Homebrew PCBs

Hi all,

I'm about 80% of the way towards getting my CNC mill up and running,
will eventually be able to feed it a JPEG/GIF/PNG file of the PCB
artwork, and it will cut the tracks, auto-detect the holes from the
artwork and drill those too, using a narrow-tapered conical diamond bit.

The completed unit will have a total desktop footprint barely larger
than a sheet of A3 paper.

All reasonable IMHO for a final parts cost circa US$150 all up.

The X-Y tracks have been a pain though. As someone who's very skilled
with software, reasonable with hardware but largely clueless with the
mechanical engineering, this has been a pain.

What I've got for each X and Y direction is an aluminium platform, with
lengths of aluminium angle bracket along the short edges, riding on top
of aluminium angle bracket, shown below as cross-sectional view, excuse
the rough ASCII graphics:

       **************************         PCB
     /////////////////////////////     chipboard to absorb drill
   ================================   upper platform
   +----                      ----+
   |              _               |
   | |           / \            | |     upper rail
   | |           \_/            | |
     |                          |
 ----+                          +-----  lower rail
 =====================================  lower platform

(circle in middle is threaded rod, driven by unipolar stepper motor).

With liberal spraying of CRC lubricant (will grease it later), the top X
and bottom Y platforms are moving freely enough for the stepper motors
to be able to turn the threaded rods (@ 1:1 gearing) through the full
movement range in both X and Y directions. Getting reasonable movement
speed, rotational/translational stability, plus precision of around 0.04mm.

But at some time, I would like to replace this crude abomination with
something similarly simple but better.

Can someone recommend better ways to do the tracks, using inexpensive
and easily available materials that don't need specialised tools to
deploy? Preferably materials which can be easily sourced in a small city
without 8 hours searching the Yellow Pages?

(The other extreme is ready-made precision linear bearing parts, but
these are hideously expensive within NZ).

All ideas appreciated.

-- 
Cheers
David

On Thu, 23 Jun 2005 15:52:47 +0200, David McNab <david@...>  
wrote:

>
> Can someone recommend better ways to do the tracks, using inexpensive
> and easily available materials that don't need specialised tools to
> deploy? Preferably materials which can be easily sourced in a small city
> without 8 hours searching the Yellow Pages?


Al on Al is bad, better make plastic on Al or even better plastic on steel.

ST

David McNab wrote:
> Can someone recommend better ways to do the tracks, using inexpensive
> and easily available materials that don't need specialised tools to
> deploy? Preferably materials which can be easily sourced in a small city
> without 8 hours searching the Yellow Pages?

Yes, I'd use sections of 1/2" steel drill rod which
can be found at machine tool suppliers (ENCO is a
good example) and they are likely cheaper than the
AL angles you are using.

Using the drill rod as a "ways" I'd use HDPE as
a bearing material.  If not readily available I'd
grab the cutting board from the kitchen.  You
can mill/drill out the bearing as a "clam shell"
from two mating pieces of PE relieved to encase
the drill rod.  You probably want to preload the
clam shell bearings to the rod to remove play.
Something like a strip spring should work well.

Depending upon the length of the drill rod used
you may need intermediate support.  If you keep
the profile of the support narrow relative to
the parallel axis of the drill rod the clam
shell bearing should be able to easily clear the
supports.  The supports ideally should mount to
the drill rod by threaded connection into holes
cut into the rod's bottom which do not penetrate
clear through the rod's diameter.

-- 
uhmgawa@...        www.gnu.org

Plastic on aluminum is a great bearing surface.  I would recomend UHMW 
(Ultra High Melecular wight) plastic.  You can purchase it at 
www.mcmaster.com and some woodworking stores carry it.  Woodworkers 
use it for making jigs and fixtures etc.

I used it for the "bearing" surfaces on the way system for my CNC 
router, works great and tolerates dust and chips really well.




