Announcement

Collapse
No announcement yet.

Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    #41
    Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

    Originally posted by mockingbird View Post
    I wonder how varied characteristics of aluminum/electrolyte quality have an effect on the cap.
    I think the electrolyte has the biggest ramifications on the overall reliability, longevity, construction, and chemistry of an electrolytic. This is how series that otherwise are comprised of high quality materials - like KZG and KZJ - are so notorious for failure that they bulge in storage.

    Let's focus on KZE-class series. I've read recently that that was the first major evolution in low ESR caps, in that it employs an aqueous electrolyte, but is stabilized with a proprietary formula. Let's examine some of the contenders:
    Actually, Rubycon ZL was the first major evolution in low ESR capacitors. Rubycon ZL (and ZA) went into production in 1998, and Chemi-con KY, KZE, and KZG all went into production in 2001, around the same time (and Nichicon HD and HC in early 1999, particularly in the nylon laminated Pb-free cases coated with lacquer and whose lead wires are CP tinned).

    Rubycon was the first company to find the proper mixture of inhibitors to suppress the aggressiveness of electrolytic solutions that contain a relatively high amount of H2O content. ZL uses ethylene glycol (an organic solvent) as its solvent and ammonium salts as its solution, working out to roughly 40% H2O content. Rubycon's ZL series was shown to have conductivity in the electrolyte that triples that of previous low ESR series (such as YXG).

    UCC KZE - Reliable, very few reports of failure (and only then under extreme conditions). Do not bulge in storage
    The most bizarre part about that is that UCC claims that KY, KZE, KZG, KZA, and KZJ all use the same electrolyte technology in their datasheets. KZH is said to use a "unique water base electrolyte" and that's it. So either they pushed the electrolyte too far with KZG and KZJ or Chemi-con are unable to control the aggressiveness of their electrolyte when the concentration of H2O reaches a certain threshold (you need to increase the water content or the amount of ions in the electrolyte in order to increase conductivity).

    Nichicon HD - Also reliable, but 8mm variants have been found to bulge in storage
    Only the 1500uF 6.3V Nichicon HDs in 8x20 case sizes are known to be unreliable. All other HDs, no matter their case sizes, are good IMO. I think some early HDs and HCs were known to fail because they lack the sleeve and that increases their heat dissipation quite a bit.

    Rubycon ZL - As reliable as KZE. No known reports of any bizarre failures
    Yup, Rubycon ZLs are almost bullet proof.

    Panasonic FM - As reliable as KZE, perhaps even superior
    Panasonic FF is actually the direct across to KZE grade capacitors, FM is a bit lower ESR I think. Panasonic FF is known to be reliable despite only being a 2,000 hour series.

    Sanyo WX - Not as reliable as KZE, sensitive to heat?
    Yes. It's certainly strange how WG seems to be more reliable, but WF is definitely worse.

    Samxon SH/SK - Do these even exist? The only Samxon capacitors I ever see in equipment are high-voltage primaries or dud series like GF. Someone should open up their X-CON polymers and compare them to UCC parts.
    Samxon GK is actually the equivalent. Known to be unreliable.

    Now the question is, when does a KZE-class cap hold up better? When the aluminum is purer, or when the electrolyte is more stable? In an aqueous cap, I would assume that the electrolyte is more important. But non-aqueous series like PW/LXZ/FC class knock-offs might hold up better with purer aluminum, even if the electrolyte is less stable.
    Presumably both, but I think the electrolyte is the biggest factor, as stated above.

    So it's not that simple. It's a shame Pete from CA isn't that active here anymore. I'm sure he'd have some great insight on this.
    His torture test results were certainly intriguing, but they don't really reflect how these series perform under real world conditions.

    Comment


      #42
      Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

      Originally posted by Wester547 View Post
      Actually, Rubycon ZL was the first major evolution in low ESR capacitors.
      I stand corrected. Speaking of YXG, I've found them to be incredibly reliable and long-lasting. For example, many Asus boards from before the 00's (which used them exclusively) still work fine with them. I have quite a few YXG in my cap box, and none of them show any signs of failure. I even used a couple of 15 year old parts on the 5VSB circuit of an FSP PSU.
      The most bizarre part about that is that UCC claims that KY, KZE, KZG, KZA, and KZJ all use the same electrolyte technology in their datasheets. KZH is said to use a "unique water base electrolyte" and that's it. So either they pushed the electrolyte too far with KZG and KZJ or Chemi-con are unable to control the aggressiveness of their electrolyte when the concentration of H2O reaches a certain threshold (you need to increase the water content or the amount of ions in the electrolyte in order to increase conductivity).
      Rubycon's conventions are also a little confusing. For example, is YXG an analog of NCC KY? Is it an aqueous cap? What's Rubycon's equivalent of LXZ/FC/PW? What's with all their newer series? What's the difference between say YXJ and ZL?

