Re: HP 9821 - display and tape transport issues #VintageHPComputers


Nice story Lee, I've enjoyed reading it, thank you.


On Fri, Feb 21, 2020 at 2:36 AM +0100, "Lee A. White via Groups.Io" <thermoman@...> wrote:

In the late 70's N S came out with a processor series designated as COPS (control oriented processor series). The company I worked for at the time had major field problems with controls using this processor - they were locking up in use. The problem was traced to extreme processor noise sensitivity. When operating next to electrically noisy equipment (like the magnetic contactor turning on a refrigeration compressor motor), the processor would see an instantaneous DC input voltage below the required "power supply good" voltage and the processor would halt. When the voltage rose back above the required "power good voltage" the processor would start an on chip start up timer and would then start the power on preset routines clearing memory, resetting the program counter,  etc.. When the on chip start up timer timed out the processor would be released to begin program execution. Problem was that the transient low voltage did not exist long enough for the on chip timer to discharge, so when the power returned to a valid level after the noise transient passed, the on chip timer had not yet fully discharged. As a result, it would time out too quickly and the processor would be released to execute program code before the power on preset routines had completed. The processor was then executing garbage and would lock up.
National was NO HELP AT ALL, denying that there could be anything wrong with their processors. After many calls and my getting very insistent, they challenged me to come out and prove it - to make it happen in their lab. On arrival I was shuttled off to a low level technician who was obviously supposed to humor me and then get rid of me. The welcoming was neither appreciative or respectful. I was clearly a PITA that they had no time for.
The tech took me to a test bench and watched as I set up the simple demo - the control with the processor in it and a compressor motor contactor set up to self cycle next to it. He connected some monitoring equipment to the processor and then turned everything on. In less than a minute the processor chip was locked up. The tech disappeared and returned with an engineer and it happened again. He got some more fresh parts - they all quickly locked up too. The engineer disappeared and returned with the engineering manager who saw it for himself and he then disappeared and returned with the VP - who was not happy. The VP suggested we try other processors in the series, and they all locked up too. It was now clear National had a design problem in the entire series that was going to require a mask change and cause a production interruption. Then the president/founder showed up and watched a processor lock up and then no one was happy. He asked if they could keep the test setup, and I happily agreed. Everyone was told to meet in a conference room in five minutes and I was asked to leave - escorted to the door actually. I didn't even get a lunch out of it.
National never replaced the bad parts in our inventory, nor did they do anything to make up for the bad product we had in the field or the costs of the field problems. They never said thank you either, nor did they make their problem known to other customers.
Care to guess how much respect I have for National Semiconductor.
A Japanese company got all our future mask programmable micro processor business - and we were a very big player in the automotive, appliance, and  heating and air conditioning controls businesses - what they call a tier one OEM supplier.
From: Rik Bos
Sent: Monday, February 03, 2020 3:27 PM
Subject: Re: [VintHPcom] HP 9821 - display and tape transport issues #VintageHPComputers

I think you missed an email, I mistook NI for NS.
I use NI software for design and simulation and am happy with it. I'm not happy with National Semiconductor logic IC's from the early 70th era.



On Mon, Feb 3, 2020 at 9:19 PM +0100, "Steve Leibson" <steven.leibson@...> wrote:

Hard to square your views of National Instruments with the 4000 to 5000 screaming fans I saw annually at NI Week in Austin. I'm sorry you seem to have gotten bad treatment from them, but my mileage definitely varied over the span of 35 years.


On Monday, February 3, 2020, 12:07:20 PM PST, <mariuszmleczko@...> wrote:

RIK: "A lot of transport problems are caused by a bad working EOT sensor, the LDR’s used in the system to detect the end of tape deteriorate and the resistance changes permanent to a low value."

I did a quick and very simple test. It is possible to run LDF, MRK, REW commands etc. with the opened cassete deck (without tape)....if at the same time top-right "cassette presence sensor" is pressed.
So I started it this way, trying to blink into a sensor by a pretty strong phone LED lighter, reaction.

Of corse I'm going
disassemble and measure EOT resistance anyway, but think also - if it was OK (and it's logic), then should be some reaction on my simple test.

Join to automatically receive all group messages.