[TriEmbed] Reasonable PCB Asembly

[Adam Haile]

John,

Thanks so much for that awesome response. Even if we never get to the
necessary volumes, it is great to have a better idea of how that all works.
I've gotten a number of quotes on such things in the past and it was never
super apparent to me where they actually got the numbers from.
Yes, we were definitely looking into SMT for the hopeful cost savings, but
maybe we need to start smaller first... this particular project was not
exactly of trivial size. The intent was also to provide a (mostly) complete
product, but we may just have to stick with through-hole based kits for the
time being. Or at least ones with only a couple SMT parts that we can maybe
solder on ourselves and have the rest be through-hole.

-Adam


On Thu, Jun 12, 2014 at 9:49 AM, John Vaughters <jvaughters04 at yahoo.com>
wrote:

> Adam,
>
> Having worked 10 years in Board Assembly as an Engineer where I ended my
> career in that business quoting assemblies in a technical sales role, I
> would like to explain this topic a little more accurately. I have not been
> in this business in over 10 years, but other than things getting better and
> more cost effective, it is doubtful much has changed.
>
> PCB manufacture can be boiled down to two major factors. Non-Recurring
> Engineering (NRE) and cost per square inch of the board. This may seem like
> an over simplification, but it is not. I will outline the details of the
> two below and that is where it gets complicated.
>
> NRE
>
> This is the cost that the manufacturer requires to put the board through
> their shop. NRE can vary greatly. Soft or hard tooling can make a big
> difference. Hard tooling allows you to build more boards with the tool than
> soft tooling. This will bring down your cost per board if the volume
> justifies it. Manufacture characteristics limits how cheap their tooling
> can be. In other words you can reduce this cost only so much before it is
> limited by their process. This is one of the reasons why board shops can
> vary widely on cost based on volume.
>
> It consists of the following:
>
> * Tooling required to run board in their machines
> * Engineering time/cost to create the programs/tools needed for manufacture
> * General Fudge factor, usually for profit
>
> Price per Square inch
>
> All board manufacturers calculate this number to tailor it to their
> process. They will give you this number off the top of their head and that
> is usually the highest of their capability. It truthfully depends on how
> your board is #panelized to optimize their machines. If your board is too
> big, panelization will waste material. Although this problem goes away as
> the board gets smaller, because you generally can panelize a small board
> well.
>
> # Panelized boards are boards ganged together in larger standardized
> panels to be cut apart later. The shop will have different standardized
> panel sizes, but it is based on their machines.
>
> This cost consists of the following:
>
> * Material
> * Labor
> * Machine time
>
> Summary
>
> There really is no way around these costs structures, although the really
> fast turn around folks are dropping their NRE with tooling that is rock
> bottom and that is why we see some of these fast turn around folks able to
> survive. Keep in mind that their tooling is not meant to last at all. This
> is great for the market, becasue it allows peole like us to get boards for
> reasonable costs, becasue the traditional PCB shops would price us out with
> NRE. They are not trying to screw you, it is just the nature of their
> process that limits the bigger manufacturers to helping the small runs.
>
> The pricing with NRE built into the board looks like this
>
> price per board = NRE/(Board Volume) + (Total standard panel square
> inches)*(price per square inch)/(Board Volume) + (Misc costs)/(Board Volume)
>
> An alternative pricing method that is hardly used is to charge NRE as a
> lump sum and price just the board manufature. Often we included the quote
> this way becasue it shows the customer the breakdown of the cost and it
> will show the customer why low volumes are expensive. Incidently this is a
> common way plastic molding is priced where tooling is very expensive.
>
> What this means to TriEmbed?
>
> Probably not a whole lot. It is doubtful most of us will reach the volume
> that matters. Just in case, I wanted those interested to be aware of how it
> works. For the most part you will be given some minmum build Sq. In. and a
> price.
>
> One thing I do want to point out and if you have made it this far without
> falling asleep, then hopefully this helps.
> The reason electronics has gotten so cheap is that the reduction in raw
> materials of Surface Mount (SMT) is significant. It will not take you long
> to start shopping around to find out that SMT gets your cost down greatly.
> The only case this fails is in Hobbyland where SMT can cause more scrap in
> assembly than you save. Many of you out there have already felt that pain.
>
> So if you are serious about a product, finding the breakpoint of cost
> saving and volume is not trivial and I wish you luck in your endeavor.
> Please keep us updated on your success.
>
> Thanks,
>
> John Vaughters
>
>
>
>
>
> _______________________________________________
> Triangle, NC Embedded Computing mailing list
> TriEmbed at triembed.org
> http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
> TriEmbed web site: http://TriEmbed.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20140613/73291bf4/attachment.htm>