Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
IT

YouTuber Figured Out Asus Z690 Hero Motherboards Melted Down Due To Backward Capacitor (theverge.com) 54

A YouTuber who goes by the name of Buildzoid on the Actually Hardcore Overlocking channel has figured out that a backward capacitor on the Asus ROG Maximus Z690 Hero motherboard is causing it to melt down, according to a report by Tom's Hardware. From a report: Asus has since acknowledged the issue in a post on its site and plans on issuing replacements to customers with affected motherboards. Problems with the Z690 Hero motherboard started turning up on the Asus support forum, as well as on Reddit, and the issues experienced by users are pretty much identical. As noted by Tom's Hardware, users reported that their motherboards started smoking in the same spot: the two MOSFETs (metal-oxide-semiconductor field-effect transistor) next to the DIMM slots and the Q-code reader. In a video on his channel, Buildzoid diagnoses the issue using only the pictures posted to support forums and on Reddit, attributing the Z690 Hero's failure to the backward capacitor installed next to the MOSFETs, not the MOSFETs themselves. Buildzoid looks closely at the images of the motherboard, pointing out that the text on the capacitor is actually upside down, a potential sign that it's installed incorrectly. As Tom's Hardware mentions, a reversed capacitor results in reversed polarity, causing the MOSFETs to malfunction and burn up.
This discussion has been archived. No new comments can be posted.

YouTuber Figured Out Asus Z690 Hero Motherboards Melted Down Due To Backward Capacitor

Comments Filter:
  • by AlanObject ( 3603453 ) on Monday January 03, 2022 @12:32PM (#62138739)

    It is a pretty common manufacturing error to install polarized capacitors backwards. Tantalum capacitors are the worst because they can last quite a while before they explode.

    Almost every case I have seen is where the silkscreen on the PCB isn't really clear. Usually there are a lot of components in the area, and the + or bar-mark that is used to indicate the correct orientation is sometimes mistaken. The assembly house then programs the pick-and-place machine wrong and the time bomb fuse is thereby lit.

    When this happens it is usually detected at the product burn-in phase where some capacitors pop, then the boards are reworked once root-cause is figured out. But if the burn-in isn't enough then at east some of them get shipped to the field. And this happens.

    • But why isnt it trivially detected by measuring connectivity across many possible pin combinations that will be effected by this error on the board?

      The capacitor has a _purpose_, right?
      • by b0s0z0ku ( 752509 ) on Monday January 03, 2022 @12:40PM (#62138769)
        Because a reversed capacitor will appear to work, until it fails. Time to failure can range from seconds to days or even months.
        • And possibly not detected by the scanning cameras intended to detect such things.

          • Possibly a camera would have detected it but components might be sourced from different companies and would have different markings. In the video, Buildzoid noted the incorrectly installed capacitors were most likely from Panasonic but there was one board that had a capacitor from a company he did not recognize.
          • Wouldn't a scanning camera for QA stuff work off of a reference image? And if the mobo used for said reference image was also put together with a backwards capacitor then the reference image and bad boards would agree ...

      • by matmos ( 8363419 )
        Because it's expensive. It's far cheaper to just take the hit and test a few boards after to see if they work “right away”. They just take the hit if there's a recall. Usually problems like this are limited to a lot or two.
      • by dissy ( 172727 )

        The capacitor has a _purpose_, right?

        This type of capacitor acts like a (very) tiny battery.
        They are placed near the power/ground pins of every component and chip they can, each acting as a small reservoir of power.
        When some component kicks up activity and draws a large but brief spike of power from the rest of the system, that is seen by every other component sharing its power bus as a huge but brief dip in power.

        Various components handle huge drops in voltage differently, and for different amounts of time.
        Their respective capacitors are ther

        • Errm no, Capacitors in this kind of application are generally used as filters to bypass noise, not to hold up rail voltages, that job is done by the power supply rail capacitors.
          https://www.youtube.com/watch?... [youtube.com]

          • by dissy ( 172727 )

            There are two much larger titanium capacitors just below and to the right of the capacitor in question, each of those others being titanium cans with two different farad markings.
            Those other two have all the signs of being noise filtering caps.

            The one in question is appears to be an aluminum capacitor providing smoothing to the switching FETs that burnt up due to the caps incorrect polarity.
            That sounds like a single smoothing capacitor.

