Shareware Super Platinum 8

home *** CD-ROM | disk | FTP | other *** search

/ Shareware Super Platinum 8 / Shareware Super Platinum 8.iso / mac / BACKUP / FLPKIT10.ZIP;1 / FLOPTOOL.DOC < prev

Wrap

Text File | 1994-07-19 | 18.1 KB | 346 lines

Floppy Tool Kit, Version 1.0 William Luitje 20 January 1994 This is a collection of 5 small utilities which will help you use your floppy disks better. Since you probably worry more about how to get better speed and storage out of your hard disk, this is not likely a matter of pressing concern to you. However, floppies are going to be used for quite a while and by spending a few minutes with these utilities you can make things go a little bit faster each time you use a floppy. FLOPPARM ~~~~~~~~ Floppies have been around a long time and have been improved since they first came out, not only in capacity but also in their mechanical speed. Unfortunately, system software can't tell what the best speed to drive a floppy is and must be compatible with old drives and so uses conservatively slow parameters. When accessing a floppy, the BIOS uses a table called the Disk Base Table for these parameters. This table can be found by looking at interrupt vector 0X1E. Since this table is usually set up in RAM by DOS, it is an easy thing to modify it for modern, fast drives. FLOPPARM will do this for you conveniently; simply run it with one of the three arguments SLOW, MEDIUM, or FAST. The parameters which are affected are: ˛ step rate and head unload time, which are parameters of the disk controller chip. ˛ head settle time, which is the amount of time taken after stepping from one track to another to wait for the head mechanism to stop vibrating. ˛ motor startup time, which is the amount of time taken to allow the spindle motor speed to stabilize after it first starts up. SLOW is provided for compatibility with really old drives; MEDIUM is somewhat faster than DOS 5; and FAST is pedal to metal! When run, FLOPPARM displays the parameters used. If run with no or an illegal argument, FLOPPARM will display a usage hint and the current values of the parameters. FLOPPARM also allows you to specify your own parameters by using the CUSTOM argument followed by an equals sign and the four parameters (step rate, head unload time, head settle time and motor startup time) separated by commas (no spaces allowed). For example a command equivalent to FLOPPARM MEDIUM would be flopparm custom=13,15,1,2 Legal values for step rate and head unload time are from 0 to 15. Legal values for head settle time (in milliseconds) and motor startup time (in 1/8s of a second) are from 0 to 255 Note that FLOPPARM is not a TSR and does not take up any DOS memory; it simply modifies a table which is already created by DOS. Once you decide which parameters to use, just put a line invoking FLOPPARM in your AUTOEXEC.BAT file and enjoy a modest increase in floppy speed. FLOPTEST ~~~~~~~~ How can you tell if you have set your floppy drive parameters too fast? I've included a utility which will seek to every track of the disk, read a sector and report any errors it finds. The seek pattern is called a "butterfly test" since it reads the first track, then the last track, then the second track, then the next to last track, etc. This will really give the drive a workout! This test requires a formatted floppy which has no bad sectors (FLOPTEST can't tell if an error is due to the drive not working properly or due to a bad disk). It is a read only test so it shouldn't change any data on the disk. There are two parameters: one is the drive letter and the second is the number of passes to run. If you get any errors and your floppy is ok then you should back off one or more of the floppy drive parameters and test again. Here is an example floptest a 20 If you are fortunate enough to have more than 2 disk drives (this applies both to FLOPTEST and to the other utilities in this package), you can refer to the third drive as c and the fourth drive as d, even if you have a hard drive that goes by either of those names. I know this is confusing but this test operates at the BIOS level and has no idea what letter DOS has assigned to your third and fourth floppies. I had a choice of confusing people who had more than 2 floppy drives or confusing everybody by making you specify the drive with a number rather than a letter. That's life. You can also use FLOPTEST to see if a floppy disk has any bad sectors. If you run it with only the drive letter specified, then it will attempt to read every sector and report any failures. Since FLOPTEST makes only one attempt to read each sector, it will warn you if a sector is getting weak. FLOPSKEW ~~~~~~~~ The most important improvement you can make in using your floppy drives is to format your disks using FDFORMAT by Christoph HochstÑtter or ATFMT