all over the world, and I can tell you without fear of contradiction that lack of soldering knowledge is more prevalent now than ever. The prevailing belief holds that a cosmetically attractive solder connection must be a good solder connection. But this is not necessarily true, especially if a soldering iron was involved in making that connection. At soldering iron temperatures, solder will stick to oxides and give the false appearance of a proper connection (you can read more in my document The Metallurgy of Heat-Induced Soldering ). The compromised integrity of the solder bond itself affects reliability under conditions of vibration or thermal cycling, but the far more serious reliability consequence consists of degradation inside components, especially ICs. Applying such high temperature long enough to achieve adherence of solder to the oxidized surface causes the wire bonds inside ICs to degrade to an extent that would require decades of product use. The phenomenon is known as the "purple plague" ( click here to see some discussions on this phenomenon). The damage is invisible, so everyone blames the component manufacturer when devices fail prematurely, but the root cause is improper soldering technique. Today's hand soldering procedure was developed in the vacuum tube era for attaching wires to sockets. Those materials could not be damaged by overheating. Rather, the challenge was how to get enough heat into big metal objects using irons that were not very efficient at turning electricity into heat. The focus was entirely on keeping the parts hot enough, long enough so the solder would flow adequately without freezing. When solid-state devices entered the picture, we began soldering the components themselves rather than their sockets. To prevent heat damage to the component, metal clips known as heat sinks were placed on leads next to the component body. This allowed the excess soldering heat to flow into the sink rather than stressing the component. As parts got smaller, however, there wasn't room for heat sinks, so they disappeared from the work instructions, but every trainee continued to be told to use the same technique developed for wiring vacuum tube assemblies. This is insane. Complicating matters is the fact that electronics "soldering" has mostly been welding. Surfaces to be soldered have mostly been tin or tin/lead and those surfaces melted at or below 450°F (232.22°C), which is much lower than the temperature reached by those surfaces during "soldering." When a surface melts during application of solder, the heavy liquid solder easily pushes oxides and even contaminants aside; the liquid metals (solder and component surface) can then flow together. It's hard to imagine a less challenging application. But the lead-free movement and fear of tin whiskers have caused the use of new component surface metals that have much higher melting temperatures. Those surfaces don't melt during soldering, which means that the industry—for the first time in its history—must actually solder. But they (including most of the people who set industry standards) don't know how to solder; they only know how to weld. In short, no one understands wetting forces and solderability. Nor do they have meaningful understanding of flux properties ranging from ionic contamination (acid residue) hazards to hygroscopic solids. Soldering is the heart of electronics manufacturing, and lack of process knowledge is killing industry. Touchup (most of it unrecognised by management; what we used to call "the hidden factory") and rework are rampant. Engineers get no coherent education about soldering. I don't know of any course aside from my company's Science of Soldering that teaches the chemistry, metallurgy, and physics of soldering. Operators and technicians get "certified" in the idiotic ritual of memorising A-610 or J-STD-001 acceptance rules so they can answer open-book multiple-choice questions. The training tells them the appearance of acceptable solder connections but provides no knowledge at all of how to meet those requirements (aside from pushing the solder around with the iron until it finally achieves a shape everyone can live with). Rather than rewarding operators who bring material and process problems to management's attention so the problems can be corrected, the industry places highest value on assemblers who can produce visually acceptable connections with un-solderable materials. The whole system is like teaching pilots how to fly simply by giving them route maps without any instructions about how to operate the plane itself. (My 2011 Assembly Magazine column about the damage done by A-610 and J-STD-001 training can be found by clicking here .) To sum up, you've opened the journalistic door to the single most important challenge facing electronics manufacturing today. Thanks for your time and the outstanding column. Best wishes, Jim. I don't know about you, but I think he is passionate about soldering. Also, since he says such nice things about me and my writings, I think it's fair to assume that he is a very clever and discerning person. However, I fear he has thrown down the gauntlet to some parties with an interest in certain industry standards. I await everyone's comments and feedback with bated breath.