Hardware Diagnostics (Was: A talent for repair)

[Colin McGregor]

On 7/27/10, Scott Sullivan <scott-lxSQFCZeNF4 at public.gmane.org> wrote:
> Dear List Members,
>
> The story quoted below reminds me how unverifiable our computer hardware
> can be.
> As such I wish to relay a story of my own experience and at the end
> propose a question.
>
> I own an OLPC XO.
>
> For awhile after I got it I was experiencing strange behaviour when
> using the keyboard. Wrong characters, the inability to use
> ctrl-sequences. Like the owner of the Toshiba from the other story I
> suspected that it might be a loose connection from the keyboard to the
> main board.
>
> The OLPC XO was designed with serviceability in mind, so taking it apart
> was rather easy with only three types of screws used in the whole design
> (with spares located inside the casing!). But I found nothing to be a miss.
>
> What eventually allowed me to track down the issue was the built in
> hardware diagnostics in the firmware (BIOS eqv.).
>
> The OLPC XO boot firmware includes visual diagnostics for all of it's
> hardware components.
> These include live trackpad data visualization, fish-eye generation on
> the USB ports (checkable with an ossiliscope) and most important in my
> case a live key press map.
> With this diagnostic I was able to watch the keys as I pressed them and
> after playing with it for a bit I noticed the Ctrl key was sticking.
>
> Now, this is not your normal keyboard in the OLPC XO, it's a fully
> sealed membrane. It is also rather silly how I repaired the problem by
> accidentally causing a rip in the membrane around the key.
>
> But to put this story in to perspective, if I had not had that tool in
> the firmware, I would have never likely figured out the cause.
>
> My question to the list is this,
>
>      What would it take to get manufactures to making their products
> serviceable again?
>
> (And I don't want to hear "Fat Chance" or "can't beat greed" answers,
> give the question a chance and come up with some constructive thoughts.)

The answer is a variation on "can't beat greed". Ways need to be found
which make it in the companies best interest to make
serviceable/modifiable products.

To offer a fairly recent real world example of how this can happen,
have a look at the Linksys WRT54G router. Initially this was released
running Linux, then Linksys found that by going to a proprietary OS
they could reduce the cost per unit (using less memory than the Linux
based version). For sales through places like "Future Shop" knocking a
few dollars off the cost of each unit matter. On the other hand, the
Linux community showed there was enough of a market to support
continuing the modifiable version, thus we got the WRT54GL router...

So, bottom line, and a hard question, how do we show the companies
that people will pay a bit extra for modifiable/repairable hardware?
Or even harder question, how would we get governments to impose rules
re: making it possible to service hardware...

Colin.

> On Sun, Jul 25, 2010 at 06:43:04PM -0400, phiscock-g851W1bGYuGnS0EtXVNi6w at public.gmane.org wrote:
>>> We have an older Toshiba laptop that's been used by various members of
>>> the
>>> family and eventually ended up being used by my wife Dorothy. A year ago,
>>> the keyboard stopped working correctly, although the rest of the machine
>>> was functional and we could use a USB keyboard instead. I took the
>>> machine
>>> as far appart as I dared, looking for a bad connection - at one point I
>>> had most of it spread out on my bench - but couldn't find anything wrong.
>>>
>>> Dorothy got tired of this and bought a Mac. We put the Tosh in the
>>> basement.
>>>
>>> Last week, Dorothy left for a vacation in Walton, NS, a former mining
>>> town
>>> *away* out in rural Nova Scotia. On a whim, she took the Tosh with her to
>>> show to a local woman who is self trained in fixing computers. Marg is
>>> about 65 and 'started in computers' a few years ago. I was pretty
>>> skeptical this was going anywhere.
>>>
>>> Dorothy showed the computer to Marg and she very quickly determined that
>>> a
>>> couple of the function keys were pushed down and locked together. Using a
>>> paper clip, she freed the keys and the keyboard worked fine.
>>>
>>> Some people, like Marg, have a real talent for repairing things. And it's
>>> not necessarily those with an engineering degree ;).
>>>
> --
> Scott Sullivan
> --
> The Toronto Linux Users Group.      Meetings: http://gtalug.org/
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists
>
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists