I haven't been active for about eight months due to work and other commitments, but it's time to get back to work.

The first project was to get rid of all of the stupid posts on this site advertising various unsavory stuff. The good and bad of the internet I guess. So unfortunately I've had to turn off visitor entries. I just returned from a trip to Florida to pick up two systems that look to be in pretty good shape, with lots of diskettes. Also looks like the original owner was a dealer of some sort with lots of copies of brochures and yes even a COLOR "Meet the digital group" bytemaster brochure! It also looks like the OS of choice was OPUS, so there is a bit of documentation as well as a set of documentation on SOS (Single user Operating System) which I've never heard of before. More as I begin to unpack...

A special thanks to Richard Brandle for providing the missing Bridge newsletters. Richard is also working on a cool Z80 Emulator to run DG software! More info in future posts!

Sitting busily cataloging diskettes from a recent find of DG related stuff, I was working at the Bytesmaster that I luckily acquired directly from Dr. Suding about three years ago when I heard the unmistakable sound of electronics frying..... I immediately (within a couple of seconds) shutdown the system (pressed the red "Power" button) but not before the "magic blue smoke" (I had an instructor at MIT who claimed that the reason that all electronic components worked was that "magic blue smoke" was captured inside each component during manufacturing and that the component would work as advertised unless you caused the blue smoke to leak out thus ruining the component.....) appeared at the back of the 8" diskette drive cabinet.

I took the cabinet apart, disconnected from the CPU cabinet and used the nose test to locate the area leaking blue smoke. Not surprisingly, the power supply. After further disassembly I located the tantalum capacitor that was indeed cooked (deflated from loss of magic blue smoke). While I'd read about many problems with DG power supplies and modification to make them less prone to fry other things should they have trouble, I've never had a power supply go bad, which got me to thinking and reading from the old newsletters. Several references were found to ruining diskette drive electronics from power supply failures, not good. I repaired and tested the power supply, put everything back together and all was well, at least with the computer......

I began thinking about the possibility of destroying one (or more) diskettes due to a bad power supply, or a CPU board, or, or, or. I also recently had a Phi-Deck ruin a couple of tapes that I had put together a couple of years ago due to a Phi-Deck board failure and had a 1702A ROM fail (never seen that before). I've decided that it's time to get serious about preserving all of this technology on more modern media, which sounds great until you try and start at the beginning with ROM images. Yes I know, I can simply write the ROM images to audio tape (which I have done), but have you tried to buy high quality cassette tapes lately? Besides, I want to store everything in files that I can publish on the web-site (other than in MP3) audio files. The easiest way would probably be to write a simple program to read the ROM data and transfer it to a PC (actually I use a MAC most of the time) via serial port. The other part of preservation is being able to restore things like a ROM image. I do not have an EPROM programmer able to write 1702A's (not many of those around since it takes a -48V pulse on each data pin to program!!!! I'm only aware of one commercial programmer that is $1,000+) so I'll also build an EPROM programmer if anyone out there needs any 1702s programmed.....

I'll keep you up to date on progress. I've started with a PIC based ROM reader that will likely become a ROM emulator at some point. Building a programmer is more problematic with the odd power requirements.

One of the complications of working with Digital Group systems was the different ROMS that could be used to startup the system (Audio, Phimon, Diskmon, OASIS, MCSO). Later in the chronology, if you had a Disk system, you could run a program called DDBOOT after booting into Diskmon to allow you to boot into the other operating systems. One solution was the ROM Switcher detailed in Bridge Vol.1 No.3 from August 1978. Another option was the UnROM outlined by Dr. Suding in this short article. This application was used to change ROMs on the TVC card (video interface). I'm certain that there are several articles about using the same method for the 1702a on the CPU card. I'll update this post and article when I locate the details.

"The UnROM"

by Dr: Robert Sudlng

Theory:

The 1702A EROM on the epu card may be remotely disabled by putting a low level signal on memory line Y. This circuit will Initially place a high level signal when turned on. An output of anything to Port 377 will disable the EROM: An output to Port 376 will return to the EROM.

Similarly, the 6S71L maybe disabled and replaced by some alternative character generator pattern. In this case, the 6571L Is removed from the D.G. TVC, Inserted on the supplemental "UnROM" board and the "UnROM's" cable is plugged Into the former 6571L on the TVC. Initial power on will enable the normal 657lL. An output of anything to Port 375 will disable the 6571L, and enable one of the 2708 EROM'S. Since the 6571L has an effective 16K bits of ROM, two 2708's are required, the first one handling the 1st 64 characters, and the second holding the last 64 characters. An output to Port 37. will enable the 6571L.

