Board building guide - SMD soldering
drasko
I was working over the weekend on MicropendousX board soldering with
my friend (Uros, also active on this list), and we almost lost our
eyes with tiny parts hand-soldering through all the night... It's very
complicated !
So, I was wondering - is there some easier way to do it and what are
the best practices ?
I was surfing around for informations, and link I got on this forum
about Project Ubertooth looks very useful, especially Building Guide :
http://ubertooth.sourceforge.net/hardware/build/. I think this
MocroopendousX, which is a wonderful project, deserves some kind of
detailed and precise build guide, with explanations and tutorials how
to struggle with small SMD components soldering.
I have found this Sparkfun's tutorial very useful :
http://www.youtube.com/watch?v=WDIqtGMROjM, and I this led me also to
the hot-plate solution : http://www.youtube.com/watch?v=Uov0SPHKcnk
for which even home iron (for clothes itoning) can be used.
First question is there - can we order the MicropendousX stencils (for
example with PCB boards over the BatchPCB).
Second question is troubling two-sides component mount. For the
examples I have found/shown all the components were on one side, so we
can place the bord itself on the hot-plate ot hot iron surface. But
how to mount processor which is from the bottom side of MicropendousX
board ?
Second question thus is - can somebody explain or write tutorial how
this two-sided SMD soldering should be done elegantly for this board ?
The only reasonable solution I have found is applying epoxy glue for
each component (http://www.youtube.com/watch?v=6M3MeADb1dE), but that
would imply oven (which I don't have, I want to use home iron) and
gluing each component (time and nerve consuming). I have in mind
stencil + iron, but I have problem with bottom side component
soldering.
What we did is putting the soldering paste by hand on top side, then
heating board in the oven. But we still suffered with the processor
soldering by hand and we destroyed one LPC1769 during the process, not
to mention all the trouble with soldering all the small components
around the processor by hand. This is ugly and not so good. There must
be a better way to do this.
Can you please also point me to the low-cost tools, like hot-guns,
ovens, etc... that can be used for this process, as well as some
examples/tutorials on how to find the best, least painful and still
affordable method for making these wonderful Opendous boards !
Thank you and best regards,
Drasko
Opendous Inc.
consuming.
http://en.wikipedia.org/wiki/Reflow_soldering
I use a cheap Convection Toaster oven for reflow.
http://www.walmart.com/ip/Black-Decker-Convection-Toaster-Oven-Black/15806193
I use ChipQuik solder paste.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=SMD291SNL-ND
http://store.curiousinventor.com/guides/Surface_Mount_Soldering/Solder_Paste_and_Toaster_Oven
http://www.sparkfun.com/tutorials/58
I buy my stencils from OHARARP.com. Six separate designs for $25
plus shipping assuming they fit less than one 8.5"x11" page including
additional 0.5" borders.
http://www.ohararp.com/Stencils.html
The process is as follows:
1) Print out the Assembly Instructions and prepare all your components
by placing them on this piece of paper. Secure the paper with tape so
you don't mess up hours of work with a twitch.
2) Tape the underside of your PCB onto a stiff board, position your
stencil and hold it down firmly with your fingers, then use a medium-
size metal putty knife to scrape solder paste onto the PCB.
http://www.homedepot.com/h_d1/N-5yc1v/R-202038679/h_d2/ProductDisplay?langId=-1&storeId=10051&catalogId=10053
3) Move all the components from the sheet of paper onto the solder
pasted PCB. By having all the components prepared you spend as little
time as possible inhaling solder paste fumes.
4) OUTDOORS OR IN A GARAGE WITH _EXCELLENT_ VENTILATION place your
board with components mounted into your toaster oven. Now it is just
a matter of following a reflow profile. You can find these in
datasheets for most components or your solder paste. It takes some
practice to get it right so start with simple boards and the
expectation that you will destroy them. Toaster ovens don't have
great heat distribution so reflow just one board at a time.
