Reader to Reader for #81


 
Send your letters for Reader to Reader to:
The Computer Journal
P.O. Box 3900
Citrus Heights, CA 95611-3900

Subject: Subscriptions; progress on various fronts
From: Herb Johnson 

Read through the TCJ Web page, glad to see your progress. I see you are
asking for subscriptions from your advertizers, so I will mail you today
a $24 check to start my subscription from issue #80, which I don't
believe I recieved. I've missed Rich Rodman's articles in particular:
his interests and mine coincide  on small networking like his "tiny TCP"
stuff - has he advanced that work lately?

I've said I would write something on "classic Macs", or compact Macs or
"9-inch" Macs, which I'm now also selling as well as my S-100 stuff. But
this is something new to me, and I've waited for my sense of that
"domain" to jell into something that might also be interesting for
your/my readers. One thing I like about the small Mac owners is their
enthusiasm for their systems. In addition, these are relatively useful
systems even by today's standards for CASUAL personal computing. I will
consider some kind of writeup in that context. As these systems are not
bus-based, few folks are doing any elaborate hardware development. There
are serial-based odds and ends however, and the old Mac CPU cards
themselves are potentially hackable: problem being the development tools
are scarce. The opposing market force to providing Mac tools and
technologies is the number of inexpensive single-board computers and
their free development tools. Still, even a Mac as terminal and
filestore is a useful acces sory application. If you see an article in
this I'd be encouraged.

I'd like you to consider referencing my Web page:
http://pluto.njcc.com/~hjohnson/  which describes the kind of S-100
documents I offer, as well as some tidbits on S-100 systems. As time
progresses I will also list the S-100 cards I have available, but of
course those interested can ask. I'm also buying S-100 cards and systems
at very modest prices, for sale at modest prices. 

The short story in the S-100 world is the obscene prices collectors are
paying for Altair 8800 systems. Most of the rest of the S-100 world
sells for zero to tens of dollars per card - or system. The increased
use of the Internet has permitted more of this stuff to be traded
around: the comp.os.cpm maillist is more active these days with people
asking the usual questions about what to get, where it is, and how to
use it. I've had more interest from those wanting documentation for
their "new" computers, less interest in cards, modest interest in
complete systems (which I rarely provide as it's a lot of work!).

The Web page also lists early Mac systems for sale, also well-priced;
and my astronomical interests. I'm spending a lot of my time on amateur
astronomy: I'm invloved with the TASS amateur sky survey project, which
was written up in Sky and Telescope magazine a few months ago. Amateurs
with CCD cameras monitor the sky around the celestial equator, looking
for variable stars and whatever is out there at 10th-12th magnitude.
It's an interesting study on shoestring science.

My regards to you and to your readers, and my previous readers.
Herb Johnson, still "Dr. S-100"

----------------------------------------------

Ed. - I've been trying to collect Digital Group systems.  Here's an
email conversation I had recently with someone who knows about them.

Subject: Digital Group System in photo
From: Rex Widmer 

I was surfing the net and saw a photo of you behind a vintage Digital
Group System - dress cabinet, in the usual style of lid removed! (on the
Vintage Computer Festival web pages, www.siconic.com/vcf/) I wonder if
any of the post bankruptcy products were in the machine. Hundreds of
Z80+ (higher clock rate, real time clock, multiple boot eprom,
heartbeat, etc) processor boards were built down here in my dungeon and
sold into the market. Many improved motherboards (one foot square, 1/8"
thick PC), Also lots of I/O Processors, really similar to the (later on
the scene) Hayes ESP cards, a fair number of TV80's (H89 emulator, but
bus connected) and a tiny number of solid state Pseudo Disks (512K in
1981), and various hard drives.

There are still some of these machines (with all of the RWS replacement
parts instead of the Digital Group) running 24 hours per day.  Feed them
a new Z80 chip every now and then and they just keep working. Sometimes
I wish they would all be museum pieces - but the customers paid real
money for them, and they (at least the RWS versions) were designed to
run "forever with the exception of things like EPROMS that may need
reprogramming and fans which need replacing.

----------------------------------------------

Dave Baldwin / The Computer Journal wrote:

I actually know very little about the Digital Group machines except that
they were comtemporary with the Altairs and Imsai's.  I read about them
first in a Byte article fifteen years ago.  After I took over TCJ, I
started collecting them but I'm short on documentation.  If you can help
with any of that, I'd appreciate it.  I have two machines/motherboards,
two quad tape drive units in cases and one dual eight-inch drive unit.
I'd love to hear about these machines, especially if you could write
some  articles for TCJ about them.

----------------------------------------------

Subject: Re: Digital Group System in photo
From: Rex Widmer 

The Digital Group machines emerged primarily from the Ham marketplace as
opposed to the computer hobbiest.  Lots of overlap between the two
areas.

I beleive they were first shown at the Dayton Hamvention in 1975 -
matching what was shown in Dr. Suding's original Byte articles -
interchangable 8080 and 6800 cpus, not much later with the Z80 which
consitituted most of the production.

The company entered Chapter 11 - reorganization in early 1979, and was
liquidated in August 1979 (one creditor would not sign up for the 100
cents on the dollar with interest recovery plan).  A liquidation auction
was held at the Digital Group facility in Denver in the fall of 1979.

Several small companies (mine included) stepped in to continue support.

The intellectual property - board layouts, copyrights etc, was very
poorly protected by the Bankruptcy trustee.  Some folks (not me) paid
tens of thousands for the original artwork, but then found others with
"working tooling" churning out boards from that artwork.  Was a mess for
those who spent the money. After seening the original artwork sell for
ludicrous prices, I executed a plan to build compatable, replacement
cards all of which had significant improvements.  The first was a new
processor card - Z80 but with many improvements, the next was a 64K
static ram card (replaced 8K static, 32K static and 32K dynamic cards)
using 2K x 8 technology, also a new motherboard - better power
distribution, more accurate geometry, and big enough to not require
expansion.  A new disk controller was produced, with a design that was
about 3/4 generation newer, and much more manufacturable.  A series of 4
peripeheral cards was produced using dedicated Z80 coprocessors with
onboard eprom and ram (made design and update work much easier).  This
series included a smart I/O card (buffering and protocol in the card), a
512 K solid state disk, a 80 x 24 (or more) video card, and a
co=processor prototype.  The last component produced was a daughter card
for the processor which provided a SCSI hard disk inter face.

A number of articles are possible, mainly an issue of getting enough time.

Speaking of your machines, the quad tapes were an on-going challange (as
were several other components).  The design was good on paper, then was
poorly executed - i.e. the voltage to the drives was not properly
regulated, the heads of all 4 drives were wired in parallel, and the
head preamp was on the end of a long noisy cable.  A chap at Stanford
published a newsletter with lots of tips on how to fix all of these
issues - Hugh McDonald (later of Tall tree systems as I recall).

The 8" disk subsystems might have one of several disk controllers
attached.  Original single density card (42 pin NEC LSI), noted for
needing mods and wiping out disks if powered down with disk in drive.
DIgital Group (or direct clone) double density board - high density card
with WD1791 chip.  Difficult to manufacture, lack of proper voltage
regulator for critical components, (made card not portable, sensitive to
system's power supply voltage), and card had too much noise sensitivity.
RWS type III controller - fixed all of that (only slight bragging), used
newer data seperator technology and proper board layout. Many operating
systems were available for the disk based units.  More for the double
density boards.  CP/M 1.4, 2.2. 3.0, MCOS (CP/M clone), OASIS, DISKMON
(also with a variant called PHIMON for the digital tapes you have) and
others.

The machines still running are typically involved in process monitoring
- a testing lab in Cincinatti has a bunch - Dow Chemical Co had a bunch
(last time I heard), and several local medical centers who have
dedicated message switch applications running on them.

----------------------------------------------

Dave Baldwin / The Computer Journal wrote:

Thanks for the info.  May I print your email in TCJ's 'Reader to Reader'
column? If you get the time, I'd like to see some articles.  My
impression is that the Digital Group equipment is rather neglected
compared to the Altair and Imsai crowd.

----------------------------------------------

Subject: Re: Digital Group System in photo
From: Rex Widmer 

You may certainly print the letter.

Digital Group stuff has certainly not attained the "collectable" status
of the Altair / IMSAI S100 machines. Many of the orginal machines were
without fancy cabinets - the kits were frequently sold that way.  Also
no fancy front panels - nice cost savings by doing all of that type of
stuff via EPROM and a mini monitor loaded from audio cassette.

Like the Altair, original machines frequently needed a tweak or two.  8K
memory boards had driver (to CPU) problems, origianl motherboards had
keystoning in the phototools which made the dual connector I/O cards act
up, motherboard voltage drop was a problem when you might have 25 Amps
of +5 running  56K of 8K memories (the RWS 32K used about 500 ma!)...

But... the systems could be made to work reliably, once you knew the
tricks, and perhaps got rid of a few first generation parts.

Biggest difference was that there was far fewer vendors of significance
until after the bankruptcy.  Godbout did make a suitable 32K static
card, that probably hit the street prior to the failure.  Doug Bell
A.K.A Bell Controls adadpted a fine Tarbell S100 floppy card to the
system prior to the failure, but that adaptation took a fair amount of
mechanical fiddle work and wire wrap to lash a S100 card into the D.G.
chassis / bus structure.

The surprising point is that there were active user get togethers at the
Dayton Hamvention for over 10 years after the bankruptcy.  Even now,
almost 19 years after, there will no doubt be at least an informal
gathering of the faithfull in 10 days at the Hamvention.


The Computer Journal, PO Box 3900, Citrus Heights, CA 95611-3900, USA.
Email: tcj@psyber.com
Ph: 800-424-8825 / 916-722-4970
Fax: 916-722-7480 / BBS 916-722-5799
Copyright © 1996,1997,1998, The Computer Journal. 7/5/98