[Gnuk-users] Failure mode of FST-01

[NIIBE Yutaka]

Thank you dkg for this attempt of clarification.

On 10/02/2015 12:54 PM, Daniel Kahn Gillmor wrote:
> On Thu 2015-10-01 02:24:25 -0400, NIIBE Yutaka wrote:
>> During preparation of NeuG on FST-01 devices (for shop.fsf.org), I
>> encountered a problem:
>>
>>      FST-01 w/ tube from Seeed Technology doesn't work at all.
> 
> yikes!  I think you're saying this happened only on one instance of this
> hardware, right?

Yes, exactly.  My wording is not good, since Japanese language don't
distinguish plural.  And the number I cited was also wrong.

I have experience to do writing firmware for approximately 300 FST-01.
I encountered this problem for only one instance of FST-01.

For your information: While I handled 300 instances by myself (more
than half are for shop.fsf.org), Seeed sold approximately 350 FST-01.

> you're not saying that all the FST-01 w/ tube from Seeed are bad,
> are you?

No, only one instance.

>> Today, I put soldering iron to the XTAL and let the solder under XTAL
>> flow.  Then, the FST-01 is back and works well again as Gnuk Token.
> 
> great, this sort of fix sounds not too scary as hardware fixes go.  Have
> you let Seeed know about the problem?  do they have any proposed
> remedies?

The fix procedure is simple.  Here is a picture to explain how to do
that.  Put soldering iron on the XTAL part and wait for melt of
solder.


                  ..-' Head of Soldering Iron
             ...''    ,
             ------'''
          +------------+
          |    XTAL    |
          +------------+  <---- Watch the solder melting.
         ooooo      ooooo
     ...--------------------------------...

           Printed Circuit Board
     ...--------------------------------...


Now, China celebrates the National Anniversary, so they are on
vacation.  Sure, I share my experience with them.

My test plan for the production of FST-01 included the test of XTAL.
So, the instance should have worked in the production process.  I
guess that the solder connection was somehow weaker in the production,
and the breakage happened during shipment to Japan.

Actually, in the procurement process (also done by Seeed), I accepted
Seeed's proposal for cheaper XTAL, which is exactly compatible to my
original choice of part.  I should have checked if recommended solder
pattern is same or similar, but most of cases, same size part has same
recommended solder pattern.

Well, quality control for physical product is probability thing.  It's
a bit different than software.
--