Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Average Circuit Efficiency

  1. Aug 5, 2014 #1
    Hi all, I've been looking for an answer to this for a while now but I can't explain to Google what I want to know. Hope I have better luck here.

    I'm an aerospace engineer working on a satellite project. In particular I'm currently working on the thermal analysis and I need to model up each of the systems as a heat source.

    My question: If I don't have the exact efficiency of each of the electrical systems, what is a good value to use? Obviously the input power can't all be converted to heat, but some is. For example, say the main onboard computer uses 1W of power, would it be a valid assumption to model it as a 0.6W heat source?

    I really have little to no idea and searching Google I can only find exact methods for each component on the board, or how to remove heat, or conduction through the plane. Not useful.

    Cheers!
     
  2. jcsd
  3. Aug 5, 2014 #2

    AlephZero

    User Avatar
    Science Advisor
    Homework Helper

    If you can't identify something else that it is converted into, then it is all converted into heat.

    "Something else" could be mechanical work done by a motor, or the electrical output from a power amplifier (the output power will be converted into heat in the load circuit, but that could be physically somewhere else in the satellite), etc.

    For a computer circuit board, assuming 100% conversion to heat seems reasonable to me.
     
  4. Aug 5, 2014 #3

    berkeman

    User Avatar

    Staff: Mentor

    Welcome to the PF.

    All of the input power goes into heat. Unless you have LED lights or something that emits radiation away from the electronics package, which is pretty unlikely on a satellite package, I would guess. There is no place for the power to go except into heat.
     
  5. Aug 5, 2014 #4

    berkeman

    User Avatar

    Staff: Mentor

    Oh, and you should also model the inefficiency of the power supply that converts the battery power (and/or solar panel power) to the power that is consumed by the modules. That inefficiency contributes to the thermal load on the satellite.
     
  6. Aug 5, 2014 #5

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award

    Hmmmm

    draw an imaginary surface enclosing it.

    As previous posters pointed out,


    Any electrical energy entering it must come back out as either electricity, light, mechanical motion, or heat. (Can't be sound in space i suppose.)

    So if a 1 watt computer doesn't transmit a radio signal or move, power, or illuminate something else, i'd assign it one watt of heat.
     
  7. Aug 5, 2014 #6

    berkeman

    User Avatar

    Staff: Mentor

    Oh, good point Jim. I spaced RF transmissions. That will definitely subtract from the heat output of the comms package. Good catch! :smile:
     
  8. Aug 5, 2014 #7
    I have already taken the 85% efficiency of the EPS into account.

    I was originally using the assumption that the power in has to go somewhere so it must become heat, but that means that the processing of information and anything else the system, for example the OBC, does uses 0 power. There are systems that do use power to generate mechanical energy, but some have no lights, no moving parts.

    If I was to measure the power going into my computer's motherboard while disconnected from everything else. I can't imagine the heat load would be equal to the input power.
     
  9. Aug 5, 2014 #8

    russ_watters

    User Avatar

    Staff: Mentor

    It isn't that it uses zero power, it is just that the power doesn't get stored. If a computer chip flips a bit from 0 to 1 and back to 0, the resulting state is equal to the initial state*, so conservation of energy demands that there has been no storage of energy. It's just like if you drive your car somewhere and then drive back home: your initial and final location are the same, so there has been no change in energy of the car, so all of the energy of combustion must have been dissipated as heat.

    *And I doubt there is much difference in energy content of a 1 vs a 0 bit either.
     
  10. Aug 5, 2014 #9

    berkeman

    User Avatar

    Staff: Mentor

    It does. Think about the processing in the microprocessor. That consists of currents flowing in switching circuits, and voltages changing on gate capacitances. The currents flowing generate voltage drops in the connecting traces, and the voltages generate leakage currents across transistors and capacitors. It all goes into heat unless there is some other mechanism to carry the energy away (as mentioned, RF transmissions, light, mechanical movement, etc.).

    Where else do you think the energy can go? :smile:
     
  11. Aug 5, 2014 #10

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award

    And I can't imagine it would be any less !

    The only two things a computer can produce are heat and frustration.


    i'm with you there

    It's just a calculating engine. Engines reject heat equal to input power minus mechanical or thermal work delivered.
    I can't ascribe mechanical or thermal work to calculating itself, that's an imaginary process .
    Calculating engines mechanically move something inside even if it's just charge, and that's where your electricity is turned into heat : I^2R in the electronic parts as they move charge around to flip bits.

    Thought experiment:
    Place your whole computer system in a perfectly insulated calorimeter and run it overnight.
    How much heat will you collect?
     
  12. Aug 5, 2014 #11
    OK, thanks all. I think I'll go back to my original assumption and have the heat load being equal to the input load for systems without obvious outputs of energy.

    My supervisor asked me to look into it, and I have. Not much else I can do. Lol

    Cheers!
     
  13. Aug 5, 2014 #12

    berkeman

    User Avatar

    Staff: Mentor

    If your supervisor has any doubts after you explain it to them, you know what link to email them... :biggrin:
     
  14. Aug 6, 2014 #13

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award



    Okay we've got consensus i think on the physics.

    As to your situation, a long time ago I heard a good saying: "One experiment is worth a thousand expert opinions."

    Now: from where do your "input load" numbers come?

    Manufacturers often specify power as what the device's package is capable of dissipating, not what it typically is dissipating. Now that's not a question of efficiency but of rating. I've seen computer power systems turn out oversized severalfold because of "published ratings".

    If you can measure the actual input power (DC volts X amps) while it's operating you'll have a sanity check.

    That'd be a really good "look in to it". :smile:
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Average Circuit Efficiency
  1. Rectifier efficiency (Replies: 3)

  2. Generator efficiency (Replies: 1)

Loading...