I strongly suggest you also consider buying a respirator to wear
while reflowing. Use the best cartridges. Your lungs are worth a lot
more than some cheap board which may not work.
http://www.uline.com/BL_992/Reusable-Respirators-Supplies
http://www.uline.com/Product/Detail/S-12527/Dust-Masks-Respirators/3M-6006-Multiple-Toxic-Gasses-Cartridge
Also buy a cheap shop vacuum and extension hose to remove fumes
while soldering. If you get one with its inlet at the right height
you can place it on your work table in front of you. Use ear
protection when it is on and you are nearby.
http://www.homedepot.com/Tools-Hardware-Wet-Dry-Vacuums/h_d1/N-5yc1vZb8pi/R-202198493/h_d2/ProductDisplay?langId=-1&storeId=10051&catalogId=10053
http://www.uline.com/Product/Detail/S-7902/Hearing-Protection/Max-Ear-Plugs-Uncorded
Through trial and error I have come up with the following reflow
profile. I place the boards into a cold toaster oven and set it to
150C. I wait until the oven determines 150C has been reached and
stops heating (heat rods turn off). I then set it to 180C and wait
120 seconds. This is the high end of the "soak" zone but guarantees
everything is ready for reflow. I then turn the oven up to full blast
and stare at the boards for reflow to occur. It happens suddenly and
all the pale solder sparkles. Once components on each side of the
board "sparkle" I immediately turn the oven off. I wait five seconds
and open the toaster oven door. I then wait another five seconds and
start carefully and slowly removing the tray. I then let the tray
cool down. At this point it is on to testing or the next set of
boards. Remember that you can always finish soldering with hand tools
but you cannot un-fry overheated components.
Reflowing the second side of a PCB is just a matter of repeating the
process for the second side. The only caveat is that the most heat
sensitive components should only be reflowed once. This takes some
planning. For the few MicropendousX boards I made I reflowed the
bottom first then the top side as it contains sensitive crystals. I
also tested one board where I reflowed the top only and DIY'ed the
bottom.
>affordable method for making these wonderful Opendous boards
another revision or two to be "wonderful". This design is the fourth
PCB revision and the main reason I put the project on hold is that the
costs and time just got away from me. If you can get some use out of
it that would be great. Feel free to rebrand it as your own and sell
it but please leave a small "Original Design by Opendous Inc." note on
the silkscreen.
Drasko DRASKOVIC
> I reflow solder the boards. 0603 is possible to DIY but very time-
> consuming.
>
> http://en.wikipedia.org/wiki/Reflow_soldering
Hi Matt (hope I have not mistaken your name, I saw it around on the lists),
firs of all thank you very much on these great explanations. It is
very kind of you to devote the time to answer and give a support and I
highly appreciate it.
As I come from the world of embedded software engineering, I am facing
for the first time board making from the scratch and trying to learn
quickly (and as much as budget allows me) the tricks of making
high-quality boards. Your answers really help a lot.
I also hope that these clarifications will be useful for all the
people who stumble upon these discussion searching for the similar
things (and I am almost sure there will be some).
I would like just to address you with one additional question, which
bothered me the most and I think it needs some more clarification. It
is reflowing of the second side of the board, once components have
been reflowed on one side, and this other side prepared with paste and
components positioned on it (but not yet soldered). My concern is that
when we turn this other (soldered) side down to heat unsoldered side,
that previously soldered components will desolder, move from the pads
or fall off. I have saw on the net (here for example :
http://colonyscout.tumblr.com/post/379347565/double-sided-pcb-reflow-soldering)
that some people actually glue the first side components with some
thermically resistent epoxy glue, so that they do not fall off when
other side is heated. Is this needed, or will previously soldered
components however be adhesive enough, so that the glue is not needed
at all ?
So far, we have done similar thing like you suggested learning on our
own, but you pointed out many important details that we were not aware
of. We will apply this on our next assemblies of Opendous boards (ant
there will be quite a few in a near future, I hope).
We first prepared and "coocked" top-side, not because of temperature
sensitivity (we did not even thought about it), but because, well, we
started from the top. And it has more components, so we tried to use
the oven for harder part of work.
But, once we made top side, we were very afraid to repeat the same
process, in the fear that this time already soldered components on the
top side (which will now be turned upside-down) will fall of the board
or move from the pads in the heating process.
As I mentioned it, I do not know if this fear is justified, or the
components will hold on to the board, once they been soldered before.
Because of this we tried to do the bottom side with soldering iron,
but since we are not experts in soldering, we destroyed one LPC1769.
Plus, we could not at all put SD card connector, as a hole on it
towards the pads is too small even for soldering iron top to pass
through it. I think that this component can never be soldered with
soldering iron, and that reflow technique must be applied here (which
implies that basically bottom side of this project must be soldered in
the oven).
