4 ms·

Power engineers have been doing this for decades (a century?); estimating the temperature of a transformer or motor from a simple ODE to see if the relays need

Power engineers have been doing this for decades (a century?); estimating the temperature of a transformer or motor from a simple ODE to see if the relays need to trip. The parameters are derived from the electrical parameters of the transformer. Aside from applying it to speakers, whats novel?

kkielhofner 4y ago
I'm not arguing the merits of this patent (see my other comments). I'm attempting to provide some background on how this messed up system "functions". That said, this process is not "estimating" anything. It's taking high resolution direct measurements[0] of driver temperature with awareness of the frequency and amplitude of the audio signal (also factoring in crossover frequency) and doing a bunch of stuff (this is well outside my area of expertise) to essentially overpower the individual driver(s) well past their intended electrical and physical specifications. If I ever ended up on a jury for this (I wouldn't - I would get dismissed after one question in voir dire) even I would have a tough time associating (and recognizing as prior art) what you described with this process. [0] - I've since been corrected on this - it is an estimation (of sorts).
- zamadatix 4y ago
  The "bunch of stuff" is precisely the part where estimation comes in. There is no direct temperature sensor to measure from, the only thing available is the voltage levels on the speaker. As the patent says it's a 2 level estimate: estimate of impedance from voltages -> estimate of temperature from modelling the speaker's physical parametrs as a function of the driving signal and impedance estimate. With patents it's not the general idea that's being patented. The patent itself cites many earlier patents which operate around the same general idea just with slightly different implementations on how the control process is fed and how it reacts.
  - kkielhofner 4y ago
    Appreciate the clarification - like I said this is well outside my area of expertise and my eyes glazed over pretty quickly in reading but at this point we're borderline in semantics. Exactly. In this case it's a "process". Generally speaking my take is there's almost never entirely new process (or design or anything else) in any "invention" - humans have been at this for hundreds of years and there's almost nothing that's entirely new. I don't think I've ever seen a filing without tons of references.
    - astrange 4y ago
      It's well outside your area of expertise which is why you started with a "well actually" post claiming the article is obviously false?
      - sbaiddn 4y ago
        Common, his post generated an interesting discussion (and netted me 10 points).
    - sbaiddn 4y ago
      "there's almost never entirely new process" I agree, but in this case the only innovation I see is applying a known solution to protect large inductor coils (transformers, motors) to protect smaller, inductor coils (speaker drivers). Apple can patent their actual model all they want (for example by adding the driven air as a cooling term), but the idea, generally, of building a model to estimate the temperature is prior art.
    - zamadatix 4y ago
      I suppose the non-semantics point of everyone in this chain is: there is nothing about this patent from Apple that explains why using a dynamic temperature/power limit model for the speakers is uncommon in laptops. They are neither the first nor last to get a patent to do this, why then should it be the "real" reason most laptops don't?
      - freejazz 4y ago
  - icegreentea2 4y ago
    Yes. In this case it's maybe a little face palmy how fill in the blanks are... but heh. It cites a Bang & Olufsen patent (https://patents.google.com/patent/US20090257599A1/en https://patents.google.com/patent/US20090257599A1/en) which implements voice-coil protection via measuring the resistance/impedance of the speaker while it's being driven. It also cites another patent (https://patents.google.com/patent/US20120020488A1/en https://patents.google.com/patent/US20120020488A1/en) which implements control of a speaker via combination of impedance measurements of the speaker while it's being driven + a baseline ("binding") measurement of impedance + ambient temperature that is taken at device power up - ie generating a per device calibration curve. From quickly staring at these, the Apple patents threads the needle by: * Implementing speaker control using the output of a temperature model (which is fed by the resistance measurement), instead of "directly" operating on speaker resistance. * Not implementing per-speaker calibration. (Note! There are other cited patents that I didn't look at, this wasn't meant to be exhaustive, just illustrative... and just to generally satisfy an itch)
analog31 4y ago
Speaker engineers have been doing this for decades too. Not necessarily by the specific method chosen by Apple (didn't read the patent), or with laptop speakers, but dynamic control of signals to achieve greater loudness within a specific power rating has been a thing for a long time AFAIK. Here's an example: https://testing.eminence.com/d-fend/ https://testing.eminence.com/d-fend/
- sbaiddn 4y ago
  Cool. The HiFi community has done such neat things. Never underestimate a smart community driven by love (and dash psychoacoustic delusion ;) )