      Take Nichicon PW for example. Slap a nylon bunge on it instead of synthetic rubber, call it "PV" series, and increase the endurance by thousands of hours. But that's usually not the case with these other series, is it?
      Only the 1500uF 6.3V Nichicon HDs in 8x20 case sizes are known to be unreliable.
      That's a very valid point you make now that I think about it. You mentioned before that they "pushed the chemistry too far". So it stands to reason that HD is in fact as good as anything else out there, if we assume that Rubycon didn't try to produce an 8x20 6.3V 1500uF part.
      Samxon GK is actually the equivalent. Known to be unreliable.
      Ok thanks. I was having trouble finding the Samxon version of ZL on their site as I was scanning their datasheets.
      His torture test results were certainly intriguing, but they don't really reflect how these series perform under real world conditions.
      And how could I forget about your own previous brilliant posts, good sir.

      There was either speculation or reference to Samxon importing their electrolyte from Japan for their GC/GD series. Now that you state that GK is a dud series as well, I'm begining to wonder if *most* of Samxon series are duds.

      Which brings me to RS series. RS is a miniaturized series, considering that their 16V 3300uF part in 10mm is only 30mm tall, should we assume its relative is something like KZH?

      Something that might shine some light on all this would be private datasheets from PSU manufacturers. I don't know if you've noticed this, but in a lot of recent PSU reviews, there are Nippon caps with "W" printed on the sleeve as the series. Now I've come across private Nippon datasheets on the web before, which is how I learned that when they get a custom order, they print "W" on the sleeve. What I don't understand then is why they didn't do this for Behemot for his custom KZN order.

      I must commend you on your speculation of the electrolyte differences between the series. It's a topic which should be discussed much more in depth.
      "We have offered them (the Arabs) a sensible way for so many years. But no, they wanted to fight. Fine! We gave them technology, the latest, the kind even Vietnam didn't have. They had double superiority in tanks and aircraft, triple in artillery, and in air defense and anti-tank weapons they had absolute supremacy. And what? Once again they were beaten. Once again they scrammed [sic]. Once again they screamed for us to come save them. Sadat woke me up in the middle of the night twice over the phone, 'Save me!' He demanded to send Soviet troops, and immediately! No! We are not going to fight for them."

      -Leonid Brezhnev (On the Yom Kippur War)

      Comment


        #43
        Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

        Originally posted by mockingbird View Post
        I stand corrected. Speaking of YXG, I've found them to be incredibly reliable and long-lasting. For example, many Asus boards from before the 00's (which used them exclusively) still work fine with them. I have quite a few YXG in my cap box, and none of them show any signs of failure. I even used a couple of 15 year old parts on the 5VSB circuit of an FSP PSU.
        YXG and YXH are pretty good series. They are not as resilient to abuse as HE is, as evidenced by Pete's torture tests, but they might actually decompose in terms of their electrolyte at the same temperature extremes (the 220uF 10V ones did on the Xbox 360 boards that were overheated by way of the oven trick or reflow machines, I believe). Nichicons are cheaper than Rubycons, though - more reason to buy HE when all is said and done.

        Rubycon's conventions are also a little confusing. For example, is YXG an analog of NCC KY? Is it an aqueous cap? What's Rubycon's equivalent of LXZ/FC/PW? What's with all their newer series? What's the difference between say YXJ and ZL?
        YXG (and YXH, a 10,000 hour series) is equal to KY and HE, yes. YXG use ethylene glycol as its solvent as well, but I think it has closer to 25% water content, so quite a bit less than ZL. From the way Nichicon and Chemi-con give description to HE and KY, their water content is probably a bit higher, closer to 30% I'd imagine. KY was once a 6,000 hour series but was "upgraded" to 10,000 hours in April 2002 (reformed electrolyte?). YXJ is minaturized YXH - ZL is much lower ESR.