--- In Homebrew_PCBs@yahoogroups.com, "Stefan Trethan" 
<stefan_trethan@g...> wrote:
> On Thu, 23 Jun 2005 15:52:47 +0200, David McNab <david@r...>  
> wrote:
> 
> >
> > Can someone recommend better ways to do the tracks, using 
inexpensive
> > and easily available materials that don't need specialised tools to
> > deploy? Preferably materials which can be easily sourced in a 
small city
> > without 8 hours searching the Yellow Pages?
> 
> 
> Al on Al is bad, better make plastic on Al or even better plastic on 
steel.
> 
> ST

Hi David,

As others have mentioned, al on al is bad. It is ok if it is lubricated 
but the lubricant will pick up abrasive dust and turn into grinding 
paste. If you don't lubricate it, it tends to stick and 'gall'. Drill 
rod/silver steel (depends where you are) with plastic bearings works 
well but even large diameter rod will flex an amazing amount if it isn't 
supported. My favourite is rectangular bright mild steel with roller 
skate bearings running on it. Bright mild steel is obtainable from most 
metal stockists and is cheap. You end up using quite a few bearings but 
they are dirt cheap so it isn't a problem.

If you want to really go to town then make some felt wipers to wipe off 
any dirt that may accumulate on the rails. Oiled felt is very effective 
at excluding dust.

Les


David McNab wrote:

>Hi all,
>
>I'm about 80% of the way towards getting my CNC mill up and running,
>will eventually be able to feed it a JPEG/GIF/PNG file of the PCB
>artwork, and it will cut the tracks, auto-detect the holes from the
>artwork and drill those too, using a narrow-tapered conical diamond bit.
>
>The completed unit will have a total desktop footprint barely larger
>than a sheet of A3 paper.
>
>All reasonable IMHO for a final parts cost circa US$150 all up.
>
>The X-Y tracks have been a pain though. As someone who's very skilled
>with software, reasonable with hardware but largely clueless with the
>mechanical engineering, this has been a pain.
>
>What I've got for each X and Y direction is an aluminium platform, with
>lengths of aluminium angle bracket along the short edges, riding on top
>of aluminium angle bracket, shown below as cross-sectional view, excuse
>the rough ASCII graphics:
>
>       **************************         PCB
>     /////////////////////////////     chipboard to absorb drill
>   ================================   upper platform
>   +----                      ----+
>   |              _               |
>   | |           / \            | |     upper rail
>   | |           \_/            | |
>     |                          |
> ----+                          +-----  lower rail
> =====================================  lower platform
>
>(circle in middle is threaded rod, driven by unipolar stepper motor).
>
>With liberal spraying of CRC lubricant (will grease it later), the top X
>and bottom Y platforms are moving freely enough for the stepper motors
>to be able to turn the threaded rods (@ 1:1 gearing) through the full
>movement range in both X and Y directions. Getting reasonable movement
>speed, rotational/translational stability, plus precision of around 0.04mm.
>
>But at some time, I would like to replace this crude abomination with
>something similarly simple but better.
>
>Can someone recommend better ways to do the tracks, using inexpensive
>and easily available materials that don't need specialised tools to
>deploy? Preferably materials which can be easily sourced in a small city
>without 8 hours searching the Yellow Pages?
>
>(The other extreme is ready-made precision linear bearing parts, but
>these are hideously expensive within NZ).
>
>All ideas appreciated.
>
>  
>

Les Newell wrote:
> Hi David,
> 
> As others have mentioned, al on al is bad. It is ok if it is lubricated 
> but the lubricant will pick up abrasive dust and turn into grinding 
> paste. If you don't lubricate it, it tends to stick and 'gall'. Drill 
> rod/silver steel (depends where you are) with plastic bearings works 
> well but even large diameter rod will flex an amazing amount if it isn't 
> supported.

Yes that's true and ideally the rod should be
mounted via a continuous support or discrete
supports at fixed intervals.

Commercial solutions typically use a continuous
extruded AL support.  If I had to fabricate an
approximation I'd probably use sections of 1/4"
AL bar of suitable height and mill the top to
receive the drill rod.  It would likely be preferable
to pre-bore the rod mounting holes in the bar
before milling the top relief.

> My favourite is rectangular bright mild steel with roller 
> skate bearings running on it. Bright mild steel is obtainable from most 
> metal stockists and is cheap. You end up using quite a few bearings but 
> they are dirt cheap so it isn't a problem.

