Electronics and Operating Temperatures

Discussion in 'Hardware' started by Steve S, Jan 23, 2012.

  1. Steve S

    Steve S Scribbler - Standard Member Super Moderator

    Messages:
    3,861
    Likes Received:
    272
    Trophy Points:
    101
    It's not unusual to see a lot of posts in the forums from members fretting about the operating temperatures of their tablets. A major worry is that higher temperatures must surely be degrading the expected operating lives of their shiny new device or will otherwise cause hardware problems. (Members are also concerned that some tablets run so hot that they are uncomfortable to hold but as you are about to see, I'm not going to address that.) Of course, very high temperatures can cause problems, but the "acceptable" operating range for commercial and industrial electronics is much wider than one might suspect.

    The following quote appeared today (23Jan12) in the online edition of Electronic Design Magazine, and I thought it was worth repeating because it comes from Bob Pease, a long time technical writer with a vast amount of experience with (primarily analog) electronic design and hardware. For many of us, Bob was a go-to expert on real world electronic design; he used theory, but he never let it get in the way of real world effects and behavior.

    In this extract, Bob is talking about analog circuit performance, but the temperatures are typical of digital electronic components, too. Take particular note of the temperatures mentioned in the second paragraph:

    "...in general, modern electronic components aren’t greatly improved by cooling or the absence of heating. In fact, those of us who remember the old vacuum-tube days remember that a good scope or voltmeter usually worked better if you help it run nice and warm all the time, because all of the resistors and components stayed dry and never got moist under humid conditions. I won’t say that the electrolytic capacitors might not have liked being a little cooler. But the mindless effort to improve the reliability by keeping components as cool as possible has been overdone. I’m sure you can blame much of that foolishness on MIL-HBDK-217 and all its versions...but in the industrial and instrument business, we don’t really have to follow its every silly quirk and whim.

    One guy who argues strenuously about –217 is Charles Leonard of Boeing, and you may well enjoy his writing (Leonard Charles, “Is reliability prediction methodology for the birds?” PowerConversion and Intelligent Motion, November 1988, p. 4). So if something is drifting a little and you think you can make a big improvement by adding a fan and knocking its temperature down from +75 to +55°C (167 to 131 degrees F), I caution you that you’ll probably be disappointed because there isn’t a lot of improvement to be had. It’s conceivable that if you have a bad thermal pattern causing lots of gradients and convection, you can cut down that kind of thermal problem. In general, though, there’s not much to be gained unless parts are getting up near their maximum rated temperature or above +100°C (212 degrees F). Even plastic parts can be pretty reliable at +100°C. I know the ones I’m familiar with are.

    Robert A. Pease / Engineer...
    "

    The bottom line here is that typical electronics are perfectly comfortable operating at temperatures well above those that we human beings would consider comfortable. By the same token, tablets shouldn't get so hot that they burn you and that's a legitimate concern, but that's a design constraint imposed by us, not by the temperature tolerance of the electronics.

    I thought that this quote was helpful in putting some of this into perspective...
     
  2. thatcomicsguy

    thatcomicsguy Scribbler - Standard Member Senior Member

    Messages:
    1,013
    Likes Received:
    202
    Trophy Points:
    81
    That's a nice knowledge tid-bit, Steve S. Thanks!

    It's sometimes hard to find specific answers to specific questions on the web, and instead get nothing but miles of generalities. I was trying to find out once, "Why do high temperatures cause chips to stop working?"

    Everybody on the web agreed that high temps cause problems, (hence the ten thousand different cooling systems and solutions available), but it was very hard to find out exactly why heat was a problem.

    I found this discussion somewhat informative:

    microchip - Why exactly do chips start malfunctioning once they overheat? - Electrical Engineering - Stack Exchange

    One aspect is that the atomic properties which make silicon a semiconductor change at around 150 C (which is far hotter than any tablet will ever run.)

    But it sounds as though simple heat expansion and contraction is the cause of many problems. Very simply, if your chip expands more or less than the circuit board it's attached to, then micro connections can break. If the silicon chip inside its clay housing expands differently from the housing itself, then it might form tension cracks and die. Of course, I'm sure the engineers have worked hard for decades to perfect the right expansion/contraction ratios with all their materials, which is why it is important to respect the heat limits on the data sheets for the various components in a computer.

    In the case of my heat-sensitive Tecra M4 tablet, it seems that at temperatures which are supposedly within operating limits, certain of the chips pop their solder joints and fail. One of the DIY solutions for this model is to actually take the mother board, chips and all, out of the computer and put it in a kitchen oven and roast it up to the melting point of solder. The soldier joints liquidize and reform, and when cooled. . , well several people have reported that their computers start working again and the problems go away.

