Xref: utzoo comp.sys.ibm.pc.misc:1622 comp.misc:10098
Path: utzoo!utgpu!watserv1!watmath!att!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!samsung!uunet!saxony!dgil
From: dgil@pa.reuter.COM (Dave Gillett)
Newsgroups: comp.sys.ibm.pc.misc,comp.misc
Subject: Re: Packard Bell doesn't really exist (long)
Message-ID: <397@saxony.pa.reuter.COM>
Date: 11 Sep 90 04:54:06 GMT
References: <1990Sep6.154721.12322@iwarp.intel.com>
Organization: Reuter:file Inc  (A Reuter Company) Palo Alto, CA
Lines: 45

In <1990Sep6.154721.12322@iwarp.intel.com> dwight@iwarpk0.intel.com (Dwight McDuffee) writes:

>This is posted for a friend.  Please respond to the net or to Duane at
>dlundste@ncube.com.

>1.  Have you gotten any better response from Packard Bell than I have?
>Is this a one-time (or two-time, I guess) problem?  Has anyone gotten
>good service from Packard Bell?

Well, I've never had any of their equipment fail on me.

I believe there is some sort of partnership between P-B and Tatung; each
sells a complete product line, some of which is made by one company, some
by the other, and some is OEM gear from third parties.  Specifically, I
think the P-B and Tatung AT clones we had were actually manuafctured by
Intel as OEM products, with P-B and Tatung each providing their own BIOS
ROMs and labels.


>2.  Is it true that you can't so much as replace a RAM chip on one of
>these boards?  And if it's true, why would anyone design a board that
>way?  

Lots of companies build boards this way.

Consider that a silicon chip is basically a rock with some interesting
electrical properties.  If it's good enough to pass initial burn-in and
testing, the odds that it will fail under normal use are pretty small.

So suppose you're building huge numbers of these boxes, only a small fraction
of which are expected to fail.  How high a failure rate would be needed to
justify the parts cost of the sockets, the asssembly cost of inserting the
chips in the sockets, the indexing cost of maintaining separate spare-parts
inventories of each single type of chip, and the diagnostic cost of tracking
flaws to the individual chip?  (Recall that these latter will be borne by the
manufacturer if the machine fails while under warranty.  Recall too that a
flaw diagnosed as being "in" a particular chip may actually be in a circuit
board trace leading to/from that chip--or in the socket!  Some proportion of
weird failures encountered by users are fixed by re-seating socketted chips...)

So when you ask why "anyone" would design a board that way, "anyone" turns out
to be practically every manufacturer in the entire microelectronics--including
microcomputer--industry.

                                           Dave