In CI No 1,018, we highlighted the findings of Omaha, Nebraska based disk copying specialist Memory Control Technology Corp that all was far from happy in the 3.5 floppy disk media world, and that even the legendary quality control measures put in place by the leading Japanese manufacturers as part of their campaign to win […]
In CI No 1,018, we highlighted the findings of Omaha, Nebraska based disk copying specialist Memory Control Technology Corp that all was far from happy in the 3.5 floppy disk media world, and that even the legendary quality control measures put in place by the leading Japanese manufacturers as part of their campaign to win world industrial hegemony seemed to be falling down badly in the rush to meet soaring demand for 3.5 floppy disks – but just what are those stringent tests on which the microfloppies are falling down so lamentably?
Missing Bit The minimum standards a floppy should meet are defined by the American National Standards Institute, and Memcon used the most recent draft of its ANSI X3.137-198x spec as the evaluation’s benchmark. The ANSI Missing Bit test is the industry’s most important and widely accepted measure of floppy disk quality. All manufacturers, without exception, advertise that their products meet or exceed this standard. What’s more, they all assert that their disks are individually tested to ensure compliance. However, only five of the 25 brands Memcon tested had 100% of their disks pass this test, namely C Itoh, IBM, Kodak, Sony and TDK. Most brands do not meet ANSI’s minimum standards of quality. Memcon says its results strongly suggest that some manufacturers are taking short cuts in testing their finished products. It could be inferred that products with Missing Bit failure rates above 5% were not tested for that parameter, naming Dysan, SKC and Wabash as the chief offenders. Because of its findings, Memcon concluded that most manufacturers are unable to prevent the single major cause of disk failure. Memcon believes the ANSI-defined Extra Bit test is just as important as the Missing Bit test, disputing experts who feel its redundant and hold that the defects Extra Bit testing turned up will be detected by the Missing Bit test. Despite a lack of unanimity on Extra Bit testing, all brands, again without exception, advertise that their products meet or exceed the standard or are individually tested for compliance. Although 15 manufacturers passed the Extra Bit test with flying colours, 10 had some of their population fail. Dysan and SKC had the worst scores, with BASF, Brown, JVC, Kodak, Opus, 3M, Wabash and Xidex showing a similar failure rate.
ANSI hasn’t devised criteria on durability but Memcon used the de facto standard of 3m revolutions, equal to several years’ normal use. SKC and Wabash failed the test miserably and could only hit 1.5m revolutions before the cookie was totally denuded of its magnetic coating at the point of failure, allowing light to seep through. The results, Memcon says, raise serious questions about longevity, security of stored data and the irreparable damage they can do to a drive. Memcon also checked all the disks visually looking for nine kinds of defects including frayed or visible liners, cookie defects and contaminates, label inconsistencies and shutter scratches. The tests are subjective, a volume buyer, but turned up interesting things like the box of Verbatims that was short a disk. None of the brands was perfect but Fuji, JVC and Kao performed the best with fewer than 10 defects. Opus failed miserably with 224 defects in 100 disks. Dysan, Memory Media, Sentinel and Xidex also did poorly, scoring over 100 visual inconsistencies. The biggest problems were shutter scratching and cookie contamination. Dysan, for instance, had three major cookie defects and in one case its identification label was stuck squarely over the shutter rendering the disk non-functional. This is not to mention the rust blotches on 28 of 100 disks. Opus racked up 26 demerits for glue. Seven out of 10 Xidex boxes were contaminated with what looked like corrugated-cardboard debris. Assuming that all disks should do the job they were designed for at least once, Memcon devised a functionality test. That meant formatting all the disks on an IBM PS/2 Model 50 and copying data onto them and then verifying it. All the test proves, Memcon reiterates, is that the disks will work once. Pass
ing doesn’t ensure long-term functionality or interchangability. However, Memcon found that a significant number of disks were non-functional to begin with, a result that didn’t surprise them given the way the physical and magnetic tests had gone. 12 of the 25 brands had at least one floppy disk fail at one point in the test. One percent of the disks from Centech, Dysan, Nashua, SKC, Sentinel, Wabash and Xidex failed to format. Opus came in with the worst score, two percent.
Failed to copy Three percent of Memory Media’s disks and one percent of Kodak’s, SKC’s and Sony’s failed to copy the data. Two percent of SKC’s and one percent of JVC’s couldn’t be verified. After analysing the data, Memcon has concluded that the ones that failed had serious Missing or Extra Bit defects proving a direct correlation between adherence to ANSI standards and usability. Memcon has also compiled an analysis of prices, which shows that IBM branded disks – admittedly the ones that performed best in the tests are much the most expensive, at an average $3.55 apiece from the distributors Memcon approached, with a high of $4.00 and a low of $2.28. At the other end of the scale, JVC is the cheapest at an average $1.37 with $1.30 the lowest price. Wabash, which didn’t come out of the tests too well, had an average price of $2.64, against Sony, which did well, but turned in an average price of $1.96. And although Kodak owns Verbatim, you seem to pay a premium for the Verbatim name: Kodak floppies came out at an average $2.28 with little variance, Verbatims had an average price $1.71 with an enormous variance, $1.34 for the cheapest, $2.14 for the most expensive. The conclusion seems to be that unless you buy IBM, you don’t necessarily get what you pay for.