What type of tolerances are you able to hold with this
approach?  At least with the mild steel of which I take
delivery the surface really can't compare with ground
drill rod.

-- 
uhmgawa@...        www.gnu.org

To tell the truth I haven't measured it. If you look closely at bright 
mild steel the slight roughness tends to be pits in the surface rather 
than bumps. Skate bearings give a fairly large contact area so they ride 
over the pits. Quickly running over the surface with a sharpening stone 
(I use a diamond hone) will take off any high spots and dings from 
handling. Although BMS isn't that accurate dimensionally any given piece 
will not vary much over it's length.

Les

uhmgawa wrote:

>What type of tolerances are you able to hold with this
>approach?  At least with the mild steel of which I take
>delivery the surface really can't compare with ground
>drill rod.
>
>  
>

For those that don't know, you can get those white plastic cutting boards at
the local Dollar+ stores.  They are cheap and will make decent sliders over
metal.

Robert
:)



As others have mentioned, al on al is bad. It is ok if it is lubricated 
but the lubricant will pick up abrasive dust and turn into grinding 
paste. If you don't lubricate it, it tends to stick and 'gall'. Drill 
rod/silver steel (depends where you are) with plastic bearings works 
well but even large diameter rod will flex an amazing amount if it isn't 
supported. My favourite is rectangular bright mild steel with roller 
skate bearings running on it. Bright mild steel is obtainable from most 
metal stockists and is cheap. You end up using quite a few bearings but 
they are dirt cheap so it isn't a problem.

If you want to really go to town then make some felt wipers to wipe off 
any dirt that may accumulate on the rails. Oiled felt is very effective 
at excluding dust.

Les



Be sure to visit the group home and check for new Links, Files, and Photos:
http://groups.yahoo.com/group/Homebrew_PCBs

If Files or Photos are running short of space, post them here:
http://groups.yahoo.com/group/Homebrew_PCBs_Archives/ 
Yahoo! Groups Links

David McNab wrote:
> But at some time, I would like to replace this crude abomination with
> something similarly simple but better.

Something else to keep in mind is using the XY
setup to assist with pick/place operations for
SMD devices.  Dealing with placement of any
significant number of 0603 or smaller components
per board can get tedious quite quickly.

Also even larger components such as high pin
count packages commonly have pin pitches of
0.5mm or less.  Accurate alignment by hand
here as well is tedious.

Another area where semi-precise placement would
be handy is dispensing of solder paste.  There
are methods which are somewhat forgiving of
the errors encountered in manual placement but
they usually involve a "check and fixup" step
which could be eliminated by accurate dispensing.

I don't foresee here the need for something as
elaborate as a commercial device feeding off
components dispensed via tape & reel, but rather
a tool allowing semi-automated placement of
paste and components onto a board.  The solder
paste could be dispensed via screw-driven syringe
substituting for the drill/mill head.  Component
placement could be made by a vacuum head able
to articulate at least +/-180* normal to the board.
The idea here is the vacuum head could under
operator control pick a component, rotate it to
position and place on the board into the previously
dispensed paste.

Getting this far would be cause for celebration but
a useful addition which quickly comes to mind would
be the addition of a small CMOS camera to the head.
This would dismiss the need to crouch over the
works donning a jeweler's loupe during the above
operations.

-- 
uhmgawa@...        www.gnu.org

These are the reasons why I am making provisions for extra room for more
tools on the head.  At first I was thinking of a tool carrousel, but I'm not
ready for that yet.  The rotation would add yet more 'slop'.

Instead, I'm going to distribute the tools as close to the center of the
work area as possible, and calculate an offset for each tool.  I believe
Mach 2 already has that feature.


"The solder paste could be dispensed via screw-driven syringe
substituting for the drill/mill head."

EXCELLENT idea!  I was racking my brains for a consistent technique of
applying paste.  I kept thinking motor and couldn't get out of that
'groove', a screw is perfect for a slow regular flow control.


"Component placement could be made by a vacuum head able
to articulate at least +/-180* normal to the board."