by Oleg Kibirev. These programs allow almost complete control over how your disks are formatted. Among other things, you can put up to 820K on 5¨" DSDD disks, 1.48 meg on 5¨" HD disks and 1.7 meg on 3´" HD disks. In terms of speeding up floppy disk transfers, these programs allow you to control the way sectors are arranged on a track. This concept is called sector sliding by FDFORMAT and sector spinning by ATFMT; in the US it's known as sector skewing. Basically, the idea is that since there is some mechanical delay in changing heads or tracks it may be desirable to put the first sector of a new head or track "later" in the track so the computer doesn't have to wait for the floppy to make a complete revolution before that sector can be read. See these programs' documentation for more details on how this works. The time savings when reading or writing a floppy can be substantial when you format it using the optimal skew parameters and, as an added benefit, formatting is faster, too. You can specify head skew and track skew in both FDFORMAT and ATFMT by the X and Y parameters, respectively. The DOS FORMAT program does not support sector skewing at all; this is equivalent to using FDFORMAT or ATFMT with X and Y both set to zero. But what values should you use for these parameters? Well, when FDFORMAT first introduced this feature, I wrote a batch file which formatted a test disk with various combinations of the head and sector skew parameters and then measured how long it took to copy a large file to the floppy. It took about an hour to run and tested 16 combinations of X and Y values for just one disk format. FLOPSKEW does the same job in just a few seconds but tests even more combinations of parameter values. To use FLOPSKEW, first decide which setting of FLOPPARM to use and then run FLOPPARM with that parameter since the step rate and head settle time will both affect skewing. Then, format a disk of the type and capacity for which you want to find the optimal skewing parameters with the X and Y parameters both 0 and interleave 1. Although FLOPSKEW is a non-destructive test, you should use a freshly formatted disk to ensure that these parameters are set correctly, otherwise the test will give strange results. If you don't specify X and Y, FDFORMAT and ATFMT default to the correct values except that they both must use an interleave of 2 for the 5¨" 1.44 meg and 3´" 1.7 meg formats (and for this reason it doesn't make sense to use sector skewing with these formats). DOS FORMAT always formats with no sector skewing and interleave 1. Then, simply run FLOPSKEW with the drive letter of the floppy you want to test as the only parameter. Here is an example command and its results for a double density (720K) 3´" disk on my system: flopskew b Find Optimal Head & Track Skew, Ver 1.0 9 Sectors/Track, 80 Tracks, 2 Sides Head Skew Times (milliseconds): X: 0 1 2 3 4 5 6 7 8 Time: 32 54 74 95 116 137 158 178 199 Track Skew Times (milliseconds): Y: 0 1 2 3 4 5 6 7 8 Time: 232 53 74 95 116 137 158 178 199 Interleave: 1 Minimum skewing parameters: X=0, Y=1 Recommended skewing parameters: X=1, Y=2 First you are shown the format of the test diskette, then some test results and finally the recommended skewing parameters. If you are interested in some technical details about the test you should read the next few paragraphs; otherwise you can skip ahead to the next section. Just use the recommended parameters to format your disks and expect your diskette throughput to become about 30% faster. Two tests are run to find the best values for the Head Skew parameter (X) and the Track Skew parameter (Y). The head skew test is run by reading the last sector of the first track using head 0 and then measuring how long, in milliseconds, it takes to read the first sector of that same track with head 1. (If the disk is formatted with only one side this test is not run.) The track skew test is run by reading the last sector of the first track using head 1 (if testing a double sided disk) and then measuring how long it takes to read the first sector of the second track using head 0. In the results above, the second line tells how long this took, in milliseconds, for each of the sectors on the second track read. Let's look at the example data to see what we can learn. First, notice that the times to read different sectors in the head skew test increase by about 22 milliseconds. This is just the time it takes to read one sector for this format. The formula is sector read time(ms) = 1000 * 60sec/min * rpm / sectors For this format the drive rotates at 300 rpm and there are 9 sectors so the time to read a sector works out to 22.2 milliseconds. This formula can be modified to figure the time to read an entire track by dropping the final division by sectors. In this case, it is 200 ms. To determine