Balloon Project PCBs.

[Brad Luyster]

Guys,

Thank you so much for the overwhelming response to the modular balloon
CPU Idea I pitched tonight. I received a total of $95 in
contributions to the project with just the idea, including $70 Cash
and $25 paypal. This will easily pay for the first run of boards.

I'm going to order the first run from Gold Phoenix
(goldphoenixpcb.biz) Tomorrow at 8PM unless someone comes forward with
a glaring PCB error. The PCB can currently be found at:
http://github.com/Zuph/LVL1-Transatlantic-Balloon

The only untested footprint on the board is for the JST connector. I
will verify this connector again tomorrow, to make sure the footprint
is correct.

Along with a PDF and PPTX of the presentation I gave tonight, along
with Schematic files. The file you're interested in is suffixed with
a "-3".

I will be purchasing the $100 package from Gold Phoenix, which will
give us 25 PCBs, Shipped 3 day from China. Turnaround should be 7-10
days, at worst.

I intend to get a Dokuwiki setup for the Balloon Project at some point
tomorrow, and Next week, during the Ballon Project Meeting time, I'll
be teaching about documentation and source code management, using
Dokuwiki, Git, and GitHub. Although I'll be focusing on the parts of
these which are most applicable to the balloon project, all are
welcome.

The Bill of Materials is as follows. Although the sensors are fairly
expensive from digikey, they can be found cheaper elsewhere. Also,
since all the sensors are on the I2C bus, they needn't all be
installed for proper board functionality.

Bill of Materials:
1x Atmel AtMega324P, TQFP Package (Digikey: ATMEGA324P-20AU-ND) $7.05
Ea. ($7.05)
1x BMP085 Pressure Sensor (Digikey: 828-1005-1-ND) $4.38 Ea. ($4.38)
8x JST XA 6 Pin Header, B06B-XASK-1 (Digikey: 455-1996-ND) $0.34 Ea.
($2.72)
1x TMP100 Temperature Sensor (Digikey: 296-19524-1-ND) $6.15 Ea.
($6.15)
1x DS3231 TXCO Real Time Clock (Digikey: DS3231S#-ND) $8.40 Ea.
($8.40)
1x Breakaway Headers (Digikey: WM4499-ND) $1.06 Ea. ($1.06)
1x Tactile Switch (Digikey: CKN9092-ND) $0.11 Ea. ($0.11)
1x 12MHz Resonator (Digikey: 478-2112-1-ND) $0.86 Ea. ($0.86)
1x 1206 Green LED (Digikey: 160-1188-1-ND) $0.11 Ea. ($0.11)
1x 1206 Yellow LED (Digikey: 160-1170-1-ND) $0.11 Ea. ($0.11)
1x 1206 Red LED (Digikey: 475-1416-1-ND) $0.11 Ea. ($0.11)
3x 0805 220 Ohm Resistors (Digikey: RMCF0805JT220RCT-ND) $0.03 Ea.
($0.09)
5x 0805 4.7k Ohm Resisotrs (Digikey: RMCF0805JT4K70CT-ND) $0.03 Ea.
($0.15)
6x 0805 .1uF Capacitor (Digikey: 478-3351-1-ND) $0.04 Ea. ($0.24)

Total: $31.54

Once again, many thanks. I hope this ends up being the correct way
forward. Please let me know before 8PM on 9/3/2010 if there are any
issues with the PCB.

-Brad

[Cindy Harnett]

Brad, is the final board lvl1_ballooncpu-3.brd? Would you like a paste
stencil?

[Brad Luyster]

Cindy,

lvl1_ballooncpu-3.brd is the final board.  The CAM files I intend to ship (unless errors are found) are located in the zip file.  If you have the opportunity, I would really appreciate any feedback you could give.  I intend to go over the 1:1 printout today with some same-footprint parts, as well as verify the JST connector footprint with a ruler.  The PDF datasheet for the connection is found here: http://www.jst-mfg.com/product/pdf/eng/eXA1.pdf  The exact part number is B06B-XASK-1.

