CP/M Frequently Asked Questions

oak.oakland.edu (unfortunately down - February, 2001)

ftp.mayn.de

wuarchive.wustl.edu

ftp.update.uu.se

ftp.demon.co.uk

reze-2.rz.rwth-aachen.de

soltrans.cr.usgs.gov

MIX C and other MIX products are available from:

The CPU280 (uses a Z280, an IDE interface is available), also written up in TCJ, issues #52 and #53. Circuit boards are available from The Computer Journal. CPU280 uses the ECB-bus which allows many other I/O cards to be connected.

Ampro LittleBoard products are no longer available from Dean Davidge nor are the SB180/SB180FX from Micromint.

Another CP/M machine is the PalmTech CPUZ180, designed and built in Australia. The complete SBC fits on a 6"x4" and runs at 18MHz. Included are floppy and IDE hard disk controllers, color/monichrome video controller, IBM PC/XT keyboard interface, printer parallel port, two serial ports, real time clock, 1 Meg ram, amd many other features.
It may be ordered from:

If you have a Kaypro equipped with an Advent TurboROM, Plu*Perfect Systems offers a program called MULTICPY that can read/write about one hundred different 5 1/4 formats.

The simplest way of converting *CP/M formats to a PC* is to use a PC with 22DISK - just copy the files from one CP/M disk to DOS, and then back to the other CP/M disk. (See Q12.) But a few older CP/M disks have what are called "hard sectors". These disks use several physical holes in the disk to mark divisions of data, instead of ONE hole which is used as a timing reference. These disks can only be read by a PC or a CP/M system with suitable hardware. The problem is NOT the diskette drive, but the controller cabled to the drive: the drives are unmodified, it's all in the diskette controller. CP/M hard-sectored disks come from some older Vector Graphics, Heath/Zenith H89, NorthStar, IMSAI and other CP/M systems.

Similarily, it is not possible to directly read/write Apple II CP/M disks on any other host machine because an Apple disk is recorded in GCR which is incompatible with FM/MFM *floppy* disk controllers. The only way to get files out of either kind of these disks is via a serial link with the original host system, or with special hardware on the PC compatible. (See Q12.)

An example of PC hardware is a MicroSolutions device called the MatchPoint PC. When used in conjunction with a MicroSolutions CompatiCard, files can be read from an Apple CP/M disk and transfer to another disk format with a special configuration of UniForm. The CompatiCard is also able to directly read some hard-sectored disk formats.

If your are *lucky* enough to have a *CP/M* B/P BIOS, it comes with a built-in disk format emulation capability, and a library of formats, including the source so that new formats may be added.

The interface on 8" drives and 5 1/4" drives are essentially the same. The 34 lines on a typical 5 1/4" controller are sufficient to control most 8" disk drives using soft-sectored disks. Here, is a diagram for a basic conversion cable to allow connection of an 8" drive to an IBM-compatible, AT-style (high density) controller.

                                                      8" disk drive
   PC-AT style controller                       Based on Shugart SA-851

   Grnd. Sig.  Sig. Name                             Sig Name  Sig Grnd
     1     2   Double/High Density ->>
                                    >>- Write Current Switch/    2    1
                                   Active Read Compensation
                                   User Customizable I/O pins    4    3
                                    "         "       "    "     6    5
    33    34 **Ready ---------------<<------------ True Ready    8    7
                                    <<-------------#Two Sided   10    9
    33    34 **Disk Change ---------<<----------- Disk Change   12   11
    31    32   Side 1 Select ------->>-----------#Side Select   14   13
     3     4   In Use/Open --------->>---------------- In Use   16   15
    15    16  *Motor On ------------>>------------- Head Load   18   17
     7     8   Index ---------------<<----------------- Index   20   19
    33    34 **Ready ---------------<<----------------- Ready   22   21
                                    <<---------------##Sector   24   23
     9    10   Drive Select 0 ------>>-------- Drive Select 1   26   25
    11    12   Drive Select 1 ------>>-------- Drive Select 2   28   27
    13    14   Drive Select 2 ------>>-------- Drive Select 3   30   29
     5     6   Drive Select 3 ------>>-------- Drive Select 4   32   31
    17    18   Direction Select ---->>------ Direction Select   34   33
    19    20   Step ---------------->>------------------ Step   36   35
    21    22   Write Data ---------->>------------ Write Data   38   37
    23    24   Write Gate ---------->>------------ Write Gate   40   39
    25    26   Track 00 ------------<<-------------- Track 00   42   41
    27    28   Write Protect -------<<--------- Write Protect   44   43
    29    30   Read Data -----------<<------------- Read Data   46   45
                                    <<------##Separation Data   48   47
                                    <<-----##Separation Clock   50   49

This diagram also works in the other direction--that is, to attach high-density 5 1/4" drives to an 8" controller.