        Take Nichicon PW for example. Slap a nylon bunge on it instead of synthetic rubber, call it "PV" series, and increase the endurance by thousands of hours. But that's usually not the case with these other series, is it?
        Nichicon PA is Nichicon's lowest non-aqueous ESR series (they employ gamma butyrolactone as their electrolyte, I believe, which comprises 1-2% water content), falling somewhere in between HE and HD.

        That's a very valid point you make now that I think about it. You mentioned before that they "pushed the chemistry too far". So it stands to reason that HD is in fact as good as anything else out there, if we assume that Rubycon didn't try to produce an 8x20 6.3V 1500uF part.
        HD might be a -bit- more sensitive to heat than ZL and KZE, but not enough to be called unreliable. I find it interesting that the custom orders of KZEs, IE smaller case sizes, are not unreliable like the smaller case size HDs are (at least that one custom size), however.

        And how could I forget about your own previous brilliant posts, good sir.
        Haha, thanks. But I don't think my posts are enough to be deemed brilliant. I just don't like to leave any stone unturned as far as information goes. :P

        There was either speculation or reference to Samxon importing their electrolyte from Japan for their GC/GD series. Now that you state that GK is a dud series as well, I'm begining to wonder if *most* of Samxon series are duds.
        Not all of them. GF, GK, possibly SF, and GC and GD are known to be beleaguered with some issues (but GC and GD are more reliable than GF for sure), so their water base series in other words. Samxon do import their aluminum foil from Japan, but everything else - the electrolyte, the cases (the cans), the sleeves, the bungs, the aluminum tabs, the lead wires, the paper separators, are all manufactured in China. So that still leaves quite a window of possibility for premature failure.

        Which brings me to RS series. RS is a miniaturized series, considering that their 16V 3300uF part in 10mm is only 30mm tall, should we assume its relative is something like KZH?
        No, RS is equal to LXY/LXV/LXF, IIRC.

        Something that might shine some light on all this would be private datasheets from PSU manufacturers. I don't know if you've noticed this, but in a lot of recent PSU reviews, there are Nippon caps with "W" printed on the sleeve as the series. Now I've come across private Nippon datasheets on the web before, which is how I learned that when they get a custom order, they print "W" on the sleeve. What I don't understand then is why they didn't do this for Behemot for his custom KZN order.
        Yup, "W" is NCC's new custom order series. A shame, because it would be very difficult to discern where they are supposed to fall ESR wise with no datasheets. You would need something like a LC102 or LC103 Senscore in order to get a truly accurate ESR reading. I think "TM", "TMZ", "TMJ", and "TMV" were Chemi-con's custom series codes before the fact. Maybe Behemot needs to order in a larger volume to receive the "W" designation on the sleeve. Hard to say for certain. However, this is different from series only available to OEMs, for which no datasheets exist - those are not custom series, just series unavailable to the consumer.

        I must commend you on your speculation of the electrolyte differences between the series. It's a topic which should be discussed much more in depth.
        Thanks, and I agree. It's hapless how tight lipped manufactures are about the electrolytic solution and materials, but understandable that they would be safe guarding such secrets. Even if we perform autopsies on them, we really need high end scientific equipment to make sense of the raw materials and reasons for failure.
        Last edited by Wester547; 11-08-2015, 09:11 PM.

        Comment


          #44
          Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

          Originally posted by mockingbird View Post
          There was either speculation or reference to Samxon importing their electrolyte from Japan for their GC/GD series. Now that you state that GK is a dud series as well, I'm begining to wonder if *most* of Samxon series are duds.
          GK and GF are known to be bad, those are both the green-sleeved caps in displays and CWT PSUs.
          Originally posted by mockingbird View Post
          Something that might shine some light on all this would be private datasheets from PSU manufacturers. I don't know if you've noticed this, but in a lot of recent PSU reviews, there are Nippon caps with "W" printed on the sleeve as the series. Now I've come across private Nippon datasheets on the web before, which is how I learned that when they get a custom order, they print "W" on the sleeve. What I don't understand then is why they didn't do this for Behemot for his custom KZN order.
          Interesting. Well, I asked specifically for KZN so they just made it new KZN case size. I think that when sufficiently pressed/having high enough order, they will make anything in D10x40 (or more) mm as it's their pencap can. And pencaps are what? Pencaps are in fashion today so they have them to. But nobody other than crapxon makes 8x25+ size. Not yet though.
          Less jewellery, more gold into electrotech industry! Half of the computer problems is caused by bad contacts

          Exclusive caps, meters and more!
          Hardware Insights - power supply reviews and more!

