Your Friday post of Capaci-Fail…
All the blown caps are Chinese fake Rubycon and Nichicon. Yes I know how to read a data sheet and confirm that these do not match the specified dimensions and case styles from Rubycon and Nichicon. Nobody at Dell Computer can, or if they can, they were forced to accept these fakes out of cost engineering nonsense. Is it any wonder why I won’t willingly buy a Dell?
Here’s an old school Audioscience and an old school Autogram. The ASI works. The Autogram likes to find new and interesting ways to keep me on my toes.
I don’t think those Rubycon (MCZs)? are fake. They may be custom orders which is why they don’t match what’s in the datasheet. As for those Nichicon HNs, there was a HUGE bad batch of HMs and HNs from 2001-2005, maybe even up to 2007. 100% failure rate on Dell machines within 1-3 years (on machine sold from 2003-2004), especially in the SFF Optiplex machines where heat was an ever expediting factor. I too have suspected if all those huge rashes of failed Nichicons (and Nippon Chemi-con with the horrible KZG and KZJ series, both of which like to fail in storage and unused at that) were fakes, and if this issue extends all the way to other good Japanese brands like Panasonic and Sanyo as some capacitors have been spotted with cheap bullseye bungs that don’t match those specified in the datasheet.
I think the real problem with these ultra low ESR capacitors is water base electrolyte. Water (unlike organic solvent and liquids such as ethyne glycol and gamma butyrolactone) has a tendency to attack the aluminum oxide layer on the anode foil (the dielectric) in an electrolytic capacitor, thinning and completely eroding the layer over time and eventually causing expansion, hydrogen gas to form, and complete dry outs. This tendency very often resulted in capacitors from Nichicon (HM/HN/HZ) showing up as leaky on the ESR or capacitance meter.
Without extreme precision and the proper inhibitors (a less than ideal formula) during the manufacturing process, this electrolyte may be unstable enough even to vent its contents on the shelf. Heat is obviously a huge factor and will greatly accelerate such chemical reactions. Using water content to lower ESR probably wasn’t the best idea but it was the cheaper method for the millions of capacitors being produced (Nippon/United Chemi-con being the largest producer of electrolytics in the world). Of course, knowing that polymers have been taking over for a while, I suppose this becomes less of an issue over time.
Also, I don’t think Dell specifies the choice of capacitors at all. IIRC Dell contracts Intel to build motherboards for them and they outsource to Foxconn. Foxconn has also in turn contracted some companies to assemble the actual boards for them such as Flextronics, at least for OEM boards.
Thank you for your insight and observations on the caps! Yeah, I do wonder about the waterbased electrolyte. It seems like the industry is starting to accept solid electrolyte a lot more, thank goodness….
Part of the problem on the Dell systems is indeed heat, but this was a big wide open machine, in a 65 degree F room. It ran nice and cool but died anyway.
Siiiigh, cost engineering.
The motherboard in this Dell actually says Foxconn on it – it’s of course NOT ATX form factor, it’s got Dell’s weirdness all over it.
Those SFF systems were such a disappointment especially since they natively run off 12vdc power, so they’d be great to install in a vehicle application….. except, surprise, enjoy replacing every cap on the board. Huuuurrrrrkkkkk.