the best values for X and Y, the program finds the smallest time and the corresponding number of sectors skipped from the line above. The data show that for my system no sectors should be skipped when changing heads so the optimal head skew parameter, X, is 0. Typical values are 0 or 1 for most systems. Now look at the results for the track skewing test; the time to change tracks without skipping any sectors is 232! This is the time it takes to read a whole track plus the sector plus some overhead. In this case the 0 sector skew provided by DOS format and the default values used by FDFORMAT and ATFMT yield a big waste of time when reading or writing multiple tracks. On my system the optimal track skew parameter, Y, is 1. Typically, it ranges from 1 to 3 on the systems I have looked at. Now notice that the program recommends values for the skew parameters which are 1 greater than the optimal values. This is because the penalty for choosing a skew value which is too small is severe: you have to wait a whole revolution, in this case 200 ms, when changing tracks. On the other hand, the penalty for having a skew value which is one too big is only 22 ms. Since some software may introduce an additional delay when changing tracks, it is a safe bet to be conservative when choosing the skew values. If you regularly use floppies on two computers you should run FLOPSKEW on both of them and choose the higher of the recommended skew parameters. Take the values of X and Y found by the above procedure and use them to format your floppies in the future. To make this easy, you can create a batch file, 4DOS or CED alias, or FDFORMAT or ATFMT configuration file with those parameters in it. Any time you format a floppy or read or write a floppy formatted this way it will go faster. How much faster? Well, assuming the most likely case that only track seek is a problem and that the sector skewing is only off by 1 or 2, then on average the computer has to wait for a full disk revolution every second track it reads. Eliminate the wait and you can reduce the time spent reading and writing by 1/3. When formatting a diskette, skewing only affects the verify operation so the time reduction is somewhat less, about 20%. FLOPRPM ~~~~~~~ If you are having trouble reading or writing diskettes, especially with a large number of sectors per track, the problem may lie with the speed at which the drive spins. FLOPRPM will test this condition for you. Simply place a formatted disk in the drive you want to test and run FLOPRPM with the drive letter as the only argument. The program will then test and display the rotation rate continuously until you press a key. Normally, all PC floppy drives will run within an RPM or two of 300 except for 5¨" HD drives which run at 360. If you are having problems and the speed of your drive is off by more than that, you may be able to get better operation by adjusting the speed of your drive. Unfortunately, no general guidelines can be given about how to do that because drives differ in their design. FLOPINFO ~~~~~~~~ This program simply reads the existing format of a disk and displays it. With all of the different formats supported by FDFORMAT and ATFMT it's nice to know what format a particular disk is. Also, since some of these formats require a TSR to read, you might want to know if a particular floppy can be read on another system which doesn't have a copy of that TSR available. Or, perhaps you want to quickly find out whether a floppy is HD or DD. FLOPINFO takes one parameter, the drive letter. Here is an example command and response, again for a 720K 3´" disk flopinfo b Floppy Disk Information, Ver. 1.0 by luitje@m-net.arbornet.org OEM Name: IBM 3.3, 512 Bytes/Sector, 9 Sectors/Track, 80 Tracks, 2 Sides 2 FATs, 1 Reserved Sectors, 112 Directory Entries Total Capacity is 730112 Registration ~~~~~~~~~~~~ While this software may be freely copied and used, I do request that you register it if you use it. Just send me either a postcard or some e-mail telling me who you are and where you are and what you like or dislike about the program. This is only partially an ego massage for the author. I enjoy writing useful utilities and have several other pieces of freeware out floating around the net so I would like feedback about which of them is worth expanding and supporting. This software is intended to be safe to use. All tests and information programs are read only. About the only problem you might have is that if you set the floppy drive parameters too fast you may get read or write errors. Never the less, in using this software you assume all risk should anything go wrong. I think that is pretty unlikely, but, hey, sometimes it rains fish! William Luitje 2677 Wayside Drive Ann Arbor, MI 48103 USA Internet: luitje@m-net.arbornet.org or luitje@m-net.ann-arbor.mi.us