On May 17, 2014, at 9:26 AM, <[email protected]> 
<[email protected]> wrote:

> I don't remember exactly how QIC tapes work, but it rings a bell that 
> they might have been different.

Quarter Inch Cartridge (QIC) tapes had 4 tracks recorded serially in one 
direction.  The device driver had to rewind the tape when each track reached 
the physical EOT (I think it was a hole in the tape) and continue.  A tape mark 
is a tape mark.

We built the first portable 16-bit digital seismograph in the late 1970s (that 
we still use!).  It uses an Intersil 6100 CMOS PDP-8 CPU and QIC cartridges, 
but we write them in serpentine format to keep up with the data rate.  (Can't 
wait for the rewind.)  That required a head with dual write gaps; standard QIC 
tape transports have heads with only one read/write gap.  (The erase gap is in 
the middle of the two.  In 7-track and 9-track tape drives, there was a 
separate erase head.  That was fine, since the tape was always written in one 
direction.  In small cartridge tapes, the erase head became an erase gap on the 
same head used to read/write, to save space I imagine.)  I modified the RSX MT 
driver to handle the serpentine format for a custom UNIBUS controller.  I 
remember when it was delivered that the microcontroller firmware timing had to 
be adjusted because the 11/70 was faster than the PDP-11 they used to develop 
the hardware and it was not always responding to device "register" read-write 
or write-read sequences properly. :)  That was quickly cured.  Other than that, 
I think we have had no problems with it.

We (Gary Maxwell) wrote the O/S for the seismograph as well.  (The O/S 
delivered by the contractor we used was unusable.)  We used the DECUS PAL PDP-8 
RSX cross assembler.  (Gary rewrote the symbol table handling to use hash 
tables which made it usable.  I rewrote it from MACRO-11 to DECUS/Microsoft/DEC 
C, CPAL, so it runs anywhere now.)  Pretty clever to move 16-bit data around on 
a 12-bit CPU.

We still have instruments in the field.  (Our second version of the instrument 
has a larger capacity tape and is played back on a PC system with a custom 
ISA-bus controller.  Most of our inventory is the second version.)  We play 
back tapes on an LSI-11/73 RSX-11M-Plus system with a QBus-to-UNIBUS adapter, 
all completely controlled by a VAX/VMS system over DECnet/Ethernet.  When our 
techs playback tapes, they log in to VMS and run a program that opens a DECnet 
remote task object on the RSX system that runs the program that actually reads 
the tapes.  I don't think they have ever logged in to RSX.  I maybe do once 
every decade or so when we have to move the equipment.  (We've been downsizing 
for years.)  The VAX and Alpha VMS systems and the LSI-11 just run and run 
(except for the CD-ROM on the Alpha, which is a PC drive that we have had to 
replace several times).  I think the LSI-11 system has a mix of hardware from 
several OEMs, including Emulex RQDX-emulating SCSI II disk controllers 
connected to CDC Wren 766 MB disk drives partitioned by the controller into 
four logical drives so RSX could use drives that big.  I think the 
QBus-to-UNIBUS adapter is an Able Qniverter.

It has all been running for decades.  DEC made rock solid hardware, as did 
their hardware add-on OEMs.

Larry Baker
US Geological Survey
650-329-5608
[email protected]


_______________________________________________
Simh mailing list
[email protected]
http://mailman.trailing-edge.com/mailman/listinfo/simh

Reply via email to