* - It seems to be a logical substitution since the vast majority of 8" drives have continuously running spindles and instead of MOTOR ON require a HEAD LOAD signal. Also, a controller sends MOTOR ON before a DRIVE SELECT.

**- Most 5 1/4" disk drives do not provide a READY signal but send a DISK CHANGE signal on line 34 of the interface. An 8" drive has provisions for both signals. Likewise, most AT-style controllers expect a DISK CHANGE signal on line 34, so lines 33 and 34 should be connected to lines 11 and 12 of the 8" disk connector. Also, some 8" drives provide a TRUE_READY signal which is more useful than the standard READY.

# - Unused on single sided drives (SA-800/801).

##- Used only on hard-sector configured drives (SA-801/851).

Some 5 1/4" disk drives have the option of providing _either_ DISK CHANGE _or_ READY on line 34 (in particular, the TEAC FD55R series). Some 8" disk controllers do not care about the DISK CHANGE signal, but must have the READY signal. If you are attaching a high-density 5 1/4" drive to an 8" controller, you may get away with making the drive always ready by shorting lines 21 and 22, but this may cause a few re-tries when switching sides. If your drive offers a READY signal that your controller can deal with, by all means use it.

The MOTOR ON/HEADLOAD dilemma may also have an alternate solution if you are connecting 5 1/4" drives to an 8" controller. Some 5 1/4" drives permit motor turn-on by means other than the MOTOR ON signal. For example, the TEAC FD55R series of drives may be configured to turn the motor on based on the state of the IN USE light. The IN USE light can, in turn, be set to turn on only on drive select. Thus selecting the drive automatically turns on the motor and neither a MOTOR ON or IN USE signal need be present.

Another way to handle 8 inch drives on a PC is with a Microsolutions Compaticard IV, if you can find one. (MicroSolutions no longer offers this product.) It has the necessary software support to properly handle 8 inch drives, and in both SSSD and DSDD. This controller can be set up as both a primary controller, or as a secondary. It can support 4 drives, of any type, including 2.8 meg. It supports two MSDOS 8 inch formats, SSSD (about 250k) and DSDD (1.2 meg). It works perfectly with 22disk, and can read and write almost any 8 inch CP/M format.

Both 3.5" and 5.25" drives have the same 34 pin interface. 3.5" disks spin at 300 RPM thus the 250k/500k data rates. 5.25" disks spin at 300 RPM for all but the 1.2 Meg capacity, which is 360 RPM, thus the ratios:

        15 sectors per track / 18 sectors per track
                   = 300 RPM / 360 RPM
                   = 1.2 meg / 1.44 Meg

Most old systems didn't use pin 2, 34. That's GOOD NEWS since modern 3.5" floppy drives place signals there that the old controllers can't handle. The ready/disk changed lines changed from the "XT" generation drives to the "AT" generation drives. Older floppy drives had jumpers for drive select 0-3 and where to place the status signals The "AT" floppy drives assume the "AT" signals and usually allow only setting the middle 2 drive selects, thus the cable twist nonsense. for completeness, here are the pinouts:

        Mini/Micro Floppy Interface

        Pin#  Description         Alternate Functions
        ----  -----------         -------------------
        1     GND                 Eject, Disk Change Reset
        3-33  Odd pins are GND
        2     High Density
        4     Head Load           In Use, Eject
        6     Drive Select 3
        8     Index Pulse +
        10    Drive Select 0      Motor On A    \  IBM twisted
        12    Drive Select 1      Drive Select B \ cable - both
        14    Drive Select 2      Drive Select A / drives are
        16    Motor On            Motor On B    /  strapped DS1
        18    Direction
        20    Step
        22    Write Data
        24    Write Enable
        26    Track Zero +
        28    Write Protect +
        30    Read Data +
        32    Select Head
        34    Disk Changed +      Ready +

              + signal from drive to controller

The following table is extracted from the CompatiCard manual:

           Card        34   37              50   8 Inch Drive
     Signal Name       Pin  Pin  Direction  Pin  Signal Name
     ============================================================
     Programmable      2    3       --->    2    Low Current
     Index             8    6       <---    20   Index
     Drive Select 1/3  12   8       --->    28   Drive Select 2
     Motor Enable 1/3  16   10      --->    18   Head Load
     Step Direction    18   11      --->    34   Direction Select
     Step Pulse        20   12      --->    36   Step
     Write Data        22   13      --->    38   Write Data
     Write Enable      24   14      --->    40   Write Gate
     Track 0           26   15      <---    42   Track 0
     Write Protect     28   16      <---    44   Write Protect
     Read Data         30   17      <---    46   Read Data
     Select Head 1     32   18      --->    14   Side Select

   The odd pins of 34 pin connector to odds of 50 pin connector.
   Pins 21/37 of the DB-37 go to the odd pins on 50 pin connector.

