RE: 2114s
Over the years i'd seen A higher than normal {2102 as comparison}
failure rate. So I had the opportunity to do some checking on why.
NOTE: the 2114 never had as good a failure rate as say 2102 or
other 1k designs and the next generation of 16k static parts were
substantially different with better failure rates.
2114s common failures:
Delayed stress damage due to ESD. Common failures are address
inputs.
Missing bit, usually a failed output due to either ESD or excessive
IO loading.
bond failures (works when hot or cold) most of those are plastic cases
and likely due to cleaners causing internal plastics to expand and lift
a bond {usually during board production}. Other likely causes, heat!
Ceramic devices seems and generally were more reliable.
The CMOS versions 6114 (nec D444) were more relaible, though ESD
induced stress failures were a major factor.
In the early 80s it was not uncommon to visit a site and NOT see proper
ESD protocal used when handeling MOS or CMOS devices. In most
cases getting the production people using would often see a substantial
drop in early failures{infant mortality}.
Allison
-----Original Message-----
From: Tony Duell <ard_at_p850ug1.demon.co.uk>
To: classiccmp_at_classiccmp.org <classiccmp_at_classiccmp.org>
Date: Monday, April 15, 2002 10:24 PM
Subject: Re: TTL computing
>>
>> That doesn't seem to me as a failure of the 2114, but, more likely, a
failure
>> in the update logic in the refresh memory circuit, possibly in a buffer
enable
>
>I was about to point out that the 2114 is an SRAM and doesn't need
>refeshing, but then I realised you were talking about the video refresh
>circuitry.
>
>> or a shortage of hold timing on the data with respect to the write line.
>> What, exactly, do you mean by "dud" character? If it appears exactly in
the
>> middle of a 2Kx8 memory array, it could, indeed be a stuck bit, and if it
runs
>> from the middle to the top/bottom that would be a candidate for a failed
>> memory bit also. If, however, it moves horizontally, or moves outside
the
>> range residing in a single device, it's clearly something else. Have you
>> tried moving the device around in the video memory array? It seems to me
that
>
>Last time I had a problem with a 2114 in a video RAM array, it appeared
>that the RAM had lost an address input. Writing a character to a location
>in RAM would affect the top 4 bits (IIRC) of a character 2 locations
>away (or something like that). It was clearly the RAM because swapping it
>over with the one next to it caused the fault to move to the bottom 4
>bits. And a new RAM cured the problem.
>
>> there have to be at least two of them, and the "dud" character, whatever
that
>> is, should follow the device.
>>
>> 2114's are just about as plentiful as any device of the era could be.
You
>> should not have a problem replacing it. I don't know what the problems
of
>> sending hardware from the U.S. to the U.K. are, but I'm willing to send
you a
>
>Very few AFAIK. I don't think I've ever had a parcel declared as
>'obsolete computer parts' even inspected by customs. Of course this might
>have changed now with the increased risk of terrorism (although we've had
>terrorist activity for years over here :-().
>
>One tip I was once told. Pack chips in transparent anti-static bags (and
>put an anti-static warning label on them. Customs officers are not
>totally clueless and won't open such a bag if they can clearly see what's
>inside (and provided the contents are what they're supposed to be, like
>'computer parts' (chips)). But if you pack the chips in an opaque bag,
>they might open it (to ensure it doesn't contain drugs, say), and might
>not know how to handle the chips without damaging them.
>
>-tony
Received on Mon Apr 15 2002 - 20:24:47 BST