Re: Future: 0603 versus 0402 parts

On 25 Mar 2007 13:34:48 -0700, "Frank Raffaeli"
<SNIPrf_man_frTHIS@xxxxxxxxx> wrote:

On Mar 25, 5:15 pm, John Larkin
<jjlar...@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
On Sun, 25 Mar 2007 11:55:36 -0700, MassiveProng

<MassivePr...@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
On Sun, 25 Mar 2007 10:23:41 -0800, John Larkin
<jjlar...@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> Gave us:

I'm growing positively fond of BGAs.

But don't know the first thing about cleaning the assemblies
correctly, or economically!
You are one of those idiots that glosses over all the tasks you
consider menial.

I already said our production manager is a genius for processes. I
have enormous respect for all our production and process people, who
*are* better than me at this sort of stuff. I like them, too.

Do you like anybody? Doesn't look like it from here.



IT IS BROKE, YOU SHOULD FIX IT.

Both your process and your fucktard mentality!

What's wrong with 100% yield? Can you do better?

We build and ship stuff. We send invoices. People send us checks, and
occasionally plaques and letters of appreciation [1]. The system
works. It's the end of the fiscal year, bonus time! I share the
profits with the people who "broke" our processes.

John

[1] and some really cool Dragon Lady posters and patches from the
Skunk Works.

In 1988-90 I was fortunate to be a part of a 5-man team the developed
the Ford car radio for model years 1992-1996, for sale in North
America. Ford was fanatical about process control, and also about
training it's designers. I'm not a big fan of training courses, but it
wasn't presented to me as an option.

Myself and my colleagues trained with Dr. Genichi Taguchi and Dr. Jake
Fremenko. They both defined a process as follows:
"A process is a *repeatable* and *quantifiable* construct of
procedures, formulae and tactics that result in a realiable finished
assembly".
Ford built 114,000 radios per week in Brazil. Only 4/1000 field
complaints (these included things like "my cassette got stuck"). The
process managers at the plant in Brazil would recommend a shut down
for diagnosis if there were more than 3 final-test failures per week.

From your description, it sounds like your process meets the criteria.

Frank

In addition to engineer-created documentation (drawings, procedures,
eco's) my mfg people keep a parallel database of manufacturing notes
for things they build. We also allow them to restructure BOMs for best
fit to their processes and procedures. For example, they can redefine
what's on various levels of subassembly, or can create a BOM and dash
number for a partial assembly to be sent out to an assembly house,
then stocked and, when orders come in, finished in-house into one of
several official dash-numbered versions. We buy them all the toys they
want.

I don't think we'd do very well on a ISO9000 audit, because a lot of
what they do is undocumented practices and plain skill.

As long as the result conforms to our drawings and tests OK, we don't
tell them how to make it. We do meet with them before designs are
solid, to get their input on packaging, cabling, manufacturability and
such. They have come up with some very good stuff.

I worked for more than one company where engineering and manufacturing
were camps at war. In one case, engineering made drawings that were as
prim and complex as possible, an unreadable mess of flag notes, find
numbers, sections, and dash-number exceptions, so that they could call
manufacturing stupid. Manufacturing, if they spotted an error on a
drawing, would build as many as possible, as fast as possible, just
like that.

John


.