MDM740 - The last of the "MDMxxx" programs.

IMP245 - This is nice, and works smoothly within what it does.
What it does, it does very well. IF you have slow floppy drives, there is a patch to cut down the receive buffer size.

MEX114 - different from the above two, but minimally functional with just a MDM740 overlay. To use all of its fine features, you need MEX overlay for your machine.

ZMP15 - This program includes ZMODEM file transfers.

KERMIT - This program may have the widest implementation base because it uses only printable characters for its file transfers. This is a plus because the MODEM7 family of protocols send binary characters that sometimes conflict with the underlying system use. It is a minus because many more characters must be sent and thus is slower. KERMIT may be found on watsun.cc.columbia.edu.

QTERM43F - This is somewhat like using QMODEM on an MSDOS machine. Qterm has VT100 emulation mode as well as XMODEM and KERMIT protocol. If you can get (or write) a good overlay, this is a nice program. (Bug fixes to 43E were released in a separate library to bring it up to 43F. The FIX library did not include a new binary; users had to do their own patching.)

For high speed transfers, you will probably need interrupt-driven routines, which are available for some these. The exact baud rate where it becomes necessary varies by system and program.

Other popular library maintenance programs are LUE, DELBR, and NULU, the latter being one of the best CP/M programs for handling LBRs. However, don't use NULU to extract and unsqueeze simultaneously. It occasionally screws up doing this, and it can trash an entire disk when it does so.

LT31 is also able to unpack libraries and also supports all current compression standards (including LZH 2.0!). It is a very useful utility and can replace several single programs.

Several things can be done to help the situation. If your Kaypro came with the MITE software package, you can use it for high speed terminal emulation. A Kaypro 2X using MITE can go as fast as 19200 bps. MITE uses interrupts to achieve this.

Sometimes the problem can be ignored. A 2X will drop characters at 300 baud using Kermit-80. File transfers work fine at 19200 bps. It is always a good ides to run file transfers in the quiet mode if terminal mode is dropping characters as then the display update time is minimized.

The graphic-equipped Kaypros can be significantly improved in terminal mode just by turning off the status line at the bottom of the screen. As most terminal programs have an initialize sequence available, just send the no status line command to the Kaypro - <ESC>, C, 7 [1BH, 43H, 37H in hex].

There are several hardware changes that can lessen or eliminate the problem. There is a speed modification for the 1983 Kaypro-II's & IV's requiring changing some chips to faster versions and outfitting the back with a toggle switch. Upgrading to a MicroCornucopia MAX-8 or Advent TurboROM also helps.

If your machine is equipped with the Advent TurboROM and you choose to run QTERM, Don Kirkpatrick can send you an interrupt driver that allows the graphic-enhanced Kaypros to work just fine to at least 2400 baud.

Tilmann Reh, an engineer in Germany, has designed an IDE hard drive interface that plugs into a Z-80 socket, and described it in The Computer Journal magazine as the Generic IDE (GIDE). He has produced a number of kits that include the circuit board, parts, and even a time of day clock chip. Several people have bought these (as of Jan 1996) and are beginning to write software to support these on various Z-80 based computers (including ADAM and TRS-80 as well as CP/M based systems).

Europeans can contact Tilmann Reh directly. In the USA, Tilmann may refer you to a US distributor. The current US distributor is The Computer Journal which has a GIDE Web page (See Q23.)

In general, we cover software and hardware that one person can work with, where you can "do it by yourself". This includes common programming languages and boards and systems where you can identify (and get) the parts and get code to make it work. Source code from the articles is posted on the TCJ Web pages and BBS so you can download it instead of typing it in.

The subscription rate is $24 for 6 issues or $44 for 12. Subscriptions may be sent to:

subscribe list-tcj <your@email.address> end

Early 8086 machines (CP/M-86 or non-IBM Sharp MS-DOS programs): MZ-5500, MZ-5600, 'SHARPWRITER', PC-5000 'Bubble' machine.

The SUC can supply hardware upgrades and documentation for many of the above machines. Contact :

ZCPR3 gives you UNIX-like flexibility. Features implemented include shells, aliases, I/O redirection, flow control, named directories, search paths, custom menus, passwords, on line help, and greater command flexibility. ZCPR3 can be found on many BBS and SIMTEL mirrors. The Z System commercial version is available for a nominal fee from The Computer Journal. Further details can be found in the text "ZCPR3, The Manual", by Richard Conn, ISBN 0-918432-59-6.

You can find a detailed history of the development of ZCPR and the Z System in Jay Sage's column in issue #54 of The Computer Journal. This article celebrated the 10th anniversary of ZCPR, which was first released on February 2, 1982. His "ZCPR33 User's Guide" also has a section on the history.