This is our experience, and it is very interesting one because we
learned a lot. Plus, it was a lot of fun !
After soldering JTAG wires (I have only Amontec JTAGKey Tiny with 20
pin connector at my disposal) to cortex header on the board I was able
to attach to the processor with OpenOCD and observe regs and memory.
Seems OK, and this is where we are now, with real debugging of both HW
and SW to follow. We will keep you and the list informed of our
experiences, as it may help others and help improve the design (if we
run into some bugs).
> Reflowing the second side of a PCB is just a matter of repeating the
> process for the second side.
Here you did not mention anything about gluing the components from the
previously soldered side... Is this needed ? How do you make them not
to move, as they are on the bottom side now, and whole board is laying
on them, pressing them towards the plate (so that they might maybe
move a little from the pads as the solder starts melting) ?
> The only caveat is that the most heat
> sensitive components should only be reflowed once. This takes some
> planning. For the few MicropendousX boards I made I reflowed the
> bottom first then the top side as it contains sensitive crystals. I
> also tested one board where I reflowed the top only and DIY'ed the
> bottom.
This is something I was not thinking about, but is a great tip. I will
have this in mind for the next assemblies.
>
>>affordable method for making these wonderful Opendous boards
>
> Thanks for the vote of confidence but these boards probably need
> another revision or two to be "wonderful". This design is the fourth
> PCB revision and the main reason I put the project on hold is that the
> costs and time just got away from me.
Well, I have seen many boards during my career as a embedded SW
engineer, and I am just delighted by the design of your boards. I am
honestly impressed with your detailed approach and neat and compact
design. Also, you provided everything - schematics, PCB and gerber
files, JTAG, whole platform up to the operating system ! Really, as we
say in France, "chapeau bas !"
> If you can get some use out of
> it that would be great. Feel free to rebrand it as your own and sell
> it but please leave a small "Original Design by Opendous Inc." note on
> the silkscreen.
Thank you. Of course, if we ever succeed to have any public
appearances with our project, you will be highly credited for your
outstanding effort and help you provided us. We are not planning to
commercialize it, but if we succeed in making something really useful
we will consider this option also - we will by all means keep you
informed on all this.
Thanks again for your great help and support, and above all thanks for
your work - as I said I am truly impressed and I became a fan of
Opendous products !
Best regards,
Drasko
Opendous Support
have been planning to create a nice tutorial but have not been able to
find the time.
The best piece of advice I can give you aside from all the safety
precautions is to budget at least 10 reflow fails. You can do this
with top of the line ICs but it would be cheaper to squander a cheap
and simple design. For example, my Micropendous-32U2 PCB can be
prototyped with the AT90USB162 which costs as little as $2.65
(ArrowNAC.com or Avnet) and is basically just the IC and the USB
connector. Each fail can set you back less than $5 plus the PCB cost.
http://batchpcb.com/index.php/Products/55435
It helps to practise reflow on something complex enough to fail
fully even with a slight reflow error versus testing a regulator or
similar that can withstand significant overheating. Something with a
USB PHY is a good bet.
>reflowing of the second side of the board ...
> ... previously soldered components will desolder
Should have mentioned to use a tray. The one that comes with your
oven. It allows the heat to focus on the top side. Also, it allows
you to easily carry the PCBs from workstation to oven. Parts will not
separate unless you over-reflow, which would destroy the other parts
anyway. Reflow is a very interesting process. Once you learn it you
will be amazed what can be done. It may be surprising to learn that
BGA technology is actually the easiest to reflow as it allows for the
most error.
>we could not at all put SD card connector
well. I have moved to other microSD connectors mostly due to
availability concerns but a simple hack is to cut the occluding metal
with pliers. Right above the signal pins is a "vent" and before that
vent is a strip of metal. Just cut it off with pliers or a small
rotary tool with a cutting wheel. It helps to have a soldering iron
tip that is long and thin and thin gauge solder.
>I think that this component (LPC1769) can
http://store.curiousinventor.com/guides/Surface_Mount_Soldering/QFP
Flux up all the pins. Solder everything over without regard for
shorts and blobs. Then flux it up again and remove excess solder with
the solder wick. Just place the solder wick onto the fluxed-up pins
and press the solder wick down using your soldering iron tip. Excess
solder will flow into the wick but solder between the pins and PCB
pads is much harder to remove and will therefore stay permanently.
