cem@intelca.UUCP (Chuck McManis) (07/17/85)
> [whurr! clank! clikety-click!] > > I am confused about floppy disk drive specifications. > I see that the major choices are: > . single sided single/double density 48 TPI > . double sided single/double density 48 TPI > . double sided single/double density 96 TPI > full or half height. > > John J. Meshna Jr. (a surplus catalog) has the Mitsubishi > no. 4853 for $175. It is double sided, double density, 80 tracks > per side. Is this the same as a DSDD 96 TPI drive? > Yes. > The way I understand the specs: > single sided - read/write heads on one side only > You cannot flip over the disk to use the other side unless > the sector hole is in the center, or the sector hole is not > used (as in Apple computers). Partially, one set of heads, but you can flip the disk if you have a "notcher" which can cut a write enable notch in other side of the jacket. > double sided - read/write heads on both sides Correct. > density - depends on the external controller, thus any drive > that can handle double density can be used single density. > Does quad-density mean another doubling of the density per track > by the controller, as long as the drive can handle it? > I think density refers to the number of bits per track. Quad density means double tracks per inch (eg 96tpi) Other observations are correct about the drive, however some *disks* can only handle single density. > > How do they specify how many tracks the head moves (and is this > necessarily all the tracks on the diskette?)? I think the TPI rating > and the number of tracks per side are related as follows: > 2.25" outer track radius > - 1.4373" inner track radius > ------ > .8125" head movement edge to edge > so 48TPI means (48 tracks/inch) (.8125 inches) + 1 = 40 tracks > (a track is a discrete head position), thus the 80 tracks > means 96TPI. Completely correct. > > If a drive claims 160 tracks - is that 80 tracks by 2 sides > at 96 TPI, or 160 tracks per side at 192 TPI? Each side has a capability of 80 tracks. Note that 96tpi aka Quad Density aka 80 track drives require the better grade diskettes, sometimes referred to as HD (high density I suppose) Standard DD diskettes are not reliable especially on the lower end drives (see Mitusbishi (sp?) above) > Let us try to read the following specs: > mfg. model tracks no.heads > ==== ===== ====== ======== > a)Shugart 400 35 1 > b)Shugart 400 35 1 > c)BASF 6108 80 (both sides) 2 > d)Pertec FD 250 70 (80 optional)2 > e)MPI B92 160 (both sides)2 > f)CDC 9409T 160 (both sides)2 > g)Mitsubishi 4853 160 (both sides)2 > > Drives a and b could be 48 TPI with the head not traveling all the > way to the hub, or (35 tracks)/(.8125 inches) = 39.4 TPI. > Which is it? They don't travel all the way to the hub. It is possible to step "apple" drives beyond track 35 up to track 40 but it can confuse the home software the sends 35 step back pulses and assumes the drive is on track 0. Some protection schemes use this feature. > > Drive c is 80 tracks total, 40 per side thus 48TPI. > > Drive d - Does the stepper mechanism have two modes, > stepping .8125/(70-1) inches or .8125/(80-1) inches per track? > With 70 tracks, the TPI is 1/(space between tracks) = 85 TPI. > With 80 tracks, it is 1/(.8125/(80-1)) = 97 TPI. > What did I calculate wrong - the head traveling distance? > How - by changing the lead screw? > Again, if you're interface will use all 80 tracks then they are there. Interestingly enough 70 = 2 * 35 which would sound suspiciously like a double sided apple format. The Apple /// might use it. > Drive e, f, and g are 160 tracks total, 80 per side thus 96 TPI. > > In conclusion: > I think it is too easy to confuse tracks total with tracks per side > (total tracks = tracks per side times number of sides). > > The TPI rating says the distance between tracks, but not how many the > heads will reach. It should not be confused with density although > a 96 TPI disk will hold twice the data as a 48 TPI disk. I wonder > if some of the 'quad density' drives are really double density > at 96TPI as compared to single density at 48 TPI. Quad density got its name because a 96 TPI disk will hold 4 times the storage of a single density 48 TPI disk. The difference in the drive data sheets are labeled "track density" and "bits/inch" or "bits/track" There is a third type of drive made by DriveTec that holds 3.3 Mbytes on a minifloppy it is up around Hepto Density. > > The number of tracks the head can position itself is almost standard > (80 for 96 TPI, 40 for 48 TPI) but other possibilities exist. > Does this mean that the tracks for a 35 track and 40 track drive > won't align due to unequal spacing (thus different TPI ratings)? > The drives are the same, current drives can access 40 or 80 tracks per side (except the drivetec above). The controller may find it more convenient to limit this to n < 40 (such as apple). Other than the fact that the recording technique (GCR without Index) on apple disks is so bizarre albiet efficient, they would otherwise be readable on any minifloppy drive. > Density is the number of bits per track, determined by the > controller. The drive dictates how fast the data can be > read/written thus limiting the density which is otherwise > determined by an external controller. The data is recorded on the disks at 250Khz. The difference between double and single density is the data encoding technique. Single density is recorded using FM and double density as Modified FM (MFM). At the same frequency MFM yields twice