Loading ...
Sorry, an error occurred while loading the content.

Re: CRT monitor dying?

Expand Messages
  • Mariss Freimanis
    That certainly will work but is impractical in production conditions. We produce a product that has 151 SMT components per board and we build about 200 boards
    Message 1 of 30 , Nov 30, 2006
    • 0 Attachment
      That certainly will work but is impractical in production conditions.
      We produce a product that has 151 SMT components per board and we
      build about 200 boards a day. These boards are visually inspected
      after reflow under a low-power stereo microscope. All that can be
      caught are missing parts, upside down parts, tombstoned parts and bad
      solder joints.

      The fractured X7R caps were only caught during electrical test of the
      assemblies. The problems appeared when we switched to ROHS compliant
      parts, board and solder paste. Our solution was to ban X7R caps from
      the entire design. NPO and COG parts were designed in. No problems
      since then.

      The Y5V problem started then too but is rare (1:10,000 or so).

      I really hate ROHS.

      Mariss



      --- In Electronics_101@yahoogroups.com, "jverive" <jverive@...> wrote:
      >
      > You can see the cracks in an SMD ceramic device by conducting light
      through the device from the side. In tight spaces, I have used single
      optical fibers. I have also custom ground and polished small glass
      rods (with square cross-sections) to serve as miniature inspection
      "periscopes". Together with conducted side illumination, I've been
      able to obtain photos and microscope images of single failed ceramic
      SMD caps buried among countless other parts. The images are quite
      striking, since the conducted light intensity drops off at the crack
      due to the refractive index changes.
      >
      > - Jeff
      >
      > --- In Electronics_101@yahoogroups.com, "Mariss Freimanis"
      <mariss92705@> wrote:
      > >
      > > Ceramic SMT capacitors have their very own charming eccentricities.
      > > Examples:
      > >
      > > X7R caps are mechanically fragile. Reflow thermal stresses fracture
      > > the caps unless the reflow process is under very tight control. Result
      > > is zero capacitance even though the cap is visually perfect. On
      > > hot-air rework the cap comes off in 2 pieces with the fracture near
      > > the metallized end.
      > >
      > > Y5V caps (multi-uF) develop shorts even when operating at small
      > > fractions of their rated voltages. Capacitance varies greatly with
      > > applied voltage.
      > >
      > > Mariss
      > >
      > >
      > >
      > > --- In Electronics_101@yahoogroups.com, "Stefan Trethan"
      > > <stefan_trethan@> wrote:
      > > >
      > > > On Thu, 30 Nov 2006 21:09:10 +0100, lcdpublishing
      > > > <lcdpublishing@> wrote:
      > > >
      > > > > Wow, That is very disturbing. For something as common as Caps
      > > > > which are in just about anything electronic, you would think that
      > > > > their useful life would be better than 10 years. Then again, what
      > > > > electronic components have an active life that long. Hmmm, I do
      > > > > have a stereo that is about 25 years old - all the TVs are
      less than
      > > > > 20, most less than 10. Computers are all less than 6 years
      old. I
      > > > > suppose, most of this stuff isn't intended to last 25 years - just
      > > > > seems strange at first glance.
      > > > >
      > > >
      > > >
      > > > The useful life often is much longer.
      > > > But because of their construction - a little like batteries - they
      > > have a
      > > > higher chance of failing than other stuff.
      > > > When nothing goes wrong and they are used right (= not hot),
      they will
      > > > last 30 years and more.
      > > > Most of the time nothing goes wrong, but they are still a high
      risk
      > > > component compared to others.
      > > >
      > > > You'll see, wait a few decades and 90% of 'lytics will be replaced
      > > with
      > > > ceramics or some other technology.
      > > > Operating frequencies and circuit speeds go up and up all the time,
      > > that
      > > > means there's less need for large filter capacitors, on the
      other hand
      > > > ceramic capacity goes up dramatically as well, i remember only a few
      > > years
      > > > ago someone was so amazed about 1u ceramic caps in 0805 that he
      > > actually
      > > > put them in the RLC bridge before he believed it's true ;-)
      > > >
      > > > Wet aluminum technology will not be around forever. They also fit
      > > badly
      > > > with modern manufacturing and SMD technology, all that can stuff
      and
      > > > sealing stuff and wet stuff and large stuff and.. nah.. not nice ;-)
      > > > You'll see, they'll eventually go away.
      > > >
      > > > Now i'm not saying that things will get more reliable because of it,
      > > but
      > > > that's not because of bad components, that's just bad
      engineering and
      > > > economics.
      > > >
      > > > ST
      > > >
      > >
      >
    • jverive
      Mariss, Indeed, the inspection procedures I mentioned were for a failure analysis lab I used to run (before having to go on disability). As you have found, the
      Message 2 of 30 , Dec 1, 2006
      • 0 Attachment
        Mariss,

        Indeed, the inspection procedures I mentioned were for a failure analysis lab I used to run (before having to go on disability).