And another great idea!  I had figured the vacuum part, I was missing the
rotating head.  I already have a Pick and Place file exported from QCAD that
is just sitting there.  The rotation is already determined in the file, the
data is all there waiting to be used.


"Getting this far would be cause for celebration but
a useful addition which quickly comes to mind would
be the addition of a small CMOS camera to the head."

Now that one I already had covered, in fact, I'm going even one step further
with this one.  I have 2 USB cameras waiting by my desk (read pile of junk).
I was initially going to place only 1, but now I intend to place 1 camera on
each axis for a perfect view of alignment.  Small lamps will come handy,
maybe high intensity LEDs.

Check EBay for USB miniature cameras.  The resolution might not be
excellent, but I've bought several magnifying devices through EBay and I'm
going to test adding a DIY mag lense on the camera.

And the last tool remaining would be the via/trackpin gizmo and I've got my
dream machine.

Robert
:)


Be sure to visit the group home and check for new Links, Files, and Photos:
http://groups.yahoo.com/group/Homebrew_PCBs

If Files or Photos are running short of space, post them here:
http://groups.yahoo.com/group/Homebrew_PCBs_Archives/ 
Yahoo! Groups Links

On Fri, 24 Jun 2005 18:05:22 +0200, uhmgawa <uhmgawa@...> wrote:

>
> Something else to keep in mind is using the XY
> setup to assist with pick/place operations for
> SMD devices.  Dealing with placement of any
> significant number of 0603 or smaller components
> per board can get tedious quite quickly.


I thought about it a while ago, but haven't got time/means right now.

One of the (i think) better ideas then was to put the components in a  
pickup area.
Few of us will have reels for most parts, but most will have short strips.
If you put little pins on the pickup area where the holes in the tapes  
line up, that could work. You need to peel off the covering tape  
beforehand.

For other components like ICs that you don't have in strips one might mill  
out small indentations in a plate.
of course this would also work for smaller parts.

This would present you the components in a known position/orientatation.
All you need to do then is write a software that tells you which position  
to put which strips, and increment accordingly the parts over several  
boards. then you only need to somehow insert the right positions in the  
pick-and-place file.

Just an idea.

ST

Robert Hedan wrote:
> "The solder paste could be dispensed via screw-driven syringe
> substituting for the drill/mill head."
> 
> EXCELLENT idea!  I was racking my brains for a consistent technique of
> applying paste.  I kept thinking motor and couldn't get out of that
> 'groove', a screw is perfect for a slow regular flow control.

The syringe should be as rigid as possible so the
post-drive flow is minimized.  I'd choose something
like a stainless tube and plunger.  The seal on the
plunger can simply be an o-ring.

Then again a minimal amount of post-drive dribble can
be handled in the control code by backing off a tad
on the screw drive at the end of a dispensing run and
pre-loading by the same amount when the next run begins.

Some thought is needed in how the syringe is loaded
with solder paste to avoid inclusion of air pockets.
Loading via reverse vacuum through the dispensing tip
might work depending upon the viscosity of the paste
but may also be slow.  If another means of creating a
larger orifice to load the paste is provided it will be
necessary to initially purge the air from the
dispensing path.  A few bubbles may be output during
the dispensing process but you will notice this and
be able to correct such.

> "Component placement could be made by a vacuum head able
> to articulate at least +/-180* normal to the board."
> 
> And another great idea!  I had figured the vacuum part, I was missing the
> rotating head.  I already have a Pick and Place file exported from QCAD that
> is just sitting there.  The rotation is already determined in the file, the
> data is all there waiting to be used.

I don't know how much a pick and place file will buy
here as the operation is going to be manual unless
component pickup is from an automated dispenser (eg:
tape and reel).  Even then commercial solutions include
machine vision to align the component on the vacuum head.

I'm just looking for something to pick up a 0603
capacitor out of a dish and give me an accurate means
to manually locate it on the board.  There is some
tolerance for placement inaccuracy in reflow soldering
as the surface tension of the solder paste upon
liquefaction will pull the component into alignment
with the pad.