          Comment


            #45
            Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

            Originally posted by Wester547 View Post
            Okay, I hate to double post, but I was wrong. Man Yue do get their aluminum foil from Japan, so the fault must lie within their electrolyte:

            http://www.jcc-foil.co.jp/eng/company/outline.html

            Taicon hold claim to having sourced their aluminum foil from the same company since 1995, by the by.
            Okay, it turns out that by Man Yue's own admission, they do actually have three factories in China for producing their own aluminum foil:



            And according to this, Man Yue was simply a distributor of JCC products as of 1998:

            http://www.scmp.com/article/245670/m...joint-ventures

            Taicon do source their paper separators, electrolyte, and aluminum foil from Japan and sometimes even Nichicon, but sometimes unfortunately do get their aluminum from China (and process all their other materials in-house, IE Taiwan and/or China):

            http://cfile222.uf.daum.net/attach/1...4C08890EC02343

            http://webcache.googleusercontent.co...&ct=clnk&gl=us

            However, it seems that none of the aforementioned companies suffer enough from aluminum foil issues for that to be the root cause of their high failure rate when it comes to a number of their series. Or maybe I'm wrong and it's just a crap shoot - IOW inconsistent quality in terms of materials.
            Last edited by Wester547; 11-10-2015, 07:40 PM.

            Comment


              #46
              Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

              Originally posted by mockingbird View Post
              Rubycon's conventions are also a little confusing. For example, is YXG an analog of NCC KY? Is it an aqueous cap? What's Rubycon's equivalent of LXZ/FC/PW? What's with all their newer series? What's the difference between say YXJ and ZL?
              Rubycon went with mixed-aqueous electrolytics fairly early in the game, with ZA/ZL. I don't know about the YXA/YXG/YXF/YXH series, but they're probably less aqueous than ZL, and hence roughly the equivalent of FA/FC/FK, LXA/LXF/LXY/LXV/LXZ and PJ/PL/PM/PR/PS/PW, which are all quarternary-ammonium electrolyte non-aqueous series.

              So it stands to reason that HD is in fact as good as anything else out there, if we assume that Rubycon didn't try to produce an 8x20 6.3V 1500uF part.
              HD is indeed very good. Rubycon did produce a 1500uF/10V 8x20 ZLH, but AFAIK it is rated at some 6000..8000 hours and I haven't seen any failures with it. It's a bit of a rare bird, since it showed up pretty late in the game when mobo manufacturers were generally shifting to polymers.

              ... there are Nippon caps with "W" printed on the sleeve as the series. Now I've come across private Nippon datasheets on the web before, which is how I learned that when they get a custom order, they print "W" on the sleeve.
              Cool, makes sense. I have some brown-sleeved NCC W-series caps rated at 1000uF/18V. It would probably be just a regular 16v series re-qualified during testing at 2V higher voltage.
              Last edited by linuxguru; 11-10-2015, 08:15 PM.

              Comment


                #47
                Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                Originally posted by Wester547 View Post
                Panasonic FF is actually the direct across to KZE grade capacitors, FM is a bit lower ESR I think. Panasonic FF is known to be reliable despite only being a 2,000 hour series.
                The FFs I've seen are all in silver-on-dark-blue sleeves, which suggests that FC/FK/FM (all in gold-on-dark-blue or gold-on-black) are lower ESR than FF. I'll measure and check the ESR on the few 1000uF/10V 8x20 FFs that I still have lying around, and compare with the 1500uF/6.3V 8x20 FJ just for reference.

                Comment


                  #48
                  Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                  Originally posted by linuxguru View Post
                  Rubycon went with mixed-aqueous electrolytics fairly early in the game, with ZA/ZL. I don't know about the YXA/YXG/YXF/YXH series, but they're probably less aqueous than ZL, and hence roughly the equivalent of FA/FC/FK, LXA/LXF/LXY/LXV/LXZ and PJ/PL/PM/PR/PS/PW, which are all quarternary-ammonium electrolyte non-aqueous series.
                  IIRC, Pete tore apart a Nichicon PW, Chemi-con LXZ, and Panasonic FC many years ago, so as to score for the presence of H2O content, and found that they have the organic solvent gamma-Butyrolactone. Generally, non-aqueous capacitors can be broken down into three types of organic electrolyte:

                  1. gamma-Butyrolactone (GBL)
                  2. Dimethylformamide (DMF)
                  3. Dimethyl Acetal (DMA)