    I've not needed to try this myself; I've taken measures to assure that the critical components don't reach those breaking point temps, but my computer does run hot. -My CPU is currently sitting at 71C. (Which means the heat sink could probably use some dusting. Hm. I'll do that today). But the CPU isn't the problem in the Tecra M4. It's the GPU, which has a history of dying at high temps.
     
  3. purplepeopledesign

    purplepeopledesign Scribbler - Standard Member

    Messages:
    359
    Likes Received:
    0
    Trophy Points:
    31
    It's not just the circuit board expansion that can be problematic at higher temperatures, but other non-electronic components are susceptible. I've seen more than enough deformed thermoplastics and delaminated adhesives to say that I always try to stay well within recommended operating temperatures. This is probably the reason why many portables have metal chassis.

    :)ensen.

    PS: For those that don't know, the quote is from the late great Bob Pease.
     
  4. Steve S

    Steve S Scribbler - Standard Member Super Moderator

    Messages:
    3,861
    Likes Received:
    272
    Trophy Points:
    101
    <<...I was trying to find out once, "Why do high temperatures cause chips to stop working?"...>>

    ...And it's not just differential thermal expansion and delamination.

    I once saw a very instructional time-lapse micro-video of a metallic trace on an IC under a typical load current. At the microscopic scale, the metal trace could be seen to be a series of metal grains, looking much like a flagstone walkway. Under current and at operating temperature, the grain boundaries moved, sort of like watching sticks float along the surface of a slow-moving stream. At one point, a series of grain boundaries started to "pile up" together; that eventually led to even more moving grain boundaries getting snared, until an actual gap started to form across the metal trace. The gap continued to snare more moving boundaries and the gap finally grew to the point that the trace failed in an open circuit condition!

    The sophistication required to build reliable micro-electronics is substantial and shouldn't be underestimated...!!!
     
  5. Steve S

    Steve S Scribbler - Standard Member Super Moderator

    Messages:
    3,861
    Likes Received:
    272
    Trophy Points:
    101
  6. thatcomicsguy

    thatcomicsguy Scribbler - Standard Member Senior Member

    Messages:
    1,013
    Likes Received:
    202
    Trophy Points:
    81
    That was one of the better articles I've now read on the subject. It was a little light on the I.C. section, but then it's fairly easy to find out the maximum operating temperature of your CPU. Intel includes it in their user manuals, and there are collected lists available.

    Here's one. . .

    Maximum CPU Temperature | Hardware Secrets

    Apparently the Pentium M in my Tecra M4 has a maximum operating temp of 100 C. The highest I've seen my machine go is around 91 C, which seemed pretty alarming at the time, but it was still below the limit. Removing dust build-up fixed this, and I now seem to top out around 75 C on average.

    I'm particularly interested in this at the moment because just last week I ordered a faster CPU on eBay, (along with some Arctic Silver compound).

    The Tecra M4 runs on the Pentium M processor, which is a user-replaceable part on that motherboard. So my Tablet will shortly be going from 1.73 Ghz to 2.13 Ghz if all goes well.

    I wonder what this will do to the operating temperature. I'm hoping my fan, which is locked at a permanent slow speed will not have to be adjusted upwards, as I like how quiet it is right now. But I can't imagine it would have to get *that* much faster in order to keep pace. I might have to remove one of the 10 ohm resistors on its power line. (Right now I've got 40 ohms on a 5 volt power line).

    Anyway, in the Toshiba section, I'll report on any differences an extra 400 million operations per second has on my compy. I don't know that I'll see any improvement, but who knows. For $15 on eBay it's a pretty affordable experiment. :)
     
  7. Hattori Hanzo

    Hattori Hanzo Scribbler - Standard Member

    Messages:
    614
    Likes Received:
    1
    Trophy Points:
    31
    User replaceable? I'd rate this service technician replaceable :D

    By paper you should not see a difference, as, iirc, both have the same TDP. Practically I'm quite sure that the faster one will emit more heat. That's my experience with Intel CPU's of a specific generation since the Pentium M era: Same TDP on paper, but the slower CPUs never reach that TDP.

    I guess you'll either have to give undervolting a shot or remove one of those 10 Ohm resistors. Iirc you're using that thing stationary, don't you?
    In that case you can just go and find an appropriate voltage for the max multiplier and throw speedstep aboard.
     

Share This Page