1. For prospective new members, a word of warning: don't use common names like Dennis, Simon, or Kenny if you decide to create an account. Spammers have used them all before you and gotten those names flagged in the anti-spam databases. Your account registration will be rejected because of it.
    Dismiss Notice
  2. Since it has happened MULTIPLE times now, I want to be very clear about this. You do not get to abandon an account and create a new one. You do not get to pass an account to someone else and create a new one. If you do so anyway, you will be banned for creating sockpuppets.
    Dismiss Notice
  3. For a trial period of one week, allowed avatar size has been increased to 256kb, up from 51.2kb. During this time, if the forum seems to be loading slow, drop by the topic in the Announcements & Rules sub to let me know.
    Dismiss Notice
  4. QQ will be down for upgrades starting at 2030 EDT. Please be prepared and save any work.
    Dismiss Notice
  5. If you wish to change your username, please ask via conversation to tehelgee instead of asking via my profile. I'd like to not clutter it up with such requests.
    Dismiss Notice
  6. To the many new members following Mr Zoat to our forum, I would suggest a quick read of our Rules. Don't worry, it's not a long post. Pay special attention to Rule 8. We are deadly serious about it.
    Dismiss Notice
  7. Due to the actions of particularly persistent spammers and trolls, we will be banning disposable email addresses from today onward.
    Dismiss Notice
  8. To make things official, I have added a proper description of the method to appeal a moderator action in the Rules thread.
    Dismiss Notice

Electrical Engineering Trivia Thread

Discussion in 'General' started by Prince Charon, Apr 9, 2019.

  1. Prince Charon

    Prince Charon Just zis guy, you know?

    Joined:
    Feb 20, 2014
    Messages:
    4,422
    Likes Received:
    12,560
    There was an odd derail on the Sanctimonious Assholery on QQ thread, relating to electrical engineering. Clearly, this thread is needed to prevent such derails.
     
    NuclearBird likes this.
  2. vyor

    vyor Oh that's cute

    Joined:
    Jul 29, 2015
    Messages:
    10,135
    Likes Received:
    25,468
    I ARRIVE
     
    Prince Charon likes this.
  3. NuclearBird

    NuclearBird Herald of Emissions

    Joined:
    Jun 6, 2018
    Messages:
    1,189
    Likes Received:
    7,790
    Let me start this off in the wake of a disastrous Energistics midterm test:

    Never forget the difference between phase voltage and line voltage.
     
    Salbazier, Ohmrond, DkArthas and 2 others like this.
  4. vyor

    vyor Oh that's cute

    Joined:
    Jul 29, 2015
    Messages:
    10,135
    Likes Received:
    25,468
    Ouch.

    For that matter, when looking at VRMs and transformers, do not assume wattage in == wattage out, you will lose energy to heat and if you don't, you're probably doing something wrong(or you're magic).
     
    Prince Charon likes this.
  5. NuclearBird

    NuclearBird Herald of Emissions

    Joined:
    Jun 6, 2018
    Messages:
    1,189
    Likes Received:
    7,790
    Don't forget the Z=R+jX impedance and the wnding layout, including grounding of different types.

    Usually the resistance is negligible at nominal conditions, leaving jX as a reactance on the line.
    Winding type (Star-Star, Delta-Delta. Star-Delta) determines behaviour when the three-phase system isn't fully symmetrical (usually due to a short circuit). There is a mtehod to solve this issue, especially when transformers are involved, but it would take too long to explain.
    Expanding on the previous point: grounded star windings are great when dealing with high voltage (120kV) systems.
    Get used to dividing and multiplying by sqrt(3).
    There is capacitance on everything, unless it's declared negligible.

    Now for something outside of Energistics, yet extremely important for Electrical Engineering as a whole:
    You solve for

    z = exp(sT)

    where T is the Sampling Period. Reducing it to a logarithm might sound appealing, but some analysis shows that once you start testing for equivalencies in stability criterions you find that things diverge wildly. Instead, when you take

    z=exp(sT/2)/exp(-sT/2) ~ (1+sT/2)/(1-sT/2)

    and "approximate" the numerator and denominator in the above way, then test each side for how stability on one side shows on the other.
    You find that no matter which side you start on, once you start with a stable s=a+jß on the continuous side where a<0 (Real component is negative) or a z=r*exp(j*Th) where r<1 (magnitide less than one) you get a result on the other side of the equation that indicates stability and vice versa.

    Testing for z when s->0+ (s approaches 0 from a higher value) you get a convergence to +1 from a higher value. This is for cases where Re{s}>0
    Testing for z when s->0- (s approaches 0 from a lower value) you get a convergence to +1 from a lower value. This is for cases where Re{s}<0

    Why is this important? You see, testing a system for stability in discrete and continuous complex frequency domain is easy: Determine the poles of the system, the roots of the denominator polynomial.

    For continuous time systems, f(t) -> F(s), then test for every single pole (root). If real component Re{pi}<0 for every i=1...n where n is the order of the polynomial, then the system is surely stable.

    For discrete time systems, f[k] -> F(z), then test for every single pole (root). If the magnitude of pi, |pi|<1 for every i=1...n where n is the order of the polynomial, then the system is surely stable.
    Despite the fractional polynomial being a mere approximation the equivalence holds perfectly outside of (1,0) on the complex plane, despite having used a fairly crude approximation of an exponential function.
     
    vyor likes this.
  6. molemole

    molemole Versed in the lewd.

    Joined:
    May 18, 2017
    Messages:
    1,116
    Likes Received:
    7,155
    For those wanting to do DSP, one word: Don't. Outside defense, there are rather few jobs in pure DSP, and most image processing and speech is run by deep learning guys with CS degrees.

    If you do do DSP, then by God read Shannon's theorem and the original derivation. If you know that well, and you know linear algebra decently well then 80% of DSP is pretty much just reading and minor understanding.
    All stuff I wish I knew before leaving with a EE undergrad, hopefully someone can make use.
     
  7. vyor

    vyor Oh that's cute

    Joined:
    Jul 29, 2015
    Messages:
    10,135
    Likes Received:
    25,468
    The only real modern jobs in DPS are the designers for the damn things, and that ain't a big field.
     
  8. molemole

    molemole Versed in the lewd.

    Joined:
    May 18, 2017
    Messages:
    1,116
    Likes Received:
    7,155
    More hardware than anything else these days. The issue is that outside of a few firms like Microsoft and MERL, most Silicon Valley companies don't believe in using traditional signal processing for data conditioning. And most DSP for telecoms has been commodified, so the big market for DSP guys is more hardware/VLSI design than algorithm design and 'pure' DSP.
    That will change once the system realizes that asking a CS graduate with no knowledge of sampling to code speech processing systems is suboptimal, but that'll take time. Until then, not much present.

    DSP is damn near everywhere, at least in the mathematical sense. Telecoms? Check. Information retrieval? Check. Image systems? Check. Robotics and localization/SLAM? Check (mostly to check data integrity). Hardware? Of course, most specialist chips are built to run things like the Viterbi algorithm. Speech processing? Hell yes, but there's little market. Etc, etc, but it's become commiditized enough that one doesn't need a DSP engineer when one can hire a CS generalist, use some libraries and get more utility out of the guy. Sure, the marginal gains in performance from the DSP specialist will be gone, but who cares?
     
    wasprider, vyor and LordVile like this.