                  There are a number of other weak, lactic acids, and ammonium salts such as those quaternary, used as organic solvents as well. Nichicon PR, PF, and Nippon LXF are some of the series that used that meticulous solvent. The advantage of non-aqueous capacitors is very good leakage current characteristics and stability in the long term - the hydrogen outgassing and pressure is less a likelihood and corrosion based on the eventual dissolving of the oxide layer into the solution a non issue during actual usage. Hillman's 2004 study found that the Japanese samples in their tests did have the phosphate ions and inhibitors to avert the H2O-based corrosion (only one of the two, though - possibly because one of the samples was a non-aqueous series) and that the Taiwanese samples did not, instead including ammonium ions in insufficient amounts with pH levels that were far too high.

                  As for Panasonic FF? Look at the datasheet here:

                  https://www.badcaps.net/forum/showpo...52&postcount=1

                  It's a direct cross to HD, ZL, KZE, WX... the different colored sleeves are likely due to the materials used. Polyester, Pb-free Polyvinyl chloride, or Polyolefin...
                  Last edited by Wester547; 11-10-2015, 09:26 PM.

                  Comment


                    #49
                    Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                    Originally posted by Wester547 View Post
                    As for Panasonic FF? Look at the datasheet here:

                    https://www.badcaps.net/forum/showpo...52&postcount=1

                    It's a direct cross to HD, ZL, KZE, WX... the different colored sleeves are likely due to the materials used. Polyester, Pb-free Polyvinyl chloride, or Polyolefin...
                    Thanks for the explanation and the link to the FF datasheet - it looks respectable, though not as low in ESR as the more recent FJ, FJS, etc.

                    Regarding the reliability of Nichicon HD, I may have spoken too soon. I just found one 3900uF/6.3V/12.5mm aluminium/lacquer-finished HD with date code HD124 which has bulged and spilled electrolyte in storage. Granted, it is over a decade old, has never had any bias applied to it to inhibit foil corrosion, and it's only one of over ~20 in that lot that has bulged, but it's still a cause for concern, since I've used that exact same value for Vcore bypass in a variety of motherboards, including a Shuttle FV24 (none of which has failed yet). The failed HD measures with very high capacitance (22,000 uF), ESR of 40 mohm and high Vloss (20%), noticeable bulging on the vent side as well as lower rubber bung. I'm guessing that it has high leakage current. The not-yet-failed cohorts in the same lot also measure with high capacitance of 5000 uF+, but with lower ESR (20..30 mohm) and Vloss in the region of 2% - I'm guessing that they'll eventually fail in storage unless they're re-formed with bias.

                    Post-2005 HM seems to hold up better, both in storage and operational usage. An unused 3300uF/6.3V/10mm HM with date code H0533 measures as 3810 uF and ESR of ~30 mohm.
                    Last edited by linuxguru; 11-11-2015, 12:44 AM. Reason: addendum

                    Comment


                      #50
                      Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                      Originally posted by linuxguru View Post
                      Regarding the reliability of Nichicon HD, I may have spoken too soon. I just found one 3900uF/6.3V/12.5mm aluminium/lacquer-finished HD with date code HD124 which has bulged and spilled electrolyte in storage. Granted, it is over a decade old, has never had any bias applied to it to inhibit foil corrosion, and it's only one of over ~20 in that lot that has bulged, but it's still a cause for concern, since I've used that exact same value for Vcore bypass in a variety of motherboards, including a Shuttle FV24 (none of which has failed yet). The failed HD measures with very high capacitance (22,000 uF), ESR of 40 mohm and high Vloss (20%). I'm guessing that it has high leakage current. The not-yet-failed cohorts in the same lot also measure with high capacitance of 5000 uF+, but with lower ESR (20..30 mohm) and Vloss in the region of 2% - I'm guessing that they'll eventually fail in storage unless they're re-formed with bias.
                      Wow, that is very disappointing. Just to be clear, were these ever soldered to a PCB or do they still retain their leads in full?