> "Getting this far would be cause for celebration but
> a useful addition which quickly comes to mind would
> be the addition of a small CMOS camera to the head."
> 
> Now that one I already had covered, in fact, I'm going even one step further
> with this one.  I have 2 USB cameras waiting by my desk (read pile of junk).
> I was initially going to place only 1, but now I intend to place 1 camera on
> each axis for a perfect view of alignment.  Small lamps will come handy,
> maybe high intensity LEDs.

I'm unsure how much usable resolution you may get
with USB cameras.  I'm also uncertain if standard
NTSC resolutions will be enough though I haven't yet
experimented.

Low power microscope cameras are available which
render NTSC which is both encouraging and a possible
solution as well.  10x seems to be more than enough
for 0603 work and likely 0402 as well.

> Check EBay for USB miniature cameras.  The resolution might not be
> excellent, but I've bought several magnifying devices through EBay and I'm
> going to test adding a DIY mag lense on the camera.

Try a jewelers loupe.  They are both inexpensive and
available in suitable magnification powers.

If actually available I'd prefer the greater resolution
and use an NTSC frame grabber if I needed to get image
data into a computer.  The advantage of NTSC is also
that you don't need a PC to decompress the video and
can simply use a standard analog monitor.

-- 
uhmgawa@...        www.gnu.org

"The syringe should be as rigid as possible so the
post-drive flow is minimized.  I'd choose something
like a stainless tube and plunger.  The seal on the
plunger can simply be an o-ring."

I was thinking an el-cheapo syringe, very easily accessible, like those used
to refill printer ink.  The metal needle has a fine hole, and can be removed
for filling up (larger diameter hole).

Someone mentionned that the paste will run easily when warmed.  What about
filling up the syringe with paste in liquid form, turn it upside down and
store at a temperature where the paste turns to the right 'thickness'?  Time
for a small fridge for the lab?  Please say yes, I need it to cool the beer.


"post-drive dribble can be handled in the control code by backing off a tad
on the screw drive at the end of a dispensing run"

Good idea!


"I don't know how much a pick and place file will buy
here as the operation is going to be manual unless
component pickup is from an automated dispenser (eg:
tape and reel).  Even then commercial solutions include
machine vision to align the component on the vacuum head."

I was planning on using a 1/2 machine 1/2 man solution.  :)

I'd preload a 'cartridge' with the components so that the pick head only has
to lift the part, polarity for capacitors, pin #1 for ICs and such would be
done by hand.  The cartridge would basically be a thin rail tilted at an
angle so that the components slide down towards the pick spot.  Maybe a
small vibration at the far end to move parts down (small rotating cam?).

One note, I was only considering SMD components with this application.
Bending wire gets WAY too complicated.


"I'm just looking for something to pick up a 0603
capacitor out of a dish and give me an accurate means
to manually locate it on the board."

WAY too difficult for me, hence the idea of manually loading cartridges.


"There is some tolerance for placement inaccuracy in reflow soldering
as the surface tension of the solder paste upon
liquefaction will pull the component into alignment
with the pad.

Yeah, I'm counting on that.  :D


"I'm unsure how much usable resolution you may get
with USB cameras.  I'm also uncertain if standard
NTSC resolutions will be enough though I haven't yet experimented.
Low power microscope cameras are available which
render NTSC which is both encouraging and a possible
solution as well.  10x seems to be more than enough
for 0603 work and likely 0402 as well."

All things I'll have to test.  Putting a magnifying lense on the USB camera
was my first option (read el cheapo).  If the resolution is not acceptable,
then I'll see what is available at that time.  I have an old 8mm video
camera that chews tapes, I could jam the 'tape detector switch' and use
that.  I could even use the zoom feature by hotwiring the contacts.  The
more I think about it, the more I like this idea.  I'll have to check how
close I can focus.


"Try a jewelers loupe.  They are both inexpensive and
available in suitable magnification powers."

Yup, that's part of the el cheapo USB camera solution.


"The advantage of NTSC is also
that you don't need a PC to decompress the video and
can simply use a standard analog monitor."

I will have a PC dedicated to this task anyways.  It will run only Mach 2 so
the CPU is not bothered by other tasks.  I'm setting up another desktop for
all my other development software; QCAD, Visual BASIC, MPLAB IDE, PBP
Compiler, PIC serial programmer, etc.  Right now I'm doing all that on my
laptop, and that's not efficient use of a laptop.