            And while we don't really know what exactly was the root of the problem

            • Try watching the linked video. Bypass functions can use up to 3 different cap values, a mix of ceramic( for small values and high frequecies) and electrolytic ( for lower frequencies) each handling a different frequency range of noise, bypass caps are also positioned physically close to a device to eliminate noise induced to the pcb. Ive found after doing many total recap jobs of older electronics, Its quite common to find electros and none polarised types used due to component availability issues. Theres e

    • Had an odd related issue with a small device we had. One batch had a few where apparently an inductor was upside down. You know, the tiny things a couple millimeters long. They possibly jiggled in the reel and a few turned over. Then the glue holds them in place, the camera scanning the board sees that they're there and oriented properly, and passes. There is solder present around all four sides so that orientation isn't normally an issue, but these had a clear plastic cap on the top intended to make i

    • Re: (Score:2, Interesting)

      by Tablizer ( 95088 )

      Why don't they add some kind of socket or extender that doesn't allow it to be connected in the wrong direction?

      A similar mistake almost ruined the Galileo Jupiter atmosphere probe because a part was installed backward. Shape things so that they don't fit the wrong way.

      • by sodul ( 833177 )

        Adding shapes to force putting the capacitor in a specific way would require additional room, and cost. I do not thing the cost benefit warrants it here. While this might be a somewhat common mistake, this happens in a very very small percentage of the production and this recall costs less to the industry overall than designing and using a new fool proof standard for capacitors.

        It does make sense if you are connecting something with a bunch of wires, or something that would be user installed, but for compon

      • by smap77 ( 1022907 )

        Sometimes referred to as poke-yoke or baka-yoke. The difference is in how expensive the malfunction turns out to be.

        https://en.wikipedia.org/wiki/... [wikipedia.org]

      • by matmos ( 8363419 )
        That would be far to expensive. Do you realize how many teeny tiny capacitors there are on a board? Thousands.
        • by ceoyoyo ( 59147 )

          To be fair, most of them don't care which way they're hooked up.

          But the big electrolytics, while fewer, are already one of the most expensive components to place on a board, especially compared to the little SMT ceramics that engineers sprinkle everywhere.

      • by Fons_de_spons ( 1311177 ) on Monday January 03, 2022 @02:17PM (#62139099)
        In my experience, reversing a component in a product is pretty rare. A lot needs to go wrong. There are a lot of different automated checks that test if everything is right, during qualification and afterwards during production. Also, adding sockets to every component that can be oriented wrong tends to be expensive and is still not foolproof. Sounds like someone did a last minute change to the board and skipped a few tests since it was only a small change. Or someone ended up with the wrong version of the PCB during the final production steps. I wish I could slap every engineer/manager who proposes to make a quick small change last minute. Especially in electronics. That's where expensive mistakes are made.
        Always fun when a junior guy learns the hard way. "What do you mean I need to repeat all those tests for this change? I am sure everything works just fine." Most of the times it is, but a lot of disasters were prevented by rerunning the tests. After a few years they become more modest.
    • by AmiMoJo ( 196126 )

      It will have been placed by a pick and place machine, using data from the CAD software. The data includes the location and orientation.

      The most likely error is that whoever loaded the caps into the pick and place machine didn't input the correct orientation (i.e. the machine thought that they were rotated 180 degrees from where they actually were).

      That kind of mistake can be detected using machine vision, but it needs setting up. Because everyone wants a matte black PCB now, even the components are black wi

  • Hard to imagine that minimal testing of the pins on the motherboard wouldnt hack picked this up as out of spec resistances between some pair of pins when unpowered, or an out of spec amps/volts/resists between some pair of pins when powered.
    • Re:Minimal testing? (Score:4, Informative)

      by b0s0z0ku ( 752509 ) on Monday January 03, 2022 @12:37PM (#62138761)
      Reversed electrolytic capacitors don't fail immediately -- it can take some time in operation for the insulating oxide later to fail and for them to go "closed circuit."
    • Testing certain pins of the capacitor would have detected it with a simple meter. Most MB manufacturers are not going to test every component like that as there are hundreds of components per board.
  • by b0s0z0ku ( 752509 ) on Monday January 03, 2022 @12:36PM (#62138755)
    Reversed electrolytic capacitors do not cause reversed polarity in the sense that they reverse current direction. They simply fail because the oxide coating on the plates is lost and they allow current to flow through them like a wire instead of storing charge.
    • I'm surprised and saddened by how far down the comments this is, and how many people above were blathering as if they understood the words.