                      I guess Nichicon had problems with their water base series until 2006-2008.... HD, HC, HV, HM, HN, HZ... I don't think HE has enough H2O content to wreak corrosion upon the dielectric. Taicon had issues with their HD and HI (equal to HM) series as well, which may include their HL (equal to HN), HF (equal to HD) and HH (equal to HV) series too, but otherwise, Taicon seems to be a very respectable capacitor manufacturer out of Taiwan. The fact that your Nichicon HD measures high capacitance (besides the fact that it's just been sitting in storage, of course, which in and of itself dissolves the dielectric) means that the dielectric film on the anode is austerely thinned, and the water base electrolyte has consumed it in storage more or less. Nichicon didn't use very good passivating mechanisms I guess. I think Nichicon HD and Suncon WX must be significantly more aqueous than Rubycon ZL, Panasonic FF, and Chemi-con KZE. 0.03 ohms and 3810 uF is a bit high for a 3300uF 6.3V 10x25 Nichicon HM IMO, though it is unused.
                      Last edited by Wester547; 11-11-2015, 12:57 AM.

                      Comment


                        #51
                        Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                        Originally posted by Wester547 View Post
                        Wow, that is very disappointing. Just to be clear, were these ever soldered to a PCB or do they still retain their leads in full?
                        Leads are intact in full, and they were never soldered or had bias applied to them.

                        I think Nichicon HD and Suncon WX must be significantly more aqueous than Rubycon ZL and Chemi-con KZE.
                        Likely - a few drops of clear electrolyte with a slight yellowish tint had leaked out of the bulged unit, but no perforation is readily apparent in the vent (to the naked eye) - it must be a small one at the intersection of the score lines.

                        Comment


                          #52
                          Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                          Well, if you want a cap which has never been used to last a decade…is it even reasonable demand? Just try reforming them while there is still time left and see how that goes…
                          Less jewellery, more gold into electrotech industry! Half of the computer problems is caused by bad contacts

                          Exclusive caps, meters and more!
                          Hardware Insights - power supply reviews and more!

                          Comment


                            #53
                            Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                            Originally posted by linuxguru View Post
                            Post-2005 HM seems to hold up better, both in storage and operational usage. An unused 3300uF/6.3V/10mm HM with date code H0533 measures as 3810 uF and ESR of ~30 mohm.
                            Interesting.

                            The 10x20 part should be 13 mOhm and the 10x25 part 12 mOhm.

                            I wonder if the ESR would go back to normal after re-forming...
                            "We have offered them (the Arabs) a sensible way for so many years. But no, they wanted to fight. Fine! We gave them technology, the latest, the kind even Vietnam didn't have. They had double superiority in tanks and aircraft, triple in artillery, and in air defense and anti-tank weapons they had absolute supremacy. And what? Once again they were beaten. Once again they scrammed [sic]. Once again they screamed for us to come save them. Sadat woke me up in the middle of the night twice over the phone, 'Save me!' He demanded to send Soviet troops, and immediately! No! We are not going to fight for them."

                            -Leonid Brezhnev (On the Yom Kippur War)

                            Comment


                              #54
                              Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                              Originally posted by Wester547 View Post
                              non-aqueous capacitors can be broken down into three types of organic electrolyte:

                              1. gamma-Butyrolactone (GBL)
                              2. Dimethylformamide (DMF)
                              3. Dimethyl Acetal (DMA)
                              SMD electrolytics use cat-piss as electrolyte.

                              Comment


                                #55
                                Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                                Been talking to former Man Yue employee and got some interesting information…

                                They use best quality foils and electrolyte imported from Japan for important manufacturer customers like Osram. As for the rest, especially resellers, I guess any combination of materials may be possible. I would expect some series (those known to be bad) to constantly use bad materials and vice versa.

                                What does not pleased me at all is that Many Yue supposedly has very fluctuating quality between batches as they do not test them enough/do not have capabilities to test properly all of them.
                                Less jewellery, more gold into electrotech industry! Half of the computer problems is caused by bad contacts

                                Exclusive caps, meters and more!
                                Hardware Insights - power supply reviews and more!

                                Comment


                                  #56
                                  Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                                  well thats not good. thats a strike against them due to poor QC and making their products inconsistent.

                                  my opinion on putting non-jap caps as second tier (due to their lack of "EQ" behind their product) still stands.

                                  Comment


                                    #57
                                    Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                                    Originally posted by Behemot View Post
                                    Been talking to former Man Yue employee and got some interesting information…

                                    They use best quality foils and electrolyte imported from Japan for important manufacturer customers like Osram. As for the rest, especially resellers, I guess any combination of materials may be possible. I would expect some series (those known to be bad) to constantly use bad materials and vice versa.