Robert
:)


Be sure to visit the group home and check for new Links, Files, and Photos:
http://groups.yahoo.com/group/Homebrew_PCBs

If Files or Photos are running short of space, post them here:
http://groups.yahoo.com/group/Homebrew_PCBs_Archives/ 
Yahoo! Groups Links

Robert Hedan wrote:

> Someone mentionned that the paste will run easily when warmed.  What about
> filling up the syringe with paste in liquid form, turn it upside down and
> store at a temperature where the paste turns to the right 'thickness'?

I believe the issue is maintaining the viscosity of
the paste medium in order to keep the solder particles
uniformly suspended.  Reducing the viscosity to the
point of having a pourable paste may compromise the
suspension uniformity.  I'd check with the vendor of
the particular paste to be used as they can offer
exact guidance.

-- 
uhmgawa@...        www.gnu.org

Stefan Trethan wrote:
> One of the (i think) better ideas then was to put the components in a  
> pickup area.
> Few of us will have reels for most parts, but most will have short strips.
> If you put little pins on the pickup area where the holes in the tapes  
> line up, that could work. You need to peel off the covering tape  
> beforehand.

I'd imagine tape reels are a non-issue for the type
of prototyping under discussion here.  As you suggest
a far more common scenario is dealing with cut tape
segments.  But even here I don't see the advantage
to gradually pull back the tape as components are
consumed nor much of an overall advantage to peeling
off the tape from a strip and feeding from the exposed
carrier.  It is likely for some unused components to
remain and you'll need to find another home for them
anyway.

I've just abandoned the tape carriers all together and
strip components wholesale from them over a funnel which
feeds into a small ziplock bag.  I can dump whatever I
need into a tray for placement and the excess goes back
to the bag via funnel.  The bag also offers a much more
hospitable means to identify component values compared
with the carrier tape.

> For other components like ICs that you don't have in strips one might mill  
> out small indentations in a plate.
> of course this would also work for smaller parts.

These days prototype IC count tends to be far lower
than days gone by.  Such that I don't have much of
an issue just sitting the packages on a foam slab
prior to placing.  Alternatively if you need to
prebake the packages before reflow soldering to avoid
popcorn damage a metal tray works as well.

-- 
uhmgawa@...        www.gnu.org

--- In Homebrew_PCBs@yahoogroups.com, Robert Hedan <robert.hedan@v...>
wrote:
> These are the reasons why I am making provisions for extra room for more
> tools on the head.  At first I was thinking of a tool carrousel, but
I'm not
> ready for that yet.  The rotation would add yet more 'slop'.
> 
> Instead, I'm going to distribute the tools as close to the center of the
> work area as possible, and calculate an offset for each tool.  I believe
> Mach 2 already has that feature.
> 
> 
> "The solder paste could be dispensed via screw-driven syringe
> substituting for the drill/mill head."
> 
> EXCELLENT idea!  I was racking my brains for a consistent technique of
> applying paste.  I kept thinking motor and couldn't get out of that
> 'groove', a screw is perfect for a slow regular flow control.
> 

I've often wondered why not use a cnc machine to cut a paste stencil?
 Bascially, its just a couple mil thick piece of plastic sheeting that
you squeegee (sp?) the solder paste over.  A stencil is pretty easy to
use - I watched a guy prep a board for reflow in about 5 minutes with one.

"I've often wondered why not use a cnc machine to cut a paste stencil?
Bascially, its just a couple mil thick piece of plastic sheeting that you
squeegee (sp?) the solder paste over.  A stencil is pretty easy to use - I
watched a guy prep a board for reflow in about 5 minutes with one."

Yup, the machine could easily do that.  That could be an alternate solution,
but an automated process is soooo much easier.  While the machine is
distributing the paste, you can be loading the cartridges with parts.  No
wasted time dismounting the board for pasting, and remounting for placing,
assuming I figure an adequate picker/placer.