I haven't been able to find much solid data on how to size traces for high-currents, but as best I can tell, the thick traces on the JST connectors' 5v and 3.3v lines should be able to carry 3A without a problem (but not more).  Sparkfun has verified the landing pattern for the DS3231, BMP085 and a package-identical Temperature Sensor in the same line as the TMP100, as well as the AtMega324.  The resistors are all standard sizes, and if the 1206 LEDs I have won't fit that closely together, they can be replaced with a slightly smaller package and still fit on the pads.

The DS3231, it's passives, and the Filter caps for the AtMega324 are located on the reverse side of the board (something I've never tried before), and called out by Solder Mask (since 2 layer silkscreen is expensive).

I won't pretend to not be nervous:  This is the first time I've ordered a bulk number of PCBs without first doing a design check on a one-off.

To anyone:  If GitHub gives you trouble saving the .brd file, simply navigate here: http://github.com/Zuph/LVL1-Transatlantic-Balloon/blob/master/Hardware/lvl1_ballonCPU/lvl1_ballooncpu-3.brd, and right-click, save-as on the "Raw" link just above the file.

Thanks so much, Cindy, and Everyone Else.

--
You received this message because you are subscribed to the Google
Groups "LVL1 - Louisville's MakerSpace" group.
To post to this group, send email to lv...@googlegroups.com
To unsubscribe from this group, send email to
lvl1+uns...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/lvl1?hl=en
For more info about LVL1 visit www.lvl1.org

[Daniel Bowen]

Cindy, a paste stencil would be really amazing to have, and we plan to teach while soldering these so that would be a great opportunity to show how to do stencil solder paste application work.  

Thanks,

Dan

On Fri, Sep 3, 2010 at 8:22 AM, Cindy Harnett <cindy....@louisville.edu> wrote:

[Brad Luyster]

Oops, forgot to answer 50% of Cindy's quesions :-)

[Cindy Harnett]

As long as the smt pads don't move, the same stencil will work if the
traces are widened or rerouted. The JST looks through hole and this
also won't affect a stencil. Let me know when you are ready to go and
I will try exporting these--C

On Sep 3, 9:10 am, Brad Luyster <bradluys...@gmail.com> wrote:
> Oops, forgot to answer 50% of Cindy's quesions :-)
>

> >> lvl1+uns...@googlegroups.com <lvl1%2Bunsu...@googlegroups.com>

> >> For more options, visit this group at
> >>http://groups.google.com/group/lvl1?hl=en
> >> For more info about LVL1 visitwww.lvl1.org
>
> >  --
> > You received this message because you are subscribed to the Google
> > Groups "LVL1 - Louisville's MakerSpace" group.
> > To post to this group, send email to lv...@googlegroups.com
> > To unsubscribe from this group, send email to

> > lvl1+uns...@googlegroups.com <lvl1%2Bunsu...@googlegroups.com>

[DE 'Tad' Heckaman III]

Not having opened up any schematics or PCB layout files before, what is the tool of choice to view these?

I'm already liking Git... I've already cloned the entire repository down to my PC. From the discussion yesterday, I assume there will be a separate LVL1 Balloon GitHub repository setup at some point, for all project related files?

Tad

--
You received this message because you are subscribed to the Google
Groups "LVL1 - Louisville's MakerSpace" group.
To post to this group, send email to lv...@googlegroups.com
To unsubscribe from this group, send email to
lvl1+uns...@googlegroups.com

For more options, visit this group at
http://groups.google.com/group/lvl1?hl=en
For more info about LVL1 visit www.lvl1.org

--
Tad Heckaman

[Brian Wagner]

How do you make your stencils?

[Cindy Harnett]

it will be laser-cut 3 mil mylar

On Sep 3, 10:42 am, Brian Wagner <br...@tegrasys.com> wrote:
> How do you make your stencils?
>
> On Fri, Sep 3, 2010 at 9:44 AM, Cindy Harnett

> <cindy.harn...@louisville.edu>wrote:

> > lvl1%2Bunsu...@googlegroups.com<lvl1%252Buns...@googlegroups.com>

>
> > > >> For more options, visit this group at
> > > >>http://groups.google.com/group/lvl1?hl=en
> > > >> For more info about LVL1 visitwww.lvl1.org
>
> > > >  --
> > > > You received this message because you are subscribed to the Google
> > > > Groups "LVL1 - Louisville's MakerSpace" group.
> > > > To post to this group, send email to lv...@googlegroups.com
> > > > To unsubscribe from this group, send email to
> > > > lvl1+uns...@googlegroups.com<lvl1%2Bunsu...@googlegroups.com><

> > lvl1%2Bunsu...@googlegroups.com<lvl1%252Buns...@googlegroups.com>

[DE 'Tad' Heckaman III]

which brings up a good question... Brian, can the CNC machine be outfitted with a laser? :) I don't see why it couldn't... we just need a powerful laser right?

