Jump to content
Search In
  • More options...
Find results that contain...
Find results in...


  • Content Count

  • Joined

  • Last visited


This user doesn't have any awards

About Nystemy

  • Title
  • Birthday September 5

Profile Information

  • Location
    Sweden, just north of most of Europe.
  • Interests
    Tech but primarily electronics.


  • CPU
    Xeon E5-1650v3
  • Motherboard
    AsRock X99 WS-E
  • RAM
    Crucial 4x 16GB DDR4 2666 MHz (Might get 4 more if I run out of RAM again...)
  • GPU
    Giggabyte GTX 660 x 2 (not running SLI, I just have too many screens, thinking of upgrading, anyone got a pair of GTX960?)
  • Case
    Some old cooler master case that I have no clue how I even got it...
  • Storage
    The usual mix. 512 GB SSD + a handful of TBs on HDDs
  • PSU
    Corsair CX750M
  • Display(s)
    Samsung SyncMaster 2494HS + SyncMaster 223BW + Envision (19" 4:3 aspect ratio) + Cintiq 13HD for the arts....
  • Cooling
    Noctua NH-U9S (Larger wouldn't fit in the case...)
  • Keyboard
    Logitech G110 Nordic version
  • Mouse
    Logitech M500

Recent Profile Visitors

1,253 profile views
  1. To cut zip ties one can take a few different approaches. One can tighten them by hand and then use a pair of flush ground side cutters. (Recommended if you don't do hundreds of them per day. Though, if you need to tighten them to a standard, get the proper tools!) Or one can go the industrial way and use a zip tie installation tool, like the gun shaped thing above. Or one from HellermannTyton. (I would like to have their MK9P-PLGF-BK on my bench, but it costs more than my last two computers combined.) Though, the computer ended up seeming fairly compact indeed. Likely beats
  2. Rendering solutions like this are interesting to be fair. It is a bit like cheating but not at the same time. But it is nice of AMD to make it a more "open" standard. But I also understand why DLSS is fairly closed, since it relies on specific hardware so it isn't much sense in shoehorning it in where it can't easily fit. (Even if some users would happily make driver level amendments to get a feature working.) Though, I do wonder how this technology will pan out in a couple of years. Even DLSS is fairly new, real time ray tracing too. So who knows where things migh
  3. "This is not a servo motor. It just has a sensor sending the information back to our computer [that then controls the motor driver that runs the motor completing the loop]" Ie, it is a servo motor. Since "Servo" just means that it is motor with some form of feedback, be it through the larger system as a whole. (Even a stepper can be a servo if it has a position sensor. A servo motor is though typically either a boring brushed DC motor, or a fancy 3 phase AC one. But people usually just draw a sharp line between "servo" and "non servo" without knowing what servo motor control is.)
  4. Even among servers, supplies like this are rare to be honest. Most power supply manufacturers seems happy with just overbuilding one big power stage instead of going with a handful of smaller ones. Though, a lot of things in society follows that logic, even when some redundancy exists. Might some day make an 12VO PSU just for fun, but build it with 4-8 power stages and see how overkill it can get. The advantage with more power stages in parallel is that it reduces current density, and also allows for lower resistance, and thereby greatly reducing power losses from both sides, all w
  5. I were mostly providing additional content on the topic. Stating the pros and cons of both. I have built both types myself. Resonance is nice if one has a mostly fixed input to output ratio. Like a PFCed power supply supplying 12 volts. And flyback is wonderful if one wants still really good efficiency, but with less hassle. Or where one has a varied input to output ratio. Like battery charging cars. Though, the industry is set on shoehorning in resonance mode supplies here regardless, preferably with one huge transformer and banks of water cooled IGBTs switching in
  6. It depends to be fair. A flyback converter has major advantages in regards to voltage conversion. Ie, choosing an arbitrary input to output ratio is simple, since it only moves over clumps of energy. A resonant mode supply on the other hand is little more than a high frequency inverter running a regular linear transformer, but at far higher frequencies and with materials with less losses, but there is methods to provide some output voltage regulation as well with some downsides in efficiency. The saving grace for resonant mode supplies is the PFC/(boost converter) circuit allowing fo
  7. As an electrical engineer myself, videos like these sometimes makes me cringe. Or most times makes me cringe to be fair... There is the small niggles like "Switch-ing mode power supply" being technically incorrect, since the technical term is just "Switch mode power supply" or SMPS for short, and Varistor aren't called "VerY-resistors". And the primary side mosfets don't actually make a square wave, it is more a saw tooth to be fair. (Not that the mosfets even generates it, it is more a side effect of flyback converters.) But I will be honest and say that this video is honestl
  8. Working in electronics manufacturing and actually doing a fair bit of cable assembly. I usually end up surprised how cheap certain cables are, like USB and network cables are practically free compared to the work needed to construct them. Meanwhile, IDE cables are trivial to make and can be fully machine assembled since it is a ribbon pushed into an IDC connector. Though, machine assembly is in itself far from a trivial thing. Though, in regards to copper cables, they could theoretically reach much further than 3 meters. To me that seems like the twisted pairs are somewhat poorly s
  9. They are however not putting out 17 LTT videos a week, but rather LMG is putting out 17 videos across its 8 channels every week. And in regards to the audience, with 8 different channels, there is 8 different audiences to target. It is more about not putting all eggs in one basket. All though, as it currently stands, most eggs are still in the Linus basket. And to be fair, the LTT channel could likely skip a video every other week or so, I don't think the majority of people would notice to be fair. And those who do are likely the most adherent of fans that would still watch. And if
  10. Truly a ton of work for a ton of videos. And even if one has a bit of a buffer between production and release, one can't technically have too much of that either, since it can create a false sense of security. Honestly more impressive that LMG can keep up with this frankly silly release schedule.
  11. So in regards to this weeks WAN show, Linus makes the argument that talking about MHz in the way he does is something that the audience already "understands."
    While dr. Ian Cutress argues that one should specify that one is talking about the transfers.