I'd also much prefer to spend those 5 minutes doing something else while the
machine is pasting the board, even if it's watching the machine work with my
3 girls.  The nice thing about a mechanical process is the repeated
precision, something my shaking hands, bad eyes and quick loss of
concentration are not helping with.

The doc has finally figured what's wrong with me; sleep apnea (you'll notice
I post at all times of day, 'cause I stay up 4 hours, sleep 2 hours, and so
on).  I stop breathing on average 25 times per hour, that's something like
every 2 minutes or so.  I never get a full night's rest and my coordination
and concentration are NOT improving as time goes by.  That's why I want to
automate as much of the process as possible.  Ideally, I'd like it so even
my wife could follow a checklist of operations and do the work when I'm
sleeping.

Robert
:)




Be sure to visit the group home and check for new Links, Files, and Photos:
http://groups.yahoo.com/group/Homebrew_PCBs

If Files or Photos are running short of space, post them here:
http://groups.yahoo.com/group/Homebrew_PCBs_Archives/ 
Yahoo! Groups Links

On Sat, 25 Jun 2005 03:00:11 +0200, uhmgawa <uhmgawa@...> wrote:

>
> I'd imagine tape reels are a non-issue for the type
> of prototyping under discussion here.  As you suggest
> a far more common scenario is dealing with cut tape
> segments.  But even here I don't see the advantage
> to gradually pull back the tape as components are
> consumed nor much of an overall advantage to peeling
> off the tape from a strip and feeding from the exposed
> carrier.  It is likely for some unused components to
> remain and you'll need to find another home for them
> anyway.
> I've just abandoned the tape carriers all together and
> strip components wholesale from them over a funnel which
> feeds into a small ziplock bag.  I can dump whatever I
> need into a tray for placement and the excess goes back
> to the bag via funnel.  The bag also offers a much more
> hospitable means to identify component values compared
> with the carrier tape.

I was thinking you cut off the required amount of parts for the number of  
boards you intend to make, and put that on the machine. I use the strips  
extensively, and like it much more than loose components. The strips can  
be written on, and the parts are in little drawers, so i see no need to  
further identify them.


>
>> For other components like ICs that you don't have in strips one might  
>> mill
>> out small indentations in a plate.
>> of course this would also work for smaller parts.
> These days prototype IC count tends to be far lower
> than days gone by.  Such that I don't have much of
> an issue just sitting the packages on a foam slab
> prior to placing.  Alternatively if you need to
> prebake the packages before reflow soldering to avoid
> popcorn damage a metal tray works as well.

What would that help? sitting them on foam would require you to build some  
kind of machine vision, which is a vast amount more difficult than just  
pick here place there.

ST

All the parts I receive on strips of paper predate the stone age.  The glue
has turned to amber and the paper is SUPER stiff.  I swear I saw a mosquito
from the age of the dinosaurs trapped in the amber.  I've soaked some parts
for days, only to get a rusted out bowl and I STILL had to scrap the crap
off the leads with my finger nails.

I prefer to buy in bulk or those modern plastic SMD strips, then I just
transfer into my storage cabinets with a label on each drawer.

Robert
:(



Be sure to visit the group home and check for new Links, Files, and Photos:
http://groups.yahoo.com/group/Homebrew_PCBs

If Files or Photos are running short of space, post them here:
http://groups.yahoo.com/group/Homebrew_PCBs_Archives/ 
Yahoo! Groups Links

On Sat, 25 Jun 2005 08:39:01 +0200, Robert Hedan  
<robert.hedan@...> wrote:

> All the parts I receive on strips of paper predate the stone age.  The  
> glue
> has turned to amber and the paper is SUPER stiff.  I swear I saw a  
> mosquito
> from the age of the dinosaurs trapped in the amber.  I've soaked some  
> parts
> for days, only to get a rusted out bowl and I STILL had to scrap the crap
> off the leads with my finger nails.
> I prefer to buy in bulk or those modern plastic SMD strips, then I just
> transfer into my storage cabinets with a label on each drawer.
> Robert
>


Then you have a bad supplier.
I get about 70% in paper and the rest in plastic strips, and never had  
problems with the paper. (tho it is harder to find the end of the cover  
tape and i often end up splitting the paper instead ;-).
Both paper and plastic could be used to feed a CNC.

