History (1988): How to Choose 3.5-Inch Diskettes?

A report released by Memory Control Technology Corp. (Memcom) says most brands of 3.5-inch diskettes fail the pertinent quality tests set forth by the American National Standards Institute (ANSI).

The best 3.5-inch disks come from C. Itoh, IBM (but they are expensive), Kao, Memorex or TDK. Those not to buy: Dysan, SKC, Wabash, and Xidex. This is the final result from comparative tests made by Memcon that certifies that “there is no correlation between disk price and quality.“

Jerry Korth, Memcom president and CEO, said: “While Apple’s Macintosh and IBM’s PS/2 have made 3.5-inch diskettes very popular, quality simply hasn’t kept up with quantity. Our extensive independently conducted tests show there is strong reason to doubt that all name brand disks are tested to ANSI specifications, even though their manufacturers claim ‘100% certification’.”

“Only 4 of the 25 brands studied for the survey had all 100 tested disks pass both the ANSI missing-bit and extra-bit tests,” he added. “The most important result of the study was that it found a direct correlation between meeting ANSI minimum specifications and working in a PC environment. Last year, some manufacturers questioned the relevance of at least part of the ANSI standard. Today’s results should make clear to those cynics that the ANSI standard is very relevant indeed. A high percentage of diskettes that failed the PC functional test also failed the ANSI missing-bit or extra-bit tests.”

Among other highlights of the survey, it was found that:

Visual flaws, which can be taken as evidence of quality control (or the lack of it), varied tremendously from manufacturer to manufacturer. Five manufacturers – Dysan, Memory Media, Opus, Sentinel, and Xidex – had particular trouble in this area, with each having over 100 visual inconsistencies. The principal problems within this category were scratched shutters and contamination of the magnetic recording surface.
Only 5 out of the total of 25 tested brands had 100% of their disks pass the ANSI missing-bit test. They are C. Itoh, IBM, Kodak, Sony, and TDK. On the other hand, 15 of the brands exhibited 100% compliance with the ANSI extra-bit test. These brands are: Centech, C. Itoh, Fuji, IBM, Kao, Maxell, Memorex, Memory Media, Nashua, Polaroid, Sentinel, Sony, Syncom, TDK, and Verbatim.
All manufacturers did very well on the amplitude, modulation and resolution tests.
Overall, the manufacturers demonstrated great sophistication in their coating techniques. There were, however, two exceptions: disks from SKC and Wabash failed the ANSI wear resistance test at 1.5 million revolutions. This test is the best indicator of a disk’s useful life. Last year, all manufacturers passed this test.

The disks used in the tests were obtained at random from various distributors across the US. Ten 10-disk boxes of each brand were purchased at distributors’ quoted prices, with no attempt made to get a discount. Prices averaged $19.60 but varied over a very wide range: from as low as $13.00 a box (for JVC and Nashua) to $40.00 a box (for IBM).

Price variation within brands was also significant. A box of IBM diskettes, which cost $40.00 from one distributor, was purchased for $22.00 from another.

The study found no correlation between disk price and quality. In fact, it found that high-quality floppy disks could be purchased for less than the market average price of $19.60.

Of the 4 manufacturers whose disks exhibited 100% conformance to the ANSI missing-bit and extra-bit tests, three could be purchased at close to the average price. They were C. Itoh at $20.00 a box, Sony at $19.60, and TDK at $17.S0. (The fourth manufacturer was high-priced IBM.)

The 1988 report is based on the latest draft (N°9) of ANSI Standard X3.137-198x, to which Memcon added a PC functionality test and a visual examination.

The tester swallows over 500,000 disks a month

Memcon (Omaha, NE), who started this survey, has one of the five largest independent duplication business in the world. Its main activity is copying programs for software editors or converting 5.25-inch into 3.5-inch format.

“Independent” means that some software editors are not included, like Microsoft that makes duplication itself. Memcon says to use between 500,000 to one million diskettes a month.

The American company is also a manufacturer of duplication equipment and disk testing, which means that as it’s giving startling results on the quality of supports it could be taking advantage by selling a little more testers. This is the only argument that could make the truthfulness of this study questionable.

Some manufacturers think that the number of samples (100 of each brand) isn’t enough to represent a production that reaches hundreds and even thousands of disks per year, but this argument is less credible.