(Click on schematic at right for full size version)

The card may be inserted into any 22 pin I/O connector.

Parts:

1 - 7400
1 - 7410
1 - 7430
1 - 7442
2 - 74125
2 - 2708 (programmed with upper and lower)
1 - 6571L (removed from TVC)
1 - 22K 1/4 W resistor
3 - :01 disc capacitor
3 - 1 mfd tantalum capacitor
1 - 22 mfd tantalum capacitor
5 - 14 pin IC sockets
1 - 16 pin IC socket
3 - 24 pin IC sockets
1 - Dual 22 pin socket card
1 - Approx: 18" ribbon cable with 24 pin plug


Operation:

After correctly wiring card, remove 6571L from TVC and Insert on card. Plug Into any available I/O 22 pin connector. Insert the 24 pin plug Into the former 6571L socket on the TVC: Turn on the system. To UnROM, load a RAM version of the desired EROM; Out 377; Move the RAM version to address 000000ff; Proceed with the program. Similarly output to Port 375 to enable the APL character set ( or other desired set). Output to Port 374 to return to the standard ASCII.

Following is a summary of a two hour questlon and answer period with Dr. Robert Suding. Included here are a number of unpublished stories, fixes and comments of interest to Digital Group system owners. For ease of use, the material has been broken down by category. Those interested in hearing a tape of the entire session should contact Rich Kurtz.

Hardware Fixes and Notes

If you are having trouble with your TV-Cassette board., be sure to check the +5 volt supply at the board...it should not exceed 5.05 volts. Higher than this will cause a multitude of welrd problems. The 5 volt supply has a pot to adjust the output voltage. Bringing the voltage down will also make the system less noise prone. Try to shoot for 4.75 volts at the farthest memory board for the best results.

Having read errors on your Phi-Deck system? ... This can be cured by running a wire from the motherboard ground plane to the nearest chassis screw. This will also eliminate many other noise problems. What happens is that the Phi-Deck head lines pick up as much as .8 volts of video noise. Installing the ground eliminates 9O% of this noise. This also will eliminate hum problems in the Votrax system.

There are now modifications to the floppy dlsk board to allow its use with dynamic RAM. This is also what has been holdlng up the disk software. It is being made compatible so that the disk will not cause wait states which will loose the data in the dynamic chips. This means that OASIS, OPUS, Disk-Mon, MINI-MON, etc. will be released in the near future with all of these modifications.

There are design flaws in the Z-80 chips themselves... when you press the reset button on the computer, the z-80 goes to invalid TTL states once in a while in the memory request line. This will mainly cause problems with dynamic memory, causing your software to self-destruct ...to reset with dynamics, it is neceseary to modify the CPU card to reset using the non-maskable interrupt, thus avoiding the problem with the hardware reset. To modify the board...move the 220 Ohm resistor to the right of the CPU chip (R6) to the holes below the Z-80 chip. You are now ready for dynamics. This will have no affect on the operation of static memory except to allow plug in compatibility with all boards.

The video graphic board being demonstrated has 128 x 128 resolution with 16 grey levels. It can be built on a proto card for about $20.00 in parts. The schematic is available from Dr. Suding at no charge or see Walt Chalberg at a meeting for a copy.

RF problems either entering or leaving the system?... 1) Is the cover on? 2) Bypass the power cord. 3) Shield the fan and ground the body of the fan to the chassis. 4) Check your cables for radiation. The keyboard cable can be bypassed with a .01 mfd capacitor between the Strobe and +5 and another between +5 and Ground at the paddle card connector. If you look It the D.G. cabinet, you will flnd a whole bunch of tiny holes around the front and back panels. Those are meant to mount bronze "fingers" to make contact with the top and bottom to make it really RF tight. Of course, the vent holes will also have to be screened. The major radiation, however, will be from the power cord. Be sure to install those .01 mfd capacitors.

Misc. Information

On a new CPU Card:

"... The one that looks the most promising is the 6809. The 6809 by Motorola..,is not compatible with the 6800. It internally works like a 16 bit processor. It has hardware multiply capability somewhat similar to the TI but will support a lot of memory..."

On a 16 Bit CPU Card:

"....My plan tentatively is to design a new motherboard that has 2 CPU slots. You can either plug in your 16 blt processor into that slot or your 8 bit processor into that slot, all the other connections would be made for you and that's all there is to It. So you would exchange your motherboard (Which represents an investment of 40 or 50 dollars) now you can run 16 bit or 8 bit world. That's the plan of attack."

