5 Mar
2003
5 Mar
'03
12:52 p.m.
John Allain writes:
.
Somebody should say here that it's pretty strange
to have a
workstation that has all SCSI connectors on the outside and
an MFM disk on the inside. I did a mental doubletake on this
("hey, this makes no sence") a short while ago.
Since I haven't seen any other responses, I'll dive in... This is
all speculation - I'd love to have someone with a line into DEC
Engineering answer this properly. To put it another way, I may be
completely off my rocker - reader beware! ;^)
The VAXstation/MicroVAX 2000 was released in 1985 (cf. R.D. Davis'
uVAX FAQ), which means development and planning may have started
as early as 1983 - I'm just guessing here.
In the early 80s large (> 20MB) ST-506 hard drives were still
hugely expensive beasts. If you did find a native SCSI device, I
imagine the costs would be much higher for the extra interface logic
on the drive. Without native SCSI on the device, you needed a
translator like the Adaptec ACB-4000 (SCSI/ST-506) or Emulex MT-02
(SCSI/QIC-02).
The base VS2000 was a pretty small box - perhaps even ground-
breaking for this timeframe, given it could run full 32-bit uVMS
or Ultrix and had internal storage, unlike the Sun-2/50. Having to
develop and wedge in a unique converter board may have been riskier
than adapting a version of a board you were already producing/
designing for the other uVAX models like the RQDX3.
Going back to that comparison to the Sun-2/50 may answer the question
of why SCSI was acceptable for the tape interface. Since the tape
drive was going to be external, there's plenty of room for that extra
logic board. And it's only a tape drive, so if we can't get decent
throughput out of our first (?) SCSI interface, well, who's going to
notice when the tape drive is as slow as the TK50? ;^)
After a few years of Apple, Sun, and others driving SCSI into the
marketplace the whole picture changed and by the time the first pizza
box workstations come out everything is SCSI.
Anyway, that's my best shot at an explanation. Again, I'd love to
hear from someone who was inside DEC (Megan?)...
--Steve.
5 Mar
5 Mar
1:37 p.m.
--- Steve Jones <classiccmp(a)crash.com> wrote:
uVAX FAQ), which means development and planning may
have started
as early as 1983 - I'm just guessing here.
In the early 80s large (> 20MB) ST-506 hard drives were still
hugely expensive beasts. If you did find a native SCSI device, I
imagine the costs would be much higher for the extra interface logic
on the drive. Without native SCSI on the device, you needed a
translator like the Adaptec ACB-4000 (SCSI/ST-506) or Emulex MT-02
(SCSI/QIC-02).
I bought my first microcomputer hard drive in 1987. It was an ST-225
that I attached to an Amiga 1000 over an 8-bit ISA bus adapter called
"The Wedge" - the drive (with WX-1 ISA controller) was around $350.
ISTR that bare ST-225 drives were right around $300-$325. Between
the PC drive/controller and the Amiga adapter plus an external (used)
enclosure, I put 20MB on my Amiga for around $500 total. At the same
time, a C Ltd SCSI interface plus external 20MB SCSI drive (probably
an ST-225 + ACB-4000, not certain) was $1,000.
I do not think that the uVAX-2000 ever shipped with anything as small
as 20MB. All the ones I've seen had either RD53s (71MB) or RD54s (154MB).
DEC, of course, charged a huge amount per megabyte for their drives,
so my comparison of the costs of an Amiga drive have to be taken into
that context.
I do not know why DEC decided to go with SCSI for the uVAX-2000 tape.
That they went with ST506/ST412 disk drives is no surprise - it was
what they had been working with since the uVAX-I/RQDX1, Rainbow and
DECmate - if you wanted 5.25" storage, your choices in 1983-1985
were limited to ST506/ST412 and ESDI (ESDI cost even more, but it
was higher performance). Even the Mac drives weren't embedded SCSI
until later, and then, 20MB was generous (ST-225N in 5.25" width,
and a number of 3.5" models from various vendors).
I'm surprised that DEC didn't make a wierd buffered cable/interface
for the raw TK50 interface. Given how much hardware is on the
board in a TK50Z, I suppose there wasn't room. The SCSI circuit
on the uVAX2000 motherboard takes up only a few square inches (the
5380 chip itself, a minimal amount of select logic, plus the connector
at the back).
So... just a couple of data points on drive costs of the era when the
uVAX-2000 was surfacing. I saw my first one at a DECUS in either Anaheim
or Cincinnati around 1987 or 1988 (not that they were new then, but they
were new to me - everywhere I'd worked, it was all Unibus and Qbus VAXen).
-ethan
7970
days inactive
7970
days old
1 comments
2 participants
participants (2)
-
classiccmp＠crash.com -
erd_6502＠yahoo.com