Get an assortment of wick sizes. "Activated" wick contains flux and
is easier to use but then you have to clean it afterwards.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=473-1062-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=473-1063-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=473-1061-ND
Cleaning consists of using an ultrasonic cleaner (cheap ~$30 denture
cleaner will suffice) and flux remover and/or IPA (99% Isopropanol/
Isopropyl Alcohol). You can buy 99% Isopropyl Alcohol at some
pharmacies. It is not as pure as that meant for electronics but MCUs
are not that sensitive. Don't use 70% and remember to let your boards
dry thoroughly before use.
http://www.mgchemicals.com/products/4140.html
http://www.mgchemicals.com/products/824.html
>I am honestly impressed with your detailed approach and neat and
>compact design. Also, you provided everything - schematics, PCB
>and gerber files, JTAG, whole platform up to the operating system!
>Really, as we say in France, "chapeau bas !"
trust company development boards because they tend to be designed to
ludicrous specifications that make no commercial sense.
Microcontrollers should work on 2-layer PCBs designed to industry
standard 8mil-8mil-20mil trace-clearance-minimum_drill. As for the
code, there is still a lot of work to be done, especially with
debugging the TCP/IP stack.
Good luck with your project and I just want to remind you again to
take your safety very seriously. Create a fume extractor (shop vacuum
and wear ear plugs) and buy a half-face respirator with good
cartridges. Swim goggles or wrap-around safety glasses are important
for eye protection. Visit your local hardware store and look into
respirators for mold, lead, and spray sealants. This covers all the
types of electronics manufacturing fumes; rosin flux fumes, metal
(tin) fumes, and chemical fumes such as solder paste off-gassing and
reflow fumes. If you have asthma you can become sensitized to flux
fumes very quickly.
http://www.homedepot.com/h_d1/N-5yc1v/R-202080144/h_d2/ProductDisplay?langId=-1&storeId=10051&catalogId=10053
On Jun 13, 8:19 pm, Drasko DRASKOVIC <drasko.drasko...@gmail.com>
wrote:
>
> > I reflow solder the boards. 0603 is possible to DIY but very time-
> > consuming.
>
> >http://en.wikipedia.org/wiki/Reflow_soldering
>
> Hi Matt (hope I have not mistaken your name, I saw it around on the lists),
> firs of all thank you very much on these great explanations. It is
> very kind of you to devote the time to answer and give a support and I
> highly appreciate it.
>
> As I come from the world of embedded software engineering, I am facing
> for the first time board making from the scratch and trying to learn
> quickly (and as much as budget allows me) the tricks of making
> high-quality boards. Your answers really help a lot.
>
> I also hope that these clarifications will be useful for all the
> people who stumble upon these discussion searching for the similar
> things (and I am almost sure there will be some).
>
> I would like just to address you with one additional question, which
> bothered me the most and I think it needs some more clarification. It
> is reflowing of the second side of the board, once components have
> been reflowed on one side, and this other side prepared with paste and
> components positioned on it (but not yet soldered). My concern is that
> when we turn this other (soldered) side down to heat unsoldered side,
> that previously soldered components will desolder, move from the pads
Drasko DRASKOVIC
<opendous...@gmail.com> wrote:
>>thank you very much on these great explanations
>
> I was in the same boat and know how steep the learning curve is. I
> have been planning to create a nice tutorial but have not been able to
> find the time.
Hi Matt,
great explanations, as always. Thank you for your time.
I will try to put the board on a tray and do reflowing 2 times for the
nexat boards, like this :
1) reflow bottom side - the one with the CPU and SD card connector, as
it contains less sensitive components
2) turn board, populate with components, put on a tray and reflow (I
hope that CPU and SD connector will not move from their places, as
they will be laying on a tray)
So far the board that we made seems to be showing signs of life - we
can connect with OpenOCD and Blinky app is working fine ;). We will
continue hacking, I am currently trying to make Makefile more proper
and cleaner, as it has some little bugs...
BR,
Drasko
Opendous Inc.
would be that you overheat the board. Watch the process very
carefully and the moment you see solder paste has reflowed throughout
most of the board, turn off the heat and seconds later open the oven
door to start cooling.
You have to be careful not to thermally shock the components but
even with the door open the cooling rate will be close to the target
2C/s.
Nice to see someone is getting some use out of this project.
On Jun 20, 6:20 pm, Drasko DRASKOVIC <drasko.drasko...@gmail.com>
wrote:
>