From: "Christopher R. Carlen"
Subject: Re: 1523 joints
Date: Tue, 22 Oct 2002 17:58:50 -0700
Organization: Sandia National Laboratories, Albuquerque, NM USA
NNTP-Posting-Date: Wed, 23 Oct 2002 00:57:46 +0000 (UTC)
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20020826
X-Accept-Language: en-us, en
Ken Taylor wrote:
> Hi Christopher. Without having too much of a dig at you, should you have
> considered the manufacture stage during the design stage? Is the board
> suited to reflow? If it really has to be hand-assembled and soldered you may
> need to take a look at a few assembly shops and their work to make yourself
> happy with their output. Outsourcing can work but it takes a fair bit of
> preparation, investigation and ultimately confidence on both sides. Can you
> provide a fully assembled one which you *are* happy with so they have a
> model? What about if you maybe only get one side done externally, leaving
> the fiddlier bit on the 'non-active' side done internally, so that maybe an
> automated process can be used for the bulk of the joints.
> What are the chances though of taking one or more of your lab tech's 'under
> your wing' and bringing them up to your standard? They have to learn
> somewhere, and if they currently are not up to spec then here's your chance
> to improve things! :)
Thank you for your response. I will try to respond to all your valuable
I didn't consider manufacturing processes because I planned to make a
small quantity only, about 4-10 or so. For this and for most of the
jobs I do this is the case, and so everything I do has been hand
assembled. For that reason as well, I know very little about
manufacturing requirements or processes, having little need to know
about them. Now however, I am interested in learning if a manufacturing
route might save me time. That would require that all the effort needed
to provide the manufacturer with everything they need to do the job,
works out to less effort than if I just do it myself.
I am planning to hand assemble a fully working unit myself, since of
course the design needs to be fully tested once implemented on the PCBs.
The shop in fact asked me to provide them with a model, so assuming
the first board works as I wish, I can provide it to the shop as a
model. I would also intend to provide them with a considerable list of
notes and "sanity checks".
The idea of having only the one side done has occurred to me. I have to
think about that further, particularly regarding the proportions of the
total effort that I think are associated with each side.
Now for your final idea, this is an interesting subject. I will tell
you some background. When I came here, the department had an
electronics tech. with superior chassis assembly skills. The man's
instruments are just beautiful to look at. But he was admittedly not
very knowledgable about electronic circuit theory, and so the types of
circuits that were considered for custom design here were very limited.
Consequently, there was sort of a backlog of old ideas, a wishlist so
to speak, on the minds of the scientists, but nobody around to make them
When I was hired, I thought I was going to be a laser tech. But after
graduating with my Chemistry degree, and having gotten bored with that
subject, I began delving back into my old favorite hobby, electronics.
Only this time I was armed with lots of mathematical tools from my B.S.,
and the willingness to use them. This attitude quickly propagated to my
job where I found that I could have a lot of fun learning about
electronic design by expanding my job from just lasers to involve all
those wish list projects that had been on the back burner. Of course,
since I have gotten into this field, I have now begun creating new lists
of ideas for the scientists to wish for.
Then our electronics tech. decided to leave, and I inherited his job.
So now I have the duties of two jobs, unfortunately, since we can't
afford to replace him at the moment.
The other thing I did when I came here was learn about making printed
circuits, since at the time everybody wire-wrapped things, and after my
first crack at that I quickly decided that once was enough, and PCBs
were the way to go. There is one other tech here now who is competent
at building stuff, but he's booked with the duties of his dept. There
are a few other electronics techs. as well but they aren't near prepared
to tackle fine electronics.
Which leads to my basic attitude about this, which is, that this is a
national laboratory, and it shouldn't be necessary to train people here.
I mean, the kind of people who should be here in the first place are
those who learn by themselves, and hold themselves to impeccable
standards. Unfortunately the reality isn't so rosy. And to make
matters really strange, the scientists and managers I work for are all
mechanical engineers, who have little idea about what I really do. They
don't know the difference between designing a PLL and putting a
connector on the end of a wire. And since putting connectors on the end
of wires is what most techs. do here, it can be rather frustrating to
hope for any sort of understanding from my superiors. Fortunately, the
final results are what really count, and so I have gotten my proper
rewards in the end. Thus I am not complaining, but rather laying the
groundwork for you to understand why I have been offered some very funny
propositions, basically to the effect of having mechanical techs. help
me to assemble electronics, and even to align sophisticated laser
systems. Mech. techs. without an iota of physics knowledge, that is.
Our experiments with those ideas have been terrible failures, leading me
to be very averse to the idea of continued attempts to train other
techs. to do this sort of work. Think about it, someone who doesn't
know *anything*, has to be brought through *everything* needed to
assemble, what end of a diode is what, how to identify every sort of
part (increasingly difficult today with SMT, unless one has the
experience of years of doing this stuff, that makes it all just intuitive).
I would be perfectly willing to hear your criticisms of this attitude,
but I have become somewhat demanding when it comes to standards. Thus I
think that in a leading research establishment a tech. who wants to do
something new should figure it out on his/her own, then present their
willingness to apply it to a project or even volunteer to solve some
problem, and then simply show up one day with the results in order to
prove their competence, as opposed to sinking the resource of time,
which is our most valuable commodity, by having another tech. train
them. That's what I've done. That's what the scientists have to do.
Nobody tutors them on tensor calculus. They just open a textbook, and
learn it. Then they use it.
Thanks again for the input, and...
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA