[Coco] [OT]: First computer stories

[Gene Heskett]

On Tuesday 23 December 2003 17:42, John Guin wrote:
>One of my co-workers sent me this link since we had been talking
> about the "good ol' days" recently:
>http://www.siliconvalley.com/mld/siliconvalley/business/columnists/g
>msv/7549559.htm
>
>Merry Christmas all,
>John

Now that was a trip down memory lane!  Particularly some of the names 
of the folks relating their stories, like Andy Finkal, he of nearly 
legendary status in the Amiga groups.

Thanks for the link, John it was nice.

Memory Lanes...   I'll toss this one out.  Warning, verbose old f--t 
at play here.

Back in about 77 or early 78, I was a reasonably fresh carrier of the 
title of Assistant Chief Engineer at KRCR in Redding CA.  After 
getting thrown to the wolves a week after walking in the door by the 
Chief having a heart attack and manageing to survive about 6 months 
without him without looking too inept, I had been watching 
productions very laborious preparation of a commercial, which 
consisted of cuts only editing using 2800-2850 type Sony 3/4"decks, 
sometimes for more than one pass if they needed an A/B/C cut.  By the 
time they were done, it was several generations of copy degradation 
and often looked worse than a home vhs.  Then they had to copy it one 
more time, to a copy of a master tape that had the tones on the 
second audio channel that triggered the next spot when being 
sequenced by an 'automatic station break machine'.  The master tape 
was itself a copy of a copy of a copy, and therefore a bit flaky.

I figured that if I could write a program that would put those tones 
on the existing tape, it would save at least one generation of 
copying.  So I, never having seen or played with any small computer, 
went to see the GM and said I could build a machine that would do 
that, guessing it might run somewhere in the thousand dollar area by 
the time I was done.  Imagine my surprise when Dick Green said to go 
for it.  I wasn't entirely sure I could pull a rabbit out of the hat 
I chose to wear, but, how else was I gonna learn about these things?

So I studied up in the rags of the day and finally ordered a Quest 
Super Elf for the whopping total of about $295.  It arrived in due 
time, and I spent the next few weeks playing with its hex monitor, 
and reading up on the cpu, an RCA CDP-1802, even getting it to play 
some one bit music.  Eventually I ordered up a S-100 backplane, a 
blank predrilled for wirewrap S-100 vector board, an S-100 4k ram 
board and their case & psu, which made it look something like a 
miniature cash register.  I wire-wrapped an interface to control the 
tape machines, and a second interface to a control pendant they could 
carry around the room trailing a cable, enough of a video display to 
show two digits in an 8.8 format, built that and wrote the program as 
I went.  Oh, and another output to trigger the tone generator which I 
also hid in that box.  I put a jellcell battery on the ram, and 
reworked an old broadcast cart recorder to use as program storage.

The pendant control had a bunch of buttons on it, like to forward and 
reverse the deck until it was positioned on the first frame of video 
that was to be aired, then a with or without a new academy countdown 
leader, then one of several other buttons told it how long this spot 
was, which when pushed, told it how long the spot was and inititated 
the sequence:

If new leader, back up 13 seconds by control track pulse counting, 
stop 1 second and roll forward.  If no new leader, backup 7 seconds, 
stop 1 second and roll forward.

If new leader, turn on the machines video insert mode at -9.9 seconds 
from first video.  It was generating the video countdown anyway.

At 5.1 seconds before first video, turn on audio insert channel 2.
At 5.0 seconds before 1st video, trigger the tone generator for 1 
second. then shut off the audio insert.  This was the cue tone so the 
machine could know where the tape was located later.

At 2.0 seconds before 1st video, turn off the video insert mode.

Time itself in positive numbers till the entered time had elapsed 
minus 2.1 seconds, at which time turn on audio insert channel 2, and 
trigger the tone generator again at 28 seconds if it was a 30: spot.  
Other spot lengths accordingly.  This tone functioned as the start 
trigger for the next machine and spot in that break sequence.

I designed on the fly and built everything but the tone generator and 
the prefab memory and cpu board.  The code, which I still have paper 
copies of, actually took up less than half of that 4k of static ram, 
only about 1200 bytes, over half of which was lookup tables.  The 
video was generated via only 6 bytes of DMA'd data per vertical 
field, and it was all timed by the looping being held or allowed by 
the status of the house vertical drive.  I had some concerns that the 
1802, running at 1.78 mhz, and 8 clocks per actual machine cycle, 
would be able to keep up, so I put a software flag that it set while 
it was done for that field, and was looping in the wait for vertical 
drives leading edge state.  Hanging a scope on that flag, I was 
amazed that it was done and waiting by the end of line 20 of each 
field.  Since it couldn't start until the end of line 6 and the 
riseing edge of vertical drive, I was quite pleased.

The video wasn't that complex, only displaying those 2 digits with a 
permanent hardwired decimal point between them. ISTR that took about 
4 chips of std ttl stuffs.  It was 103 lines high and about 40 
u-seconds wide, so you could read the time off a 5" monitor from 
anyplace in the control room.  The dma was sequenced by a string of 
decade counters counting h-sync as it went by.

The 1802 architecture is quite unusual, and its major fault was a lack 
of a builtin dram controller, forcing one to use static ram.

I left there after about 2 years, and was only back there once more, 
about 8 years later, and it was still in daily use abeit with much 
newer machines.  That can be an eon in the broadcast business, so the 
fact that it was still in use surprised me.  To them it was 'just how 
it is' :)  When I was intro'ed to their newest tech who had changed 
the plugs to fit the new machines and adjusted a couple of delays in 
the code to fine tune things, he didn't act as if he was the least 
impressed.  So he was either pretty good, or didn't have a clue of 
the overall design.  I did leave good docs on it though.  I didn't 
stick around long enough to find out which, what with a car full of 
"are we there yet?" kids...

I did an automatic transmitter controller using Z-80 based 
Micro-professor boards for the next station I was at, but thats 
another story, for another time.

And most of you are familiar with my coco2 project that replaced a 
$20,000 Grass Valley system and did the job far better, which also 
set some sort of a record for longevity in this business as it was in 
continuous use here at WDTV for 13 years.  When I retired, they were 
in the process of buying a new production switcher, so when it was 
finally installed, & the old one fully ripped out, they gave the old 
coco2 system back to me.  Its in a box in the basement.

Anyway, thats my trip down memory lane.  My mental "golden years" so 
to speak.

And to those that have survived this bit of boredom/trivia:

Merry Christmas everyone!

-- 
Cheers, Gene
AMD K6-III at 500mhz 320M
Athlon1600XP at 1400mhz  512M
99.22% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.