                                    What does not pleased me at all is that Many Yue supposedly has very fluctuating quality between batches as they do not test them enough/do not have capabilities to test properly all of them.
                                    Yup, QC testing is everything, and I suspected that is the #1 issue with all the "bad" brands. It isn't enough just to import materials from Japan or use high purity materials. Most poor brands don't even bother with QC testing - hence the very inconsistent quality. Samwha's QC testing must also be poor despite using Japanese imported aluminum foil. Makes me wonder how good Taicon's QC testing is, despite the fact that they import a handful of their materials from Japan and that several Nichicon employees do oversee their manufacturing process according to one of the documents I linked. Toshin Kogyo's QC testing (or the complete lack thereof) must be poor for sure - hence the exceedingly poor quality despite the high purity of materials (they claim their QC testing is great, but that's marketing).

                                    So, the upshot is, the Samxons found in consumer equipment - Hipro PSUs, CWT PSUs, Lite-on PSUs, LCD monitors, old "EVGA" rebadged nVidia motherboards, Intel/Foxconn/Pegatron motherboards, are of flaky quality at best, but the Samxons sold to Behemot, bigpope, Topcat, are of excellent quality? At least that's what I can chalk this up to. Man Yue do claim that their QC testing process is excellent in their publicly available documents - that must be marketing again.

                                    This makes me wonder about the Japanese brands, though. Nippon Chemi-con is the biggest producer of electrolytics on the planet - maybe they just didn't have the time or resources to properly QC test all those KZGs and KZJs (hence that KZJs and KZVs were discontinued as early as late 2007 because of "production" issues), or something slipped them by in the QC testing process for sure, hence the crap quality. I wonder if this was really the issue with early HMs/HNs/HZs (and some early HDs/HCs) and if Nichicon corrected that later on. This may extend to some series from some other "good" brands as well - I wonder if the demand for some of these bad series was so high at a time that even the good brands had to bypass QC testing altogether so as to meet the demand of the customers.

                                    Comment


                                      #58
                                      Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                                      Originally posted by Wester547 View Post
                                      This makes me wonder about the Japanese brands, though. Nippon Chemi-con is the biggest producer of electrolytics on the planet - maybe they just didn't have the time or resources to properly QC test all those KZGs and KZJs (hence that KZJs and KZVs were discontinued as early as late 2007 because of "production" issues), or something slipped them by in the QC testing process for sure, hence the crap quality.
                                      I think that some NCC KZGs were outsourced - maybe to Korea Chemicon or Samyoung or similar. Whether that was the cause for the KZG failures, I don't know.

                                      I have some 820uF/6.3V/8x24 mm, with sharp legend print, shiny aluminium body, and fine vent scoring (ending before the sleeve edge). This is similar to some known Japanese-production LXZ that I've used earlier, so I'm confident that this is authentic. Date code 58 E(3), measures at ~20 mohm ESR. No apparent failures in a lot of ~12 in storage.

                                      OTOH, I also have some 1500uF/6.3v/8x20 mm, with larger and less distinct legend print, slightly more matt aluminium body, wider vent score lines that extend below the edge of the sleeve. Date code 38 C(3), measures at ~40 mohm ESR. A couple of failures (bulged) in a lot of about 50 in storage.

                                      I'm inclined to believe that the latter are outsourced or produced in a "lesser" Asian production facility, though nominally controlled by NCC, and have lower production and/or QC standards. The bulk of the KZG failures in mobos may be accounted for by the latter type - the ones I've seen on an Asus M2A-MX motherboard are of similar appearance, with date codes of 7(T) 90 and 7(T) 9K, but no visible failures yet.
                                      Last edited by linuxguru; 11-14-2015, 09:19 PM. Reason: addendum

                                      Comment


                                        #59
                                        Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                                        Originally posted by linuxguru View Post
                                        I think that some NCC KZGs were outsourced - maybe to Korea Chemicon or Samyoung or similar. Whether that was the cause for the KZG failures, I don't know.
                                        Samyoung use different aluminum cans than Chemi-con. There are subtle but evident and consistent differences in the finer, more studious details of each individual case. Samyoung cases tend to have thinner vent stamps and a more silvered finish in terms of color than NCC's. The difference in stenciling clarity could mean anything, and that is bound to vary by any manufacturer with millions being produced at any given time, especially when the actual sleeve quality may vary based upon the request for PET or PVC sleeves.

