On Sat, 15 Sep 2018 16:56:02 +0100, Martin Brown
<'''newspam'''@nezumi.demon.co.uk> wrote:
>On 15/09/2018 04:28, krw@notreal.com wrote:
>> nOn Fri, 14 Sep 2018 09:06:33 +0100, Martin Brown
>> <'''newspam'''@nezumi.demon.co.uk> wrote:
>>
>>> On 14/09/2018 03:13, krw@notreal.com wrote:
>>>> On Thu, 13 Sep 2018 06:31:21 -0700 (PDT), pcdhobbs@gmail.com wrote:
>>>>
>>>>>> IEEE-488 is terrible. A typical committee product.
>>>>>
>>>>> Except that it wasn't--it was developed in-house by HP in the '70s.
>>>>
>>>> That may be but it doesn't change the fact that it's terrible.
>>>
>>> At the time it was invented it was really quite good. HP even used it to
>>> connect their external hard disks. The handshaking meant that you knew
>>> all devices had seen the commands and data (but also that things ran so
>>> that the slowest device set the workable transfer speed). ISTR Good for
>>> 1Mbyte/s peak or more realistically on a practical setup 500kbyte/s.
>>>
>>> The main alternative for instrumentation was poxy RS232 speeds 9600baud
>>> =1kbyte/s and the risk of buffer overrun (remember those?).
>>>
>>> It is still used in some places there is a lot of GPIB kit still out
>>> there and the connectors were virtually indestructible (although they
>>> didn't tolerate metallic swarf or strong acids very well).
>>
>> It was flaky as all hell. Sometimes it demanded to be a star
>> configuration (I had 8 488 connectors stacked on some equipment) and
>> sometimes *had* to be daisy-chained. Sometimes something between.
>
>There were line length limitations and only a fool would stack
>connectors more than three or four deep on one unit.
Then only a fool could get the POS to work. That was the *ONLY*
configuration that worked!
>No-one could ever
>call them pretty but they did the job. The worst connector I have ever
>seen was a Belgacom phone plug which looked more like it was designed
>for 3 phase mains than anything else.
If you had days and days to fuck around with it and could then leave
it alone forever. Maybe.
>
>> What an ungodly mess, particularly when there was more than one
>> manufacturer of equipment in the string. No, it was complete shite!
>
>I never had any problems with it and we made full implementation system
>controllers. There were flakey partial implementations by some
>instrument manufacturers but if you avoided the cowboys it was generally
>very reliable. You could even break the rules a little bit if you didn't
>want absolutely maximum speed (back then very few devices could actually
>send bulk data at 1MB/s like frame stores and transient recorders).
That's exactly what I was doing but they had their own 488 bus. It
was the power supplies and volt meters that were really picky. I'm
*so* glad I never have to use 488 again.
Reply by Jeff Liebermann●September 15, 20182018-09-15
On Fri, 14 Sep 2018 05:09:53 -0700 (PDT), pcdhobbs@gmail.com wrote:
>>In its day it was a reliable fast method of data transfer. HP even put
>>their disk drives on it for some of the laboratory micro/mini computers.
>
>I had a couple of those, a 9826 and a 9816 with a math coprocessor
>board and a Rocky Mountain Basic-compatible compiler (from some outfit
>called Infotek iirc). I still have a couple of instruments that
>have Instrument Basic installed and so can control other ones.
>
>Cheers
>
>Phil Hobbs
I remember those (fairly) well. At Granger Assoc in 1981(?), I was
designing parts of a 900MHz radio system. For production test, we
bought an HP8953XT RF test system consisting of a 5 ft rack full of
HP-IB connected RF instruments, with an HP8954A programmable
interface[1]. The computer that ran it was a Model 200 series HP9816
with HP-IB connected HP9121 dual 3.5" floppy drive. It was the first
3.5" floppy I had seen:
<https://www.hpmuseum.net/display_item.php?hw=4>
<http://www.old-computers.com/museum/computer.asp?st=1&c=879>
<https://www.google.com/search?q=hp9816&tbm=isch>
In those daze, we didn't have a mouse. The 9816 used a left handed
shaft encoder disk on the keyboard. A software switch controlled
which direction the cursor moved. It came with a book full of very
useful test routines in Rocky Mountain Basic and an addictive game
called Reversi. Only about $36,000 for everything in 1981.
