2 Aug
2001
2 Aug
'01
9:36 a.m.
I have had a problem with some of my ESDI disk drives for a number of years
which may have a very simple solution.
When I use the drives in a BA123 box with an 11/73, the system functions correctly
while I am using 7 quad boards or less. Whenever I add the extra boards, the READ
operations on the Hitachi ESDI DK515-78 disk drives (Sigma RQD11-EC quad ESDI
controller) take about twice as long. The WRITE operations are still just as fast.
I have been looking at the problem again recently and while it has not yet been resolved,
I now begin to suspect that the cables between the RQD11-EC controller and the drives
may be too long. Just a guess is that the total length of the 34-pin cable is six feet
(with
the first portion to the first header being four feet and three more pieces of eight
inches
making a total of six feet to the last header) and the 20-pin cables (there are four with
three currently in use) are about four feet each. In addition, the three ESDI drives
currently
in use run off their own PC power supply (in fact it is the actual power supply that came
with the drives within a large raid set-up to be used with a VAX). I also salvaged the
fans
inside the box and each drive has its own fan. Normally, I only turn on one ESDI drive
when the other ESDI drives are not being used for backup purposes. In addition, I
have installed a heavy ground cable between the PC power supply and the BA123 box.
In case anyone asks, the last drive on the 34-pin cable is correctly terminated and
the other drives are not terminated.
Since everything runs VERY well in the BA123 with the DK515-78 drives while I
use seven or less quad boards, I have not tried to investigate much until recently since
the extra boards are used so rarely it has not been a problem 99% of the time.
And the fact that everything works so well has not led me to suspect the length of the
34-pin/20-pin cables. Does anyone have any suggested maximum cable lengths?
Sincerely yours,
Jerome Fine
2 Aug
2 Aug
11:57 a.m.
You shouldn't be having a cable length problem with the ESDI drives. The fast
signals on the data cables (1 per drive) are driven with differential drivers
capable of considerable length at that bandwidth. In fact, the number 40 feet,
sticks in my recollection. The signals on the 34-conductor cable aren't
particularly fast, but aren't driven, typically, with drivers capable of what
the diff drivers can do. I doubt you use that much cable, that you'd see the
impact of cable length. If you really think it's the cable, I'd recommend
switching to a twisted pair ribbon from the typical flat ribbon. It won't help
much, and what you might try is simply using a different cable. They do fail
sometimes.
Dick
----- Original Message -----
From: "Jerome Fine" <jhfine(a)idirect.com>
Newsgroups: alt.sys.pdp11,vmsnet.pdp-11
To: <classiccmp(a)classiccmp.org>
Sent: Thursday, August 02, 2001 8:36 AM
Subject: Length of Cables for ESDI Hard Disk Drives
I have had a problem with some of my ESDI disk drives
for a number of years
which may have a very simple solution.
When I use the drives in a BA123 box with an 11/73, the system functions
correctly
while I am using 7 quad boards or less. Whenever I
add the extra boards, the
READ
operations on the Hitachi ESDI DK515-78 disk drives
(Sigma RQD11-EC quad ESDI
controller) take about twice as long. The WRITE operations are still just as
fast.
I have been looking at the problem again recently and while it has not yet
been
resolved,
I now begin to suspect that the cables between the
RQD11-EC controller and the
drives
may be too long. Just a guess is that the total
length of the 34-pin cable is
six feet (with
the first portion to the first header being four feet
and three more pieces of
eight inches
making a total of six feet to the last header) and the
20-pin cables (there
are four with
three currently in use) are about four feet each. In
addition, the three ESDI
drives currently
in use run off their own PC power supply (in fact it
is the actual power
supply that came
with the drives within a large raid set-up to be used
with a VAX). I also
salvaged the fans
inside the box and each drive has its own fan.
Normally, I only turn on one
ESDI drive
when the other ESDI drives are not being used for
backup purposes. In
addition, I
have installed a heavy ground cable between the PC
power supply and the BA123
box.
In case anyone asks, the last drive on the 34-pin
cable is correctly
terminated and
the other drives are not terminated.
Since everything runs VERY well in the BA123 with the DK515-78 drives while I
use seven or less quad boards, I have not tried to investigate much until
recently
since
the extra boards are used so rarely it has not been a
problem 99% of the time.
And the fact that everything works so well has not led me to suspect the
length of
the
34-pin/20-pin cables. Does anyone have any suggested
maximum cable lengths?
Sincerely yours,
Jerome Fine
3 Aug
3 Aug
6:40 p.m.
Richard Erlacher wrote:
You shouldn't be having a cable length problem
with the ESDI drives. The fast
signals on the data cables (1 per drive) are driven with differential drivers
capable of considerable length at that bandwidth. In fact, the number 40 feet,
sticks in my recollection. The signals on the 34-conductor cable aren't
particularly fast, but aren't driven, typically, with drivers capable of what
the diff drivers can do. I doubt you use that much cable, that you'd see the
impact of cable length. If you really think it's the cable, I'd recommend
switching to a twisted pair ribbon from the typical flat ribbon. It won't help
much, and what you might try is simply using a different cable. They do fail
sometimes.
Dick
Jerome Fine replies:
I am trying again with the same overall configuration, but except for the same
DK515-78 hard drives, the hardware is different in a different BA123. I am
now up to 9 quad boards and it is still working. I expect to add 3 other quad
boards soon.
Except for the 20-pin data cables of which I have only four which are 4 feet long,
right now I am using only 18 inch cables, but the 34-pin cable is a duplicate which
is actually 8 feet long. After I have tested everything with 12 quad boards, I will
put everything from the other BA123 into this one and see if it still works. If it
fails on a specific board, that should identify the problem. Otherwise, it must be
the BA123 box itself.
The biggest part of the problem is that it is being done in the dining room where
there is room to move. (Let's not get back to that problem!) And I have a deadline
of the next time when my kids are coming for supper on Friday!!
Sincerely yours,
Jerome Fine
2 Aug
2 Aug
12:44 p.m.
New subject: PCI to NUBUS adaptation
I came across an interesting item today...an external expansion box that
contains 4 NUBUS slots, with their own PSU, with a heavy ribbon cable coming
off of it. According to the manual, it also had a PCI card that plugged
into the early PowerMac's, such as the 8500, and allowed them to use NUBUS
cards. Unfortunately, the PCI adapter is missing but it does still have one
of the National Instruments GPIB cards installed. The box was made by
Second Wave.
Jeff
--
8558
days inactive
8559
days old
3 comments
3 participants
participants (3)
-
edick＠idcomm.com -
jhellige＠earthlink.net -
jhfine＠idirect.com