as much data density as FM. These are not the only recording methods however, Intel uses MMFM (modified modified FM) on their 8" disks and as mentioned before apple uses GCR (Group Coded Recording). > One unit of measure is the number of bits per inch along > the track. Most controllers write the same number of bits on the > outer track as the inner track, thus the number of bits per track > is constant. This is changing since many manufacturers have found you can store more data on a disk if you write fewer bits on the inner tracks than on the outer tracks. Before the outermost track was under utilized to allow for the short length (and required recording density) of the inner track. Therefore, this currently the most common case but it is changing. > Since the circumference of the outer tracks is larger than the inner tracks > and assuming the density in bits per inch along the track being > constant, some controllers > therefore put more bits on the outer tracks so the density of > bits per track can vary from the outside to the inner track. True but still relatively rare, (see above) current controllers use write precompensation to make up for the smaller size of the inner tracks. > > Height has no effect on the drive's operation except that > some full height drives can lift the heads when the disk is > not in use and even shut off the motor. The half height drives > apparently don't have the room to do this and it makes for a > less expensive head mechanism at the expense of additional > disk wear. > Most minifloppy controllers turn off the motor when the disk is deselected. This prevents massive wear, and leave the heads loaded all of the time. I can't think of any 5-1/4" drive that loads and unloads the heads during operation and leaves the motor running but there are so many someone's may. This was a "standard" feature of 8" drives and it did make them somewhat faster. (It can take half a second to spin up a minifloppy whereas the heads could be loaded on an 8" drive in less than a tenth. > Speed of rotation is constant, isn't it? That would affect > the density. Are the speeds (in RPM) standard? Usually constant although some are not. All "standard" drives are and run at 300 or 360 RPM (its one of these and I have forgotten which one) Again some manufacturers have experimented with changing the rotational speed with respect to head position. > I suppose a disk written at one speed can be read at another speed if > the format is self-clocking and the controller is adaptive enough. > Is wow and flutter a consideration, or does the controller compensate > for speed variations. Even so, does this affect the reliability of > the drive (a more stable motor meaning less read errors?)? Since MFM and FM are self clocking there can be some variations from the disk speed without generating errors. The more difficult task is keeping the Phase Locked Loop on track if the speed changes by any more than +/- 10%. > > Density of the medium - the finer the magnetic particles in the coating > the more magnetic fluxes per inch can be stored. Is the density > uniform in all directions or only in a particular orientation > (such as along the tape's or disk's tracks)? Tapes are rated > in FCI (flux changes per inch, which determines BPI and TPI). > Why not disks? Diskettes are rated in FCI but you have to get those specs from the original manufacturer. I can't imagine a consumer would need to know this stuff, but if you do ask the manufacturers for it they will supply it. > > There are several mechanisms for moving the heads. One I have seen > moves the heads via a metal band directly to the motor pulley. > It is quite fast and quiet. Others use a lead screw, which > needs occasional cleaning and lubrication. Do any use linear > motors, or is that an unwarranted expense for a floppy drive? > This might appear as a spec on head seek time. Lead screws are inexpensive and bands are fast. You can always trade dollars for speed. The only voice coil floppies I have encountered were the PerSci 2xx series of 8" drives. They are/were extremely fast for floppies and unfortunately incredibly expensive. Not enough people bought them so I guess they were unwarranted. And yes it does influence the head step time. Even most band driven heads only step, the PerSci are one of the few that can truely seek. > > My Teac drives have a 'button type' head. I have seen DSDD > drives that have small heads on a leaf spring. Are there others? > Unless you get the manufacturer's specs, this is not usually > part of the drive specification. > There are as many head designs as there are flavors of Ice Cream. The "button" type are less expensive but more massive than the thin film or other types of heads. They also get munged a lot less when you insert and remove disks. Light heads are easier to move, hence a faster step time, heaver heads are more durable and thus yield a longer MTBF. Take your pick. > My fingers grow weary, so I conclude with a request for someone > knowledgable to offer a dissertation on disk drive specifications. > Thank you for reading all this, and in advance (for whatever). > And mine are even more tired. Maybe I should post this to net sources or something. :-). > > | Jeff 'I am sooooooooo confused' Skot | --Chuck -- "Unix, the Teco of Operating Systems." - - - D I S C L A I M E R - - - {ihnp4,fortune}!dual\ All opinions expressed herein are my {qantel,idi}-> !intelca!cem own and not those of my employer, my {ucbvax,hao}!hplabs/ friends, or my avocado plant. :-}