I didn't get to play with the computer beyond getting the RF stuff
interfaced and running. Suffice to say that we never had any problems
with the system that could be attributed to HP-IB, except for one
defective HP-IB connector, that easily repaired. We had plenty of
problems with the instruments and in particular, the HP8954A interface
box, the HP-IB stuff worked as expected. Of course, this was an all
HP system. I would expect more entertainment from a mix of equipment
from different vendors.
[1] I couldn't find any photos online, but if anyone is interested, I
can scan some pages from the 1980 and 1999 HP Test and Measurement
catalogs:
<http://www.learnbydestroying.com/jeffl/crud/HP-Catalogs.jpg>
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
Reply by Martin Brown●September 15, 20182018-09-15
On 15/09/2018 04:28, krw@notreal.com wrote:
> nOn Fri, 14 Sep 2018 09:06:33 +0100, Martin Brown
> <'''newspam'''@nezumi.demon.co.uk> wrote:
>
>> On 14/09/2018 03:13, krw@notreal.com wrote:
>>> On Thu, 13 Sep 2018 06:31:21 -0700 (PDT), pcdhobbs@gmail.com wrote:
>>>
>>>>> IEEE-488 is terrible. A typical committee product.
>>>>
>>>> Except that it wasn't--it was developed in-house by HP in the '70s.
>>>
>>> That may be but it doesn't change the fact that it's terrible.
>>
>> At the time it was invented it was really quite good. HP even used it to
>> connect their external hard disks. The handshaking meant that you knew
>> all devices had seen the commands and data (but also that things ran so
>> that the slowest device set the workable transfer speed). ISTR Good for
>> 1Mbyte/s peak or more realistically on a practical setup 500kbyte/s.
>>
>> The main alternative for instrumentation was poxy RS232 speeds 9600baud
>> =1kbyte/s and the risk of buffer overrun (remember those?).
>>
>> It is still used in some places there is a lot of GPIB kit still out
>> there and the connectors were virtually indestructible (although they
>> didn't tolerate metallic swarf or strong acids very well).
>
> It was flaky as all hell. Sometimes it demanded to be a star
> configuration (I had 8 488 connectors stacked on some equipment) and
> sometimes *had* to be daisy-chained. Sometimes something between.
There were line length limitations and only a fool would stack
connectors more than three or four deep on one unit. No-one could ever
call them pretty but they did the job. The worst connector I have ever
seen was a Belgacom phone plug which looked more like it was designed
for 3 phase mains than anything else.
> What an ungodly mess, particularly when there was more than one
> manufacturer of equipment in the string. No, it was complete shite!
I never had any problems with it and we made full implementation system
controllers. There were flakey partial implementations by some
instrument manufacturers but if you avoided the cowboys it was generally
very reliable. You could even break the rules a little bit if you didn't
want absolutely maximum speed (back then very few devices could actually
send bulk data at 1MB/s like frame stores and transient recorders).
--
Regards,
Martin Brown
Reply by ●September 15, 20182018-09-15
nOn Fri, 14 Sep 2018 09:06:33 +0100, Martin Brown
<'''newspam'''@nezumi.demon.co.uk> wrote:
>On 14/09/2018 03:13, krw@notreal.com wrote:
>> On Thu, 13 Sep 2018 06:31:21 -0700 (PDT), pcdhobbs@gmail.com wrote:
>>
>>>> IEEE-488 is terrible. A typical committee product.
>>>
>>> Except that it wasn't--it was developed in-house by HP in the '70s.
>>
>> That may be but it doesn't change the fact that it's terrible.
>
>At the time it was invented it was really quite good. HP even used it to
>connect their external hard disks. The handshaking meant that you knew
>all devices had seen the commands and data (but also that things ran so
>that the slowest device set the workable transfer speed). ISTR Good for
>1Mbyte/s peak or more realistically on a practical setup 500kbyte/s.
>
>The main alternative for instrumentation was poxy RS232 speeds 9600baud
>=1kbyte/s and the risk of buffer overrun (remember those?).
>
>It is still used in some places there is a lot of GPIB kit still out
>there and the connectors were virtually indestructible (although they
>didn't tolerate metallic swarf or strong acids very well).
It was flaky as all hell. Sometimes it demanded to be a star
configuration (I had 8 488 connectors stacked on some equipment) and
sometimes *had* to be daisy-chained. Sometimes something between.
What an ungodly mess, particularly when there was more than one
manufacturer of equipment in the string. No, it was complete shite!