        As you have found, the high-k dielectrics (used in X7R, Y5V, and Z5U caps, for example) are much more brittle than the low-k dielectrics used in C0G (also known as NP0, for Negative-Positive zero, which refers to the low thermal coefficient). However, ALL ceramic SMD caps have the potential for thermal or stress cracks. Of the various app notes I've read, I like those put out by KEMET. These cracks are going to be even bigger problems with lead-free solders, and post-assembly testing is never foolproof, and neither is optical inspection (even if it could be automated). We had some success with bed-of-nails testing, though these have strict mechanical requirements to keep from causing cracks themselves. Cracked power plane decoupling caps are notoriously hard to detect because of the low node impedance, yet these have the greatest potential to cause nasty damage when they eventually short out, due to the vast amount of energy available.

        I could go on and on about the difficulties in dealing with cracked ceramic caps, but I want to keep this posting from getting too long.

        - Jeff

        In Electronics_101@yahoogroups.com, "Mariss Freimanis" <mariss92705@...> wrote:
        >
        > That certainly will work but is impractical in production conditions.
        > We produce a product that has 151 SMT components per board and we
        > build about 200 boards a day. These boards are visually inspected
        > after reflow under a low-power stereo microscope. All that can be
        > caught are missing parts, upside down parts, tombstoned parts and bad
        > solder joints.
        >
        > The fractured X7R caps were only caught during electrical test of the
        > assemblies. The problems appeared when we switched to ROHS compliant
        > parts, board and solder paste. Our solution was to ban X7R caps from
        > the entire design. NPO and COG parts were designed in. No problems
        > since then.
        >
        > The Y5V problem started then too but is rare (1:10,000 or so).
        >
        > I really hate ROHS.
        >
        > Mariss
        >
        >
        >
        > --- In Electronics_101@yahoogroups.com, "jverive" <jverive@> wrote:
        > >
        > > You can see the cracks in an SMD ceramic device by conducting light
        > through the device from the side. In tight spaces, I have used single
        > optical fibers. I have also custom ground and polished small glass
        > rods (with square cross-sections) to serve as miniature inspection
        > "periscopes". Together with conducted side illumination, I've been
        > able to obtain photos and microscope images of single failed ceramic
        > SMD caps buried among countless other parts. The images are quite
        > striking, since the conducted light intensity drops off at the crack
        > due to the refractive index changes.
        > >
        > > - Jeff
        > >
        > > --- In Electronics_101@yahoogroups.com, "Mariss Freimanis"
        > <mariss92705@> wrote:
        > > >
        > > > Ceramic SMT capacitors have their very own charming eccentricities.
        > > > Examples:
        > > >
        > > > X7R caps are mechanically fragile. Reflow thermal stresses fracture
        > > > the caps unless the reflow process is under very tight control. Result
        > > > is zero capacitance even though the cap is visually perfect. On
        > > > hot-air rework the cap comes off in 2 pieces with the fracture near
        > > > the metallized end.
        > > >
        > > > Y5V caps (multi-uF) develop shorts even when operating at small
        > > > fractions of their rated voltages. Capacitance varies greatly with
        > > > applied voltage.
        > > >
        > > > Mariss
        > > >
      • Mariss Freimanis
        Jeff, Thanks for the reply and I agree. We use the best components possible, International Rectifier for MOSFETs and certainly Kemet for capacitors. Same for
        Message 3 of 30 , Dec 1, 2006
        • 0 Attachment
          Jeff,

          Thanks for the reply and I agree. We use the best components possible,
          International Rectifier for MOSFETs and certainly Kemet for
          capacitors. Same for all other parts as well; at top-line vendor's
          parts are 10% more expensive than the also-rans. It makes no sense to
          "save" 10% up front only to pay back 50% in rework or worse.

          The biggest challange has been adapting to ROHS. We have things pretty
          well worked out now but it wreaked havoc in the beginning. No matter
          how you cut it, it is an inferior process compared to SnPb.

          I just hope it satisfies the mud-hut and wattle mentality of the green
          type's vision of Vallhala until the next 'thing'. There will always be
          a next 'thing' with these anachronists until technology is eliminated.

          Mariss



          --- In Electronics_101@yahoogroups.com, "jverive" <jverive@...> wrote:
          >
          > Mariss,
          >
          > Indeed, the inspection procedures I mentioned were for a failure
          analysis lab I used to run (before having to go on disability).
          >
          > As you have found, the high-k dielectrics (used in X7R, Y5V, and Z5U
          caps, for example) are much more brittle than the low-k dielectrics
          used in C0G (also known as NP0, for Negative-Positive zero, which
          refers to the low thermal coefficient). However, ALL ceramic SMD caps
          have the potential for thermal or stress cracks. Of the various app
          notes I've read, I like those put out by KEMET. These cracks are
          going to be even bigger problems with lead-free solders, and
          post-assembly testing is never foolproof, and neither is optical
          inspection (even if it could be automated). We had some success with
          bed-of-nails testing, though these have strict mechanical requirements
          to keep from causing cracks themselves. Cracked power plane decoupling
          caps are notoriously hard to detect because of the low node impedance,
          yet these have the greatest potential to cause nasty damage when they
          eventually short out, due to the vast amount of energy available.
          >
          > I could go on and on about the difficulties in dealing with cracked
          ceramic caps, but I want to keep this posting from getting too long.
          >
          > - Jeff
          >
        Your message has been successfully submitted and would be delivered to recipients shortly.