On New Products:

"Those of you who have the HAM board...you have a diagnostic in there that allows the system to self-diagnose (problems). ...Not only will you have the hardware perfect, you will have all the support software and you will have self diagnostic software (on all future products). What we are seeing is a vastly superior support in software for products being delivered. This means fewer products being introduced."

In bringing up a brand new out of the NEVER OPENED BOX phi-deck drive I was looking through some old tapes and ran across a copy of PHIDOC.TX which contains the updated documentation for PHIMON. This version "communicates" in hexidecimal instead of octal as well as supporting updated commands.

Enjoy


<PHIMON-H>: Hexidecimal PHIMON Ver 3.0
An addendum to the PHIMON OPERATING SYSTEM Manual

P.J. Ferrell K7PF


I. INTRODUCTION

<PHIMON> is a superior tape-based Operating System, exceeding the capability of man y disk-based systems. The modifications described herein are only extensions, and retain the design philosophy so ably expressed by David Bryant, the originator of <PHIMON>.

Some internal addresses have been changed due to 8080 code being replaced by more concise Z-80 instructions. Where possible, however, compatibility with previous versions was maintained. With the exception of HEX communication instead of Split-Octal, the user will notice little in the way of difference between <PHIMON> and <PHIMON-H>.


II. COMMAND EXTENSIONS

1) The DAte command replaces the HElp command.

2) The LOad command allows other than .GO extension, and permits optional starting address override using "@":

Format: >LO#n NAME.EX @addr

where addr is a two-byte hexidecimal address used instead of FPAGE (normally obtained from the header block).

3) SAve command allows other than .GO extension.

Format: >SA#n NAME.EX xx-yy*sa

Note that if .EX is not specified, a .GO default is assumed for LOad and RUn as well as SAve. If repeated saves of the same file are desired, all after the first can be accomplished by:

Format: >SA#n

NAME.EX and xx-yy*sa then default to their former values.

4) The ZEro command allows cassette length specification.

Format: >ZE #n Cxx
or: >ZE! #n Cxx

Legal values for xx are 30, 45, 60 and 90. If the Cxx is omitted, a C30 is assumed. Note the space following the #n is not optional, since nCxx is a perfectly legal HEX quantity, and GETNUM will return and xx for drive number!

5) The DName command allows changing (or adding) a cassette directory tape name. Automatic date insertion is optional.

Format: >DN#n

Tape drive selection is made with the #n (defaults to 0). the directory is read and the current name (if any) is displayed. A new name may then be entered. At any time, the "ESC" key will causes a return to PHIMON without modifying the present directory name. If name input is terminated with the RETU?N key, the current PHIMON date is included automatically. This autodating feature may be avoided by terminating input with the LINE FEED key rather than wit the RETURN key.

6) The DRefresh command will re-display the currently memory-resident directory listing. It will load the DIRECTORY overlay if necessary.

7) DTO/DTH has been replaced by STEPER. If the top 8K of memory is available, a DT command will execute STEPPER if resident or load and execute it if not.

III. ADDITIONAL FEATURES

1) All numeric communication with PHIMON-H is in HEX, including page numbers for the SAve command.

2) Upon booting up, <PHIMON-H> display as the current version's build date and asks for today's date. Date entry is free field, with any delimiter echoing as a /. Each program saved under this system includes the date in the header block. The LOad and RUn commands will display the save date after reading the file header block.

3) When initializing a directory with the ZEro command, a tape-name is requested. This may include up to 48 characters, and the current date is automatically appended. The DIrectory command will display the tape name along with the content listing. To correctly list a directory, <PHIMON-H> requires a name (try an old tape and see), but adding a name does not otherwise affect the old directory (try that too). Use the "DN" command described above to add a name. The name resides in the last half of the last block of the directory, reducing the maximum number of directory entries from 101 to 88.

4) The new PIP is one page smaller (8080 to Z-80 again) but has some new features. Rather than blindly continue after an error, it requests: "(SPACE) TO CONTINUE". Any other character will cause PIP to finish up with what it already has transferred.

5) An NMI (non-maskable interrupt) will force a re-boot of PHIMON without otherwise affecting memory content. This feature is best appreciated with trying to save an hour's worth of input with a clobbered copy of PHIMON.

6) Existing system tapes may be converted to <PHIMON-H> without disturbing their content. Place an old tape in drive #1 and a new <PHIMON-H> tape in drive #0, and type:

>BU #1

The boot-code and overlays of the old tape will be replaced by that of <PHIMON-H>.