                                        I have some 820uF/6.3V/8x24 mm, with sharp legend print, shiny aluminium body, and fine vent scoring (ending before the sleeve edge). This is similar to some known Japanese-production LXZ that I've used earlier, so I'm confident that this is authentic. Date code 58 E(3), measures at ~20 mohm ESR. No apparent failures in a lot of ~12 in storage.
                                        That date code would translate to August 5th of 2005, factory "3", according to UCC's date decoding system. Those 8x25 820uF 6.3V KZGs were a custom order made for Foxconn, I believe. They were used interchangeably with 8x20 820uF 6.3V MCZs and 8x20 820uF 6.3V HNs in the VRM output of older Socket 478 motherboards. The HNs of the timeframe were the most prone to failure, followed by the KZGs and MCZs. 0.020 ohms is a bit on the high side - an unused capacitor of that class should measure just a bit lower IMO, but then again they are KZGs...

                                        OTOH, I also have some 1500uF/6.3v/8x20 mm, with larger and less distinct legend print, slightly more matt aluminium body, wider vent score lines that extend below edge of the sleeve. Date code 38 C(3), measures at ~40 mohm ESR. A couple of failures (bulged) in a lot of about 50 in storage.
                                        These were never soldered or had bias applied to their plates, I take it? Do they measure high capacitance or low capacitance (or in between)? Sounds like those were manufactured on August 3rd of 2003, factory "3" again. It's common for KZGs to bulge in storage, especially the 3300uF 6.3V 10x25 ones. Those are the worst, and will bulge in storage whether the print on the legend is distinct or not. The ones you found on the ASUS board were manufactured on September 11th and 15th of 2007, factory "T" (Japan, Tokyo likely).

                                        I'm inclined to believe that the latter are outsourced or produced in a "lesser" Asian production facility, though nominally controlled by NCC, and have lower production and/or QC standards. The bulk of the KZG failures in mobos may be accounted for by the latter type.
                                        Well, I think KZGs are just poor quality, truth be told. I've seen KZGs whose sleeves are almost identical to KZEs in every respect - the materials used, the quality of the print, the values, the factories from whence they came, but only KZGs are known to be so unstable that they outgas without ever having ripple current applied to them or a bias across their plates. It seems strange to me that UCC would allot such respectable resources - a full team of QC chemists, and do the proper research - for all their other series, but just let a couple of series go to crap. In simple terms, they screwed something up.
                                        Last edited by Wester547; 11-14-2015, 10:04 PM.

                                        Comment


                                          #60
                                          Re: Samxon capacitor quality discussion (spin-off from OCZ Bug Zapper thread)

                                          Originally posted by Wester547 View Post
                                          0.020 ohms is a bit on the high side - an unused capacitor of that class should measure just a bit lower IMO, but then again they are KZGs...
                                          My measurements may be off (higher) by 10..20 mohms, due to contact resistance at the leads that I cannot null out in my setup.

                                          These were never soldered or had bias applied to their plates, I take it? Do they measure high capacitance or low capacitance (or in between)?
                                          Never soldered or had bias applied, but these are a "cut lead" lot. A failed 1500uF measures with C of about 2500uF and anomalous high leakage current - the tester claims that is a cell (i.e. battery) in one direction and diode with Vf=274 mV in another (typical for thinned or non-existent oxide - it probably shares some characteristics of metal-semiconductor diodes).

                                          A good (non-bulged) 1500uF/6.3 measures as 1805uF/40 mohm and Vloss of 1.8%. Slightly high, but still almost within the 20% spec for capacitance after 12 years, so it's not too bad.


                                          Sounds like those were manufactured on August 3rd of 2003, factory "3" again. It's common for KZGs to bulge in storage, especially the 3300uF 6.3V 10x25 ones. Those are the worst, and will bulge in storage whether the print on the legend is distinct or not.
                                          Yes, I have a few of those, but they had bulged in use.

                                          Well, I think KZGs are just poor quality, truth be told. I've seen KZGs whose sleeves are almost identical to KZEs in every respect - the materials used, the quality of the print, the values, the factories from whence they came, but only KZGs are known to be so unstable that they outgas without ever having ripple current applied to them or a bias across their plates. It seems strange to me that UCC would allot such respectable resources - a full team of QC chemists, and do the proper research - for all their other series, but just let a couple of series go to crap. In simple terms, they screwed something up.
                                          Clearly, they missed something in the formula for KZG versus KZE. It should have been easy to fix, unless the two electrolyte formulations are vastly different in their electrochemical reactions, which seems unlikely.

                                          Comment

                                          Working...
                                          X