    So what is the difference between SDR (single data rate), DDR (double data rate), and QDR (quad data rate)?
    The answer is the clock to data ratio.

    In a 3200 MT/s DDR kit, our clock runs at 1600 MHz, and our data lines run at 3200 MHz. Saying that the data is sent at 3200 MHz is fully correct.
    To a degree it is even correct to say that the data is clocked at 3200 MHz.

    And for everyone now saying, "But the data is clocked at half that, it is DDR after all!"

    Well, the CPU in your computer is typically having a 100 MHz reference clock with some generally arbitrarily picked ratio for the core clock.
    Your CPU isn't running at 100 MHz. It is the same story for the clock on a DDR bus, the 1600 MHz clock in the example above is just the reference for the transceivers.

    Now, RAM manufacturers could have stuck with 100 MHz as their reference clock and used their own PLL internally to get it to whatever frequency the system requires. But phase stability and clock skew is generally worse at larger clock ratios, and high speed memory is fairly sensitive to such issues. This is why the reference clock is synthesized in the CPU's memory controller and then sent over to the RAM chips at a much higher frequency. Since doubling the frequency can be done with relatively little issues.

    But why go with half the frequency then?
    Here the answer is signal integrity, it is simply easier to ensure a nice clean clock signal at a lower frequency. (Noise and other junk can create timing issues for our transceivers.)

    If the Industry moves to QDR or not is a debatable question to be fair.
    Increasing the ratio to 1:4 is not directly beneficial from a signal integrity standpoint, since if we have enough noise issues to not be able to run a simple clock at those higher frequencies, then we likely would struggle to send our data in the first place. And since increasing the ratio leads to worse phase stability in our PLL, then this introduces more signal noise for our data lines to handle. (Ie, sender and receiver aren't as tightly locked with each other creating timing issues that can in turn lead to data loss.)


    But in the end.
    It doesn't matter what the reference clock is, but rather what bitrate the data lines are operating at.
    To a degree, I personally think it is more interesting to talk about average throughput for the channel as a whole. 3200 MHz or 3100 MHz doesn't say much to be fair, RAM timings can make the 3100 MHz memory achieve a higher throughput than the on paper faster 3200 MHz one. It all depends on the memory in question.

  12. To make actual comparisons between different coolers, then one would need far more controlled thermal loads than a test bench. (One can still get rough indications of where something is, but asking the difference between the Hyper 212 and the U12S Redux is going to be blurry without a proper test setup, considering how the difference between the two coolers in the video is already not all that major to start with.) Software reported temperatures, power consumption, fan RPM, etc is fairly loose as far as tolerance is concerned. Not to mention how those measurements are being done in hardware
  13. Removing the glass filler would drastically reduce the wear on the injection mold. Saving a lot of recurring tooling costs. But I also suspect that the fan might use one of their older molds that were collecting dust on a shelf, or used for lower end products already.
  14. And here I thought that, "nah, I will have nothing to comment on in this video..." The thing that blew up were an MLCC cap. This one to be specific: Multi layered ceramic capacitors are actually somewhat notorious for failing as a dead short. It is their primary failure mode, and it is induced by a crack forming and from there a short is usually immanent. It typically comes from mechanical stress. Either through flexing of the board, or thermal cycling. In some industries (Automotive, aerospace, medical, etc) MLCCs have soft terminations to mitigate this issue (or just avoid ML
  15. I will be honest here and say that Linus made the more realistic looking human to be fair. Take bits from each of the two and a better overall result can likely be achieved. But I will have to say one thing, Matthias did have 1 major advantage against Linus. I am willing to bet that he glanced over at least once.... Or were the writer clever and asked Linus to stand there earlier so that they could get a shot from that perspective? (this would though be pointless since one can still move one's head around to get a better view and Linus does have to look around occasionally, not