Jerry Korth explains the reason of this study:” In our business, we often had trouble with some brands. When we mentioned it to the manufacturers nothing happened, that’s how we decided to conduct tests and publish them.”

Their first publication, in 1987, concerned 5.25-inch floppy disks. The results published in some newspapers like USA Today or computer magazines like Byte, Computer World or PC Week, did disturb the American computer business.

Of course, losers like 3M reacted more than winners (BASF or Memorex). Memcom spotted 190 visual defects on the 1,800 5.25-inch diskettes that were tested.

On one of them, from Verbatim, they even found something looking like a piece of hair. One jacket had two magnetic disks (cookies)! and all this without mentioning large variations in prices: $8.40 to $23.90 for the same box of 5.25-inch disks from Dysan rather it was bought in Boca Raton or in Troy, UT.

What 3.5-inch diskettes did Memcon end up buying for its personal use?

“We use to buy from Kao, said Jerry Korth, but now we buy from Memorex. Itoh and TDK don’t produce enough for our demand, and IBM doesn’t self bulk diskettes, necessary in duplication.”

Test Descriptions

Overall: performances summed up, prices not taken in consideration.
Visual: each floppy disk was examined by nine types of defects which included: frayed or visible liners, cookie defects and contaminants, label inconsistencies, shutter scratches.
Torque: it is a measure of the rotational force encountered as the cookie spins in the disk shell.
Tension: shutter tension is a measure of force required 10 open the disk shutter as well as a measure of the force keeping it closed.
Missing Bit: it occurs when the amplitude (or signal level) of any flux transition (or bit) is less than 45% of the average signal level (sometime referred to as dipping level) of the entire track when written at 250,000 flux transitions per second (FTPS).
Extra Bit: It occurs whenever a pulse (which could erroneously be interpreted by the computer as a legitimate bit) is detected while reading an erased track. To be considered as an Extra Bit, this pulse must be greater in amplitude than 20% of the average signal level of the track when written at 250,000ftps.
Peak shift: the peak of each voltage pulse generated at the read head defines the position of each magnetic flux transition. If two adjacent flux transitions are relatively far apart, then the voltage signals (one positive and one negative) will not overlap. As the distance between flux transitions decreases, the two voltage signals begin to overlap and subtract from each other, causing a reduction in signal amplitude and a shift in the position of the voltage signal peaks.
Amplitude/Resolution/Modulation: amplitude and resolution are functions of coating formulation and thickness. They must be tightly controlled within ANSI standard limits to ensure that the floppy disks are compatible with the drive and associated electronics. Insufficient or excessive amplitude or resolution beyond the limits established in the ANSI standard may result in unreliable data recovery. Floppy disk amplitude and resolution must be maintained within these limits in order to ensure their compatibility with the disk drive. Modulation, which is a measure of amplitude variation on a given track must be maintained within ANSI standard limits to insure data recoverability and reliability in the demanding computer environment.
Overwrite: floppy disk overwrite characteristics are a function of coating formulation and thickness. Whenever a signal (data) is written over an existing signal (data) on a disk, the overwrite characteristics of that disk will determine the extent to which the new signal (data) will be discernable by the personal computer. A floppy disk with poor overwrite characteristics can lose data if it is written more than one time. Overwrite characteristics outside the ANSI specified limits may result in unreliable data recovery.
Wear: Wear resistance and abrasivity are a function of the coating formulation and media surface processing. In the manufacture of floppy disks, a delicate balance between wear resistance and abrasivity must be maintained. The disk coating must be hard and adherent enough to withstand prolonged and continuous use and yet unabrasive to the fragile core of the recording head. Wear resistance is measured by the%age change in amplitude of the sample floppy disk after 1.5, 3.0, and 5.0 million disk revolutions on the same track. Abrasivity is measured by the%age change in amplitude of a standardized drive after five million disk revolutions on the same track (00).
PC: PC functional tests were performed on an IBM PS/2 Model 50 personal computer. It is reasonable to expect that the new disks would be able to perform the task for which they’ve been designed – at least once. A disk which is unable to be formatted and recorded with data upon initial use is of no value to the user and should be discarded. For the purpose of this test, disks with bad sectors are considered to be format failures.

This article is an abstract of news published on the former paper version of Computer Datorage Newsletter on issue ≠11, volume ≠1, published on November 1988.