ST

Phil wrote:
> I've often wondered why not use a cnc machine to cut a paste stencil?
>  Bascially, its just a couple mil thick piece of plastic sheeting that
> you squeegee (sp?) the solder paste over.  A stencil is pretty easy to
> use - I watched a guy prep a board for reflow in about 5 minutes with one.

Stencils are really a production assembly tool used
where volumes of boards are being produced.  The
amount of solder paste waste/contamination otherwise
is excessively high.

For contemporary solder paste formulations, once the
paste is brought to room temperature and handled in a
stencil screening process the shelf life is drastically
reduced.  With the cost of solder paste in prototype
quantities somewhere between $25-40 per 250gm a more
conservative approach is quite attractive.

-- 
uhmgawa@...        www.gnu.org

Stefan Trethan wrote:
> What would that help? sitting them on foam would require you to build some  
> kind of machine vision, which is a vast amount more difficult than just  
> pick here place there.

Indeed -- I'm that machine.

My interest wasn't in fully automated pick and
place.  Rather a tool to allow more accurate
manual placement.  Even in the case of feeding
components from tape/carrier there are situations
where enough play exists in the carrier well to
cause significant variation between successive
component picks and some form of additional
alignment/registration is needed.

Especially in the case of ICs which will be
reflow soldered, it makes sense to keep them
discrete.  If the packages have any moisture
content they require baking before reflow
to avoid popcorn damage.  And the number of
baking cycles per unit should be minimal
(eg: 1) to avoid tarnishing the leads.  So I
pre-pick all ICs destined for a single board
onto a metal tray for preassembly baking.

I don't disagree keeping components in tape/
carrier won't buy locating precision.  But the
overhead of mounting and unmounting bits of
tape or IC carrier in a prototype scenario
doesn't seem to be worth the effort.

The scope of the tools being discussed here I'd
hazard apply to the >10 unit prototype scenarios.
While it is a common characteristic of a well
engineered design to scale beyond the original
application it is unlikely needed in this case.
For higher volume assembly runs it would typically
make more sense to farm-out the work to an
assembly shop.

-- 
uhmgawa@...        www.gnu.org

>With the cost of solder paste in prototype
>quantities somewhere between $25-40 per 250gm a more
>conservative approach is quite attractive.

So you believe this is the same thing board manufacturers use in bulk 
hence more affordable to them? I doubt it.


--- In Homebrew_PCBs@yahoogroups.com, uhmgawa <uhmgawa@m...> wrote:
> Phil wrote:
> > I've often wondered why not use a cnc machine to cut a paste 
stencil?
> >  Bascially, its just a couple mil thick piece of plastic sheeting 
that
> > you squeegee (sp?) the solder paste over.  A stencil is pretty 
easy to
> > use - I watched a guy prep a board for reflow in about 5 minutes 
with one.

> 
> Stencils are really a production assembly tool used
> where volumes of boards are being produced.  The
> amount of solder paste waste/contamination otherwise
> is excessively high.
> 
> For contemporary solder paste formulations, once the
> paste is brought to room temperature and handled in a
> stencil screening process the shelf life is drastically
> reduced.  With the cost of solder paste in prototype
> quantities somewhere between $25-40 per 250gm a more
> conservative approach is quite attractive.
> 
> -- 
> uhmgawa@m...        www.gnu.org

derekhawkins wrote:
>>With the cost of solder paste in prototype
>>quantities somewhere between $25-40 per 250gm a more
>>conservative approach is quite attractive.
> 
> 
> So you believe this is the same thing board manufacturers use in bulk 
> hence more affordable to them? I doubt it.

It is the same solder paste formulation.  Otherwise
why sample it for pre-production/prototype qualification?

Note the amount of solder paste consumed in a printing
operation is relatively small compared to the bulk
required to load the stencil before the print is made.
New paste is added to the bulk as it is consumed.

The process of warming the paste to room temperature
as well as the printing operation results in gradual
degradation.  This is acceptable in a production
operation where the waste is amortized amongst a series
of boards but inefficient for prototype work.

-- 
uhmgawa@...        www.gnu.org