There still are active Z-nodes supporting Z-system and many RCP/M's supporting CP/M as well as some special interests. As of November 7, 1995, the known BBS's supporting the Z-System are:

The Z180 was not an outgrowth of the Z800. It was a joint effort between Zilog and Hitachi. The first two versions of the HD64180 were slightly different from the current Z180. The current HD64180 and Z180 are identical, and both have flags in one of the control registers to emulate the earlier versions. The changes are mostly bus timing, as the HD64180 was designed to interface with Motorola 6800 style peripherals as well as Intel and Zilog, which wasn't too strange since Hitachi second sources some Motorola 6800 series products.

While not the first microcomputer or microcontroller to be offered for public sale, the Altair 8800 is often cited as the "first personal computer" as it was a widely accepted and visually recognized product; it recieved a lot of press coverage inside and outside the electronics industry; and it set a manufacturing standard for a new industry. It and its successors were certainly early yet enduring leaders in affordable personal, business, and industrial computers. Only the IMSAI 8080 compares in recognition value among hobbyists, but the Altair is often cited by the popular press.

At first, MITS (and almost immediately others) produced cards which were compatible to the Altair bus. Soon, IMSAI and others followed with the production of competitive yet (somewhat) compatible systems. The S-100 bus evolved as other manufacturers, notibly IMSAI, made slight changes to the Altair bus signals and improved the front panel. Yet other manufacturers used digital designs that either depended on special signals from their own cards, or had signal timing requirements that varied between manufacturers. Over time, these differences and the limits of the original Altair/IMSAI produced a number of manufacturer-specific bus variations for extended addressing, bus operations, memory refresh and so on. MITS, IMSAI, Cromenco, Compupro, Ithica Intersystems and Northstar were among the major S-100 systems manufacturers of the time. Card manufacturers are too numerous to list. Most S-100 systems used the 8080, Z80, or 8085 processors, but some companies produced cards with almost any available 8 or 16-bit processor.

Bus signal differences were finally addressed in 1983 with the publication of the IEEE-696 standard by the Institute of Electrical and Electronic Engineers. The standard was previously in use primarily by Compupro and Ithica. As CP/M personal systems went to single-board designs with no bus at all, the introduction of new S-100 designs peaked. Further competition, price pressures, and finally the IBM PC caused new S-100 system designs for business and personal use to drop in the mid-1980's. A notible system of the era was the Heath\Zenith Z-100, a dual processor 8085/8088 system that could run CP/M 80, CP/M 86 and MS-DOS: and *very* similar to the popular Compupro 8\16 system. Zenith sold thousands of Z-100's to the military. Incidently, many systems of the mid-1980s began to run other operating systems, such as CP/M-compatible Turbodos, ZCPR and Z-system; and various UNIX-compatible OS's on 68000's, 80286's, and other processors.

New IEEE-696 systems were subsequently developed through the end of the 1980's, primarily for industrial and/or development (non-CP/M) applications, particularly where multiprocessing or speed were important. Up to at least 1993, Compupro and Cromemco still supported these systems at commercial prices, but apparently they did not support their prior CP/M systems except as cards and documentation for sale. New S-100 cards were also introduced throughout the 1980's, but declining through the end of the decade. Zenith's Z-100 system is supported by some active user groups and on-line maillists such as Usenet's comp.sys.zenith.z100. Northstar systems owners correspond occasionally on comp.sys.northstar.

One person who provides S-100 cards, documention, and some support is Herb Johnson. As "Dr. S-100" he wrote (1994-96) a regular column in The Computer Journal (www.psyber.com/~tcj/) and corresponds with S-100 and IEEE-696 owners. As of 1996 he can be reached via The Computer Journal or:

The Circuit Cellar BBS is on-line 24 hours per day with some cross development tools, particularly for CPU's that are commonly used as controllers. They have a Courier HST running 2400/9600 bps at 203/871-0549, and another line that will do up to 14.4k bps (8N1) at 203/871-1988. Both of these numbers are in Connecticut.

The Motorola BBS is in Austin, Texas, on 512/440-3733. They have downloadable cross development products mostly for the 68xx and 68xxx architectures. Like the Circuit Cellar BBS, this BBS seems to specialize in micro-controller development. Many of these files can also be accessed over the network on bode.ee.ualberta.ca (129.128.16.96).

2500AD software lists a Z80 assembler, a Z80 C compiler (that includes the assembler in the package), a Z280 assembler, a Z280 C compiler (that includes the assembler), and a Z380 assembler.

Don't forget to look in the old familiar places, such as oak.oakland.edu and wuarchive.wustl.edu.

The Walnut Creek CDROM has some tools from some of the sources listed above on the CP/M CDROM.