The most recent revision date of this page is Aug 12 2011. To order or to send email @ me follow this link.
I maintain some Web pages about technical features of floppy disk drives and media which includes statements about issues of use, repair, and standards. One such statement is about the use of High Density (HD) diskettes with Double Density (DD) diskette drives. You can "fake" a 1.2MB diskette to "look" like a 720K diskette, by covering the hole in the upper left corner on the 1.2MB diskette. The 720K diskette does not have that hole, the drive "reads" the corner to decide what kind of diskette is in the drive. This has become popular to do, because DD diskettes (or media) are not in production, while HD diskettes are still available.
This Web page has a detailed technical analysis from Chuck Guzis in 2009 at my request. There's also a quote from a statement by J. G. Harston. References are included. - Herb Johnson
In mid-July 2009, I wrote to Chuck Guzis of Sydex Inc., about the use of high-density (HD) 3.5" diskettes as double-density (DD). I have Web pages which discusses his previous work in the Z80/8080 computing world and Sydex today. Here's my correspondence with Chuck on the subject. My minor edits of Chuck's words are in 's. - Herb Johnson
From Herb Johnson: I'm writing today, because over time a problem has emerged, among those who are trying to use microcomputers of the 1970's and 80's with some of today's 3.5" floppy drives and available media. There are two general problems, but on one of them I'd like your professional opinions and experiences.
The primary reason I wrote you, is to get specifics from you about running high-density 3.5" diskettes (at 300RPM) on a 3.5" drive which is run by a controller writing and reading at double-density rates. One can tape over the "density" hole in the HD diskette, so the drive "thinks" it is working with a double-density diskette. But, my general impression is that writing DD on a HD media will produce a unreliable diskette.
I have a technical impression but not a definitive analysis, to offer to people who want to operate in this way. DD diskettes are hard to find today; the issue did not come up until that situation occurred. Part of the issue may be with data rates, which relate to density per inch as written, and/or to frequency compensation in the read/write circuits of the floppy drive. Part of the issue may be, with how strong the magnetic signal is, or the bias frequency used by the drive, or both - again, between DD and HD use. There may be some other issues.
The other problem I'll mention briefly. When replacing 5.25-inch floppy drives with 3.5-inch drives, beyond the problem of correct connections between drive and controller, is the matter of setting up the correct rotation speed of the 3.5" drive to 300RPM - trying to avoid running the 3.5" drive at 360RPM. That, and making sure the drive "knows" to operate at double density and not high density, usually with some jumper changes if available. That problem requires knowing what to do with specific brands and models of 3.5" drives. I don't expect you to provide details of that sort, I have some information about that on my site.
Do you have a "good" analysis of the DD/HD situation, in hand? or can you refer me to one elsewhere? I'm afraid on my site, all I have are bits and pieces of this "puzzle". For example, when formatting a DD diskette to HD, the "format" FAILS when the innermost tracks at highest density are written. This discourages use of DD diskettes at HD. But, when one formats in DD on HD disks, the format succeeds, and later the data writes succeed! This gives people some "confidence" they can use HD disks "as" DD diskettes.
Finally, do you have anecdotal information that these "DD written" HD disks fail in the long run? Or fail when transferred between drives of different models, brands, ages?
- Herb Johnson
[Note: Chuck refers to double density as "2D", I used the term "DD". He also uses the trade description "cookie", to describe the circular disk inside the diskette case. "Domains" are the magnetic elements inside each magnetic particle in the coating.- Herb]
From Chuck Guzis: Hi Herb. Years ago, I obtained some "standards" from the National Media Laboratory that's run by 3M (now Imation) about this very subject.
3.5" DS2D and DSHD media are very close in terms of the magnetic properties of the coating. DS2D media generally is coated with an oxide coating with a coercivity of about 660 Oersteds (Oe); DSHD media uses a coating with coercivity of 720 Oe, so it's only a bit "stiffer" magnetically speaking. DSHD also generally uses a smaller particle size, which isn't bad for DS2D use.