Reply by bitrex●September 14, 20182018-09-14
On 09/14/2018 04:06 AM, Martin Brown wrote:
> On 14/09/2018 03:13, krw@notreal.com wrote:
>> On Thu, 13 Sep 2018 06:31:21 -0700 (PDT), pcdhobbs@gmail.com wrote:
>>
>>>> IEEE-488 is terrible. A typical committee product.
>>>
>>> Except that it wasn't--it was developed in-house by HP in the '70s.
>>
>> That may be but it doesn't change the fact that it's terrible.
>
> At the time it was invented it was really quite good. HP even used it to
> connect their external hard disks. The handshaking meant that you knew
> all devices had seen the commands and data (but also that things ran so
> that the slowest device set the workable transfer speed). ISTR Good for
> 1Mbyte/s peak or more realistically on a practical setup 500kbyte/s.
>
> The main alternative for instrumentation was poxy RS232 speeds 9600baud
> =1kbyte/s and the risk of buffer overrun (remember those?).
>
> It is still used in some places there is a lot of GPIB kit still out
> there and the connectors were virtually indestructible (although they
> didn't tolerate metallic swarf or strong acids very well).
>
I'm still using mid-80s SCSI-1 with some stuff because...reasons.
The cables are more expensive than the PC adapter cards.
Reply by ●September 14, 20182018-09-14
>In its day it was a reliable fast method of data transfer. HP even put
>their disk drives on it for some of the laboratory micro/mini computers.
I had a couple of those, a 9826 and a 9816 with a math coprocessor board and a Rocky Mountain Basic-compatible compiler (from some outfit called Infotek iirc). I still have a couple of instruments that have Instrument Basic installed and so can control other ones.
Cheers
Phil Hobbs
Reply by Martin Brown●September 14, 20182018-09-14
On 14/09/2018 03:13, krw@notreal.com wrote:
> On Thu, 13 Sep 2018 06:31:21 -0700 (PDT), pcdhobbs@gmail.com wrote:
>
>>> IEEE-488 is terrible. A typical committee product.
>>
>> Except that it wasn't--it was developed in-house by HP in the '70s.
>
> That may be but it doesn't change the fact that it's terrible.
At the time it was invented it was really quite good. HP even used it to
connect their external hard disks. The handshaking meant that you knew
all devices had seen the commands and data (but also that things ran so
that the slowest device set the workable transfer speed). ISTR Good for
1Mbyte/s peak or more realistically on a practical setup 500kbyte/s.
The main alternative for instrumentation was poxy RS232 speeds 9600baud
=1kbyte/s and the risk of buffer overrun (remember those?).
It is still used in some places there is a lot of GPIB kit still out
there and the connectors were virtually indestructible (although they
didn't tolerate metallic swarf or strong acids very well).
--
Regards,
Martin Brown
Reply by ●September 13, 20182018-09-13
On Thu, 13 Sep 2018 06:31:21 -0700 (PDT), pcdhobbs@gmail.com wrote:
>>IEEE-488 is terrible. A typical committee product.
>
>Except that it wasn't--it was developed in-house by HP in the '70s.
That may be but it doesn't change the fact that it's terrible.
Reply by Martin Brown●September 13, 20182018-09-13
On 13/09/2018 14:31, pcdhobbs@gmail.com wrote:
>> IEEE-488 is terrible. A typical committee product.
>
> Except that it wasn't--it was developed in-house by HP in the '70s.
When it was originally called HPIB before IEEE adopted it.
http://www.hp9845.net/9845/tutorials/hpib/
In its day it was pretty fast and robust when compared to serial RS232
before much faster Ethernet and token ring topologies came along.
I found the later revisions 488.2 of the standard a bit top heavy. YMMV
In its day it was a reliable fast method of data transfer. HP even put
their disk drives on it for some of the laboratory micro/mini computers.
Commodore also adopted it for external hard disk interface.
--
Regards,
Martin Brown
Reply by piglet●September 13, 20182018-09-13
On 13/09/2018 15:00, Jeroen Belleman wrote:
> pcdhobbs@gmail.com wrote:
>>> IEEE-488 is terrible. A typical committee product.
>>
>> Except that it wasn't--it was developed in-house by HP in the '70s.
>
> It's clumsy, complicated and slow. I've used it many times
> over the years and still do, yet never quite managed to fully
> understand it. I even wrote working driver routines for this
> silly TMS9914 chip. I never got deeper than just enough to get
> the job done.
>
> Jeroen Belleman
I never got to use the TMS9914 but got on OK with the MC68488.
I very successfully implemented an HP-IB subset using discrete logic
that was actually great fun to design :>
piglet