For more info about LVL1 visit www.lvl1.org

--
Tad Heckaman

[Brad Luyster]

These Schematics and Board files were all written in EagleCAD.

We need to discuss the project workflow at the next balloon meeting, but I'm imaging that everyone will create their own account, and I'll simply add them as collaborators on the existing repository.  As a result, other balloon project members can push to the remote as much as they like.  If the project grows large enough, we may use a manager/peon workflow, where everyone has their own fork of the project, and submits pull requests to a maintainer or team of maintainers.  In this case, there would be a separate LVL1 Balloon GitHub account.

-Brad

[Brian Wagner]

I guess so, probably need to have something in place so it does not burn the table and catch it on fire.  I am game.  I think this would be a fun thing to do.

b

[DE 'Tad' Heckaman III]

especially with an A axis... I can picture the laser spinning around and putting a nice burn mark from floor to ceiling :)

[Brad Luyster]

Oh, I just noticed something:

Currently, the GPS connector is simply on .1" headers (the same kind used for everything).  My initial logic for this was that any flight-ready computer configuration would have the headers for the GPS unit soldered into the board.  If we don't think this will be the case, I can add another JST connector and alter the board layout accordingly.  This won't have any significant impact on ordering time.  The only issue is that during prototyping, we can't use .1" headers to develop a GPS system:  The JST connectors are 2.5mm, while .1" = 2.54mm (with fewer than 6 pins, we could make it work.  Pins >=6, though, you can't force it without breaking something.

Let me know your thoughts.

-Brad

[Brad Luyster]

Also just noticed this:   For added flexibility, I'll put some GPIO on the currently unused ADC pins, and bring the JTAG port out to pin-headers.  We may or may not use the more advanced debuggers, but the only overhead is a few more holes in the board.

[DE 'Tad' Heckaman III]

The board on the last page of the PDF has lots of wording... Will the wording be on the production boards? In the PDF, Ballon shows up twice instead of the word Balloon. Didn't know if you had fixed that or not. I don't see the wording in Eagle... but I'm probably doing something wrong.

[Brad Luyster]

Good catch.   The spelling is still messed up on the PCB, but I'll fix it before sending it off.  The PCB will have all that wording included.  To see it in eagle, make sure you have the tValues and tNames layers turned on.

[DE 'Tad' Heckaman III]

Ah... I see the wording now, fixed it, saved it as a new file, and did a git-commit -> master. Can you tell me if that worked properly? I don't see it on the web so I'm not sure it worked... :) I'd thought I would try out this fancy GIT thing.

[Brad Luyster]

Git Commits just commit to your local repository.  You don't have write access to my remote repository (Yet!).  Start reading through progit.org for more information about  where stuff is held, and how.

[DE 'Tad' Heckaman III]

bah, read documentation!?! Heresy! :)

[Brad Luyster]

Alright, I have updated the PCB.  Adding the JTAG connector required a complete re-route, so please double check me for error again.  I'm going to be at LVL1 around 6pm to finish up work on the PCB, and get it sent to Gold Phoenix.  The current version on GitHub is accurate.

The reworked PCB is a little messier, but the power distribution traces are a little thicker, and more smartly run.

-Brad

[DE 'Tad' Heckaman III]

One thing, and I'm probably missing it... I see that the switch, when pressed, grounds the reset pin on the cpu. I don't see where the switch gets ground from... is the entire board 'ground'?

[Brad Luyster]

Yep, the entire board, both sides, is a ground plane.  The ground reference will be provided by a the Ground Support Package, or Battery Management System.

[stea...@gmail.com]

Brad,

    I'm on the road now so I may not be able to see the latest revision - darn eagle for not having an iPhone version :-P.  