So far, so good--it looks as if the HD media will serve passably even if the write current in double-density mode isn't quite as high (lower-coercivity = lower write current needed to "flip" the domains). By the same token, a 2D diskette should work passably in an HD application (bring up any memories of the "puncher" accessories that used to be popular?). [These punches were sold for people to convert then-popular DD diskettes into "looking" like then-rare and expensive HD diskettes, with mixed results. - Herb]
Where things start to get interesting is in the coating thickness. DS2D coats the cookie with a layer about 1.9 microns (um) thick. By comparison the average fine human hair is about 25-30 um thick. (Hair diameter depends on the person and the type of hair, the range is pretty wide; from about 12 to over 100 um for a single hair diameter).
However, the problem arises that DSHD media uses a coating about half that thickness (0.9 um). This is why the cookie in a DSHD 3.5" diskette seems almost transparent if you hold it up to a strong light, while the DS2D variety is opaque. [The diskette's plastic material is transparent. - Herb]
Write current is directly related to the coercivity of the medium, not its thickness. The magnetic field strength must be strong enough to saturate the medium--and this is related to the write current. Try writing a DSED diskette (extended density, coercivity 1060 Oe) in a 720K drive and see if it works, in spite of the thinner than 2D coating.
There is an additional aspect that I did not refer to and that is "loading"--the ratio of magnetic particles to binder in the coating. Most modern 1.44M media in addition to having a thinner coating, also uses a lower "loading". Lower loading, as with thinner coatings make for better definition of domain boundaries, but also result in much lower read signal levels.
As a side note, a powerful influence [on the read signal] is the velocity of the medium [rotating] under the read head. The energy of the induced signal [into the read head] is proportionate to the square of the media velocity.
The thinner coating and the smaller particle size [of 3.5" DSHD diskettes] make for better-defined domain "edges" and generally better [resolved] higher-density recording. Unfortunately, [these factors] also make for a lower recovered data [read] signal, and that, coupled with the lower write current [when writing at] 2D, is what makes for Trouble, particularly on older drives.
As an aside, contrast this situation with 5.25" 2D and HD media; 290 Oe/2.5um for 2D and 660 Oe/1.3 um for 5.25" HD. If only we could take "1.2MB" media and stick it into a 3.5" shell, we'd be in Fat City. But this explains why the media incompatibility in the 5.25" world is so pronounced. By the way, VHS tape is remarkably similar to 3.5" DS2D media: 650 Oe/2um is fairly typical.
On a personal note, I think that the overall quality of DSHD 3.5" media isn't what it used to be, so that might contribute to the general impression that 3.5" HD diskettes used as 2D aren't reliable. I have problems enough finding reliable 3.5" DSHD floppies used as such. But then, we're not paying $40 a box for them either.
I've gotten plenty of conversion work where the 3.5" floppies were a mixture of HD and 2D media written in 2D drives ("720K" drives are blind to the media sense hole in a 3.5" jacket). Usually, they're just fine, with the error rate approximately the same, whether or not 2D or HD media was used.
Every once in awhile, I'll get a job where the same situation is applied to 5.25" floppies--2D (double density) data is written to a mixture of HD and 2D 5.25" diskettes. My heart usually ends up in my throat when this happens, but recovery is largely successful. My guess is because truly blank (degaussed) HD floppies were used, it didn't matter if a lower write current was used. Had the floppies been recycled from an HD application [without degaussing], I doubt that 2D writing would have been successful. [Note: degaussing is discussed in the "Solutions" section below.]
Herb wrote: When formatting a 2D 3.5" diskette as HD, it will FAIL EVERY TIME. It invariably fails when formatting the innermost tracks, where the data density, physical bits written per inch, is highest - and so it's the hardest test of the lesser-density media. Comments? - Herb
Chuck wrote: Your experience with 720K media used in a 1.44M drive shows the effect of the thicker coating and larger particle size. The inner tracks at the 500K data rate simply aren't capable of accurately registering closely-spaced domain boundaries.
One of several reasons why the inner tracks are problematical, is the velocity of the medium under the read head. [That velocity is slower, on the tracks closer to the center.] However, [the results were] good enough for many people, when a box of 3.5" DSHD floppies sold for $70. I recall going to the then-very-new Fry's and grabbing a box of Fuji DSHD 3.5" floppies because it was priced at a bargain $44.