kkielhofner 4y ago

I'm not arguing the merits of this patent (see my other comments). I'm attempting to provide some background on how this messed up system "functions". That said, this process is not "estimating" anything. It's taking high resolution direct measurements[0] of driver temperature with awareness of the frequency and amplitude of the audio signal (also factoring in crossover frequency) and doing a bunch of stuff (this is well outside my area of expertise) to essentially overpower the individual driver(s) well past their intended electrical and physical specifications. If I ever ended up on a jury for this (I wouldn't - I would get dismissed after one question in voir dire) even I would have a tough time associating (and recognizing as prior art) what you described with this process. [0] - I've since been corrected on this - it is an estimation (of sorts).

zamadatix 4y ago
The "bunch of stuff" is precisely the part where estimation comes in. There is no direct temperature sensor to measure from, the only thing available is the voltage levels on the speaker. As the patent says it's a 2 level estimate: estimate of impedance from voltages -> estimate of temperature from modelling the speaker's physical parametrs as a function of the driving signal and impedance estimate. With patents it's not the general idea that's being patented. The patent itself cites many earlier patents which operate around the same general idea just with slightly different implementations on how the control process is fed and how it reacts.
- kkielhofner 4y ago
  Appreciate the clarification - like I said this is well outside my area of expertise and my eyes glazed over pretty quickly in reading but at this point we're borderline in semantics. Exactly. In this case it's a "process". Generally speaking my take is there's almost never entirely new process (or design or anything else) in any "invention" - humans have been at this for hundreds of years and there's almost nothing that's entirely new. I don't think I've ever seen a filing without tons of references.
  - astrange 4y ago
    It's well outside your area of expertise which is why you started with a "well actually" post claiming the article is obviously false?
    - sbaiddn 4y ago
      Common, his post generated an interesting discussion (and netted me 10 points).
  - sbaiddn 4y ago
    "there's almost never entirely new process" I agree, but in this case the only innovation I see is applying a known solution to protect large inductor coils (transformers, motors) to protect smaller, inductor coils (speaker drivers). Apple can patent their actual model all they want (for example by adding the driven air as a cooling term), but the idea, generally, of building a model to estimate the temperature is prior art.
  - zamadatix 4y ago
    I suppose the non-semantics point of everyone in this chain is: there is nothing about this patent from Apple that explains why using a dynamic temperature/power limit model for the speakers is uncommon in laptops. They are neither the first nor last to get a patent to do this, why then should it be the "real" reason most laptops don't?
    - freejazz 4y ago
- icegreentea2 4y ago
  Yes. In this case it's maybe a little face palmy how fill in the blanks are... but heh. It cites a Bang & Olufsen patent (https://patents.google.com/patent/US20090257599A1/en https://patents.google.com/patent/US20090257599A1/en) which implements voice-coil protection via measuring the resistance/impedance of the speaker while it's being driven. It also cites another patent (https://patents.google.com/patent/US20120020488A1/en https://patents.google.com/patent/US20120020488A1/en) which implements control of a speaker via combination of impedance measurements of the speaker while it's being driven + a baseline ("binding") measurement of impedance + ambient temperature that is taken at device power up - ie generating a per device calibration curve. From quickly staring at these, the Apple patents threads the needle by: * Implementing speaker control using the output of a temperature model (which is fed by the resistance measurement), instead of "directly" operating on speaker resistance. * Not implementing per-speaker calibration. (Note! There are other cited patents that I didn't look at, this wasn't meant to be exhaustive, just illustrative... and just to generally satisfy an itch)

analog31 4y ago

Speaker engineers have been doing this for decades too. Not necessarily by the specific method chosen by Apple (didn't read the patent), or with laptop speakers, but dynamic control of signals to achieve greater loudness within a specific power rating has been a thing for a long time AFAIK. Here's an example: https://testing.eminence.com/d-fend/ https://testing.eminence.com/d-fend/

sbaiddn 4y ago
Cool. The HiFi community has done such neat things. Never underestimate a smart community driven by love (and dash psychoacoustic delusion ;) )