One thing I would do, based on this morning's rev, is label pin 1 of the module connectors, GPS, FTDI, JTAG.  There isn't much room, but it would be nice to have the pins of each connector labelled, as SparkFun does.  Also you may want to run it through freedfm.com for a final check.

Thanks,

Dan

Sent from my iPhone 4

[Brad Luyster]

Dan,

Already done, for the labeling of each pin.  Unfortunately, there definitely isn't room enough to legibly label each pin.

[Brad Luyster]

Alright, I fixed some final problems with the silkscreen (Silkscreen overlapping Vias, Overlapping Pads, etc), and cleaned up the solder mask on the bottom of the board.  The version located on GitHub as of right now is the version I will submit to Gold Phoenix for production in 50 minutes.  If there's something press-stopping, let me know in the next 50 minutes, by email or phone (5025339013).

I tried to put labels on the pins of most of the connectors.  I don't know if they're large enough to be legible, though.

-Brad

[stea...@gmail.com]

Thanks for the hard work Brad!

Thanks,

Dan

Sent from my iPhone 4

[Brad Luyster]

After putting in the order, I still haven't heard back from the board house on the price or a paypal link.  This is probably understandable, due to the weekend.  I found another fab house in China which seems to be slightly better overall.  http://www.pcbwing.com/

They only have 2 not-bogus reviews in English (that I can find), but both are positive.  The website is friendlier to navigate, and their prices are in line with Gold Phoenix.  Furthermore, they give you more options to customize the order.  Furthermore, they charge tooling and board costs separately, so if you need to reorder, tooling costs are greatly reduced.

I've been working on and off on the associated ground support board, with 0-3A adjustable loads on 5V and 3.3V rails, along with 0-3A switch-mode power supplies.  I've verified the design models for the adjustable load using PWM in SPICE, as well as the power handover circuitry.  For the Switch-mode power supply, the NCP3163 is the simplest model I could find.

I'm honestly worried that I won't be able to fit all this stuff inside the design limits of the free version of Eagle, since this board will have thick traces, and big heatsinks.

At this rate, I'm never going to finish grad school!

[Brad Luyster]

I got an email back this morning.  The order has been paid for, and everything is now in Gold Phoenix's hands.  Hopefully, within 8 business days, we'll have 25 balloon computers in our hands.

[Brad Luyster]

Parts are on the way! PCBs should be here by Wednesday. Join me in
anticipating their arrival: Fedex Tracking: 464294078925E

http://www.fedex.com/Tracking?action=track&tracknumbers=464294078925

[Brad Luyster]

The PCBs are HERE! We ordered 25, and received 27. After inspecting
them all, I found 2 goofs that are my fault, and 2 goofs that are Gold
Phoenix's fault:
- I connected the FTDI cable's 5v line to the 5v bus. We'll have to
cut this trace.
- I didn't mark pin one on the RTC.

- A few of the boards have kind of crappy silkscreen registration.
This is only a big issue on one of the boards, which might be unusable
as a result.
- On two of the boards, the solder mask appears to be delaminating.
This might just be a visual defect, though.

I will be at the space around 4:30 or 5:00 to try and solder one of
these things together.

-Brad

[Grace Simrall]

AWESOME!

[Brad Luyster]

Looks like I screwed up pretty bad. I owe you all an apology. The
footprint for the Temperature Sensor and the RTC got pretty badly
mangled. Go figure, these were 2 of the 3 parts for which I
configured the pinout (rather than Sparkfun or someone else).
Unfortunately, these parts won't be able to be reliably soldered to
the flight computer board.

The good news is that we have a platform to develop for and use on the
test flight.

As a result of this, I'm going to fix the errors, and get another run
of boards made. I'll fund this run out of my own pocket.
Fortunately, the rest of the board checks out, so these errors should
be the only problems which need fixing.

Again, I'm really sorry for these problems. Hopefully, we can get
another set by next week, and not have to worry about flight computers
for the future.

-Brad

[Jeff Johnson]

Brad, I'm really not part of the balloon group (other than an interested observer), but I don't think anyone would expect you to front the costs for another run. Seriously, its a group