  • or at least none that i can see.

    what a bunch of amateurs

    also too, "text is upside down" ? wtf does that mean. what's "right side up" ? The interesting thing is, at least in the picture in the article, both capacitors are in the same orientation.

    Furthermore, an electrolytic in backwards ? that thing would let out the smoke much, much faster than those MOSFETs would. But let's say for the sake of argument that the MOSFETs did blow up due to the backward capacitor. The capacitor should absolutely be damaged

    • When he says "upside down" he's talking about the orientation of the text on the capacitor for bad boards vs known good boards. Also, polymer caps (a type of electrolytic cap), which these are, "can withstand for short instants a type dependent reverse voltage for a limited number of cycles". [https://en.wikipedia.org/wiki/Polymer_capacitor#Reverse_voltage]
    • If you watched the video, Buildzoid clearly points out that there is a polarity stripe on the capacitor. In every example of a damaged board, the polarity stripe is on the left whereas it is on the right on working MBs. Also Buildzoid suspects that the MOSFETs that blew up are connected to the capacitor. Thus the MOSFETs would be getting the wrong power inputs thus causing explosive failure. Buildzoid does state there is the possibility that Asus has two versions of the board which have different layouts bu
      • Thus the MOSFETs would be getting the wrong power inputs thus causing explosive failure.

        False. "Thus the capacitors would eventually fail shorted, causing excessive MOSFET output that overheats and destroys the MOSFETs."

        • Umm what? The OP is arguing it cannot be the incorrectly installed capacitor because the capacitor should also should have been severely damaged. My point is that installing the capacitor incorrectly does not necessarily damage the capacitor. The MOSFETs dependent on the capacitor could be damaged and the capacitor remain intact.
          • The idiot says a bunch of stupid shit that is wrong. That shouldn't stop you from understanding my correction.

            Source: am engineer

    • by gweihir ( 88907 )

      Furthermore, an electrolytic in backwards ? that thing would let out the smoke much, much faster than those MOSFETs would.

      It clearly is a polymer capacitor.

  • If you do not follow his channel, he does a lot of in-depth analysis of all sorts of PC circuit boards. I watched a few of his videos on GPU boards where he would go over mundane details of layout, components, vendor sources, etc. Much of it is beyond me but worth watching for the electronic geeks.
  • It sounds like the designers, component manufacturers, and board-assembly plants need to get their heads together so situations like this don't happen.

    If the board-assembly plant's inspection-bots or human inspectors can't look at the finished board and tell when things are backwards, they need to redesign things so it's easier to tell.

    Either that, or do enough burn-in testing so mistakes that "poison" a batch never reach the end user.

  • Rumulans? (Score:3, Funny)

    by Tablizer ( 95088 ) on Monday January 03, 2022 @01:29PM (#62138929) Journal

    This is why Star Trek control rooms explode: they hired a dyslexic race to manufacture the consoles.

    • I have been noticing the same thing. Always wondered why the EPS conduits were routed under the stations on the bridge. Probably as a way to punish people for pressing the wrong button? They are also no longer made as they used to. Kirk's ship did pretty good. Now discovery? That thing is bleeding sparks and rupturing conduits all over the place. Always wonder why the crew does not go on strike untill they fix the damn thing. To many dogooders in the future. ;-)
    • high power at the controls and not remote relays

  • Off to the re-education camp for someone!
  • What you have to do is to compare the very clear marks on the capacitors to the very clear marks on the mainboard. But I guess "youtuber" makes this somehow interesting?

    The description referenced has quite a bit of ignorance in it. Best one is that they noticed the "text" on a capacitor is "upside down" and identify that as a sign of "incorrect installation". That is nonsense on an advanced level as there is no "correct" alignment for the text on a capacitor in the fist place. In actual reality, what capaci

  • by schwit1 ( 797399 ) on Monday January 03, 2022 @02:07PM (#62139055)

    I'm shocked it didn't work here. Every other Star Trek Episode has a scene where reversing the polarity of something magically fixes the problem.

  • So someone reversed the polarity, which was actually a polarized capacitor installed backwards.

    The Q-code reader got me though, I read it as QRcode reader and wtf is that doing on a motherboard, turns out it is the 2 digit LED status code readout that Asus calls Q-codes.

    When the journalists mangle everything like this how the hell can they be called a journalist?

Technology is dominated by those who manage what they do not understand.

Working...