Herb wrote:So what is your experience with HD 3.5" diskettes written at 2D? Looks like old stocks are available for about $1 each.
I'll volunteer that it works as a stopgap, but since we still have many 2D 3.5" diskettes left in storage, it hasn't been an issue for us and I can't say much about the long-term data retention issues.
It all depends on the eventual use of the diskette. If it's archival storage and it were my data, I'd definitely go for the right media, even at a dollar or more each. I'm disappointed enough with current new 3.5" HD media used in HD mode as it is.
[Herb notes: Chuck mentioned "degaussing". "Degaussing" is the process of using an AC electromagnet or "bulk eraser" to throughly remove all traces of earlier recording. Chuck is referencing the fact that a disk drive's "erase" during reformatting is not as complete as degaussing. The residual magnetic signal from prior use, amounts to an increased noise level when reading newly-written data, even on a "formatted" diskette. Much the same applies to video tapes and audio tapes, and the same degausser device for those can be used on diskettes.]
Chuck wrote:For some time, I've had an idea to take one of the 3.5" DSHD drives that can be jumpered for host-density-select (most are set for media-density-select) and tie the density-select pin to the Write Gate pin. Thus, writing will occur at a higher current, but reading will be as usual. I've not tried it, but it might improve the performance of HD media in 2D applications. I suppose that the first thing to do would be to try it on a 5.25" HD drive, recording 2D on HD media. One more project for the list!
As a side note, thickness of the coating, while it does have an effect on signal strength and "crowding", a far more powerful influence is the velocity of the medium under the read head. (One of several reasons why inner tracks are problematical). Since the energy of the induced signal is proportionate to the square of the media velocity, we've had good results recovering marginal floppies by simply increasing the speed of the spindle (and using a higher data clock rate to compensate). Why the industry abandoned the 600 RPM Sony/HP 3.5" model will ever be a mystery to me.
The good news [about adapting modern 3.5" drives to replace 8-inch floppy drives] is that a commonly-available drive, the Samsung SFD321B, can be jumpered for 1.6MB 360 RPM operation--and also provides an optional READY signal. See the Samsung SFD321B OEM manual at this Web link.
Of course, ultimately, what's needed is a good substitute for floppy drives. There are several simulators produced by hobbyists as well as at least one very expensive commercial drop-in replacement. Some of my older archived 5.25" floppies are suffering from oxide shedding after only 25 years--I won't even go into what a mess 1/2" tape has become.
In the meantime, we should all probably back up our important information in as many forms as possible, stone tablets preferred.
At any rate, you've got my thoughts. Feel free to employ them.
Charles P. (Chuck) Guzis
Sydex, Inc., Eugene, OR
In the course of researching the subject of DD/HD, I came across a document written by J. G. Harston and Web-published on his own Web domain, mdfs.net. He describes himself as someone with 30 years of experience with a range of microprocessors and software; his site is a collection of documentation and software. In a text file named "Densities" and titled "Disk Density Differences" and dated 31 Aug 2009, Harston states the following: quote
"What about storing DD data on a HD disk? The grains are small enough for the data, in fact smaller than they need to be. However, DD recording strength is twice HD recording strength. Recording DD data on a HD disk will force the data into the grains, like using a pencil with too much pressure. It will be difficult for the recording head to erase old data to write new data over it. The data on the disk will deteriorate every time something is written to it." - end quote
"Recording strength" means the strength of the magnetic field from the head, into the media. That depends on the strength of the recording current, the distance from the recording head to the media, and the features of the head design. "Smaller grains" refers to the increased density Chuck discussed. Broadly speaking, Harston questions how well a DD drive can operate a HD diskette over repeated uses. To verify these considerations, I'd have to find specs for how DD vs HD drives write and erase diskettes. But this affirms one idea, of using a bulk eraser before formatting a diskette, to throughly remove "old data" - Herb Johnson, Aug 2011.
Copyright © 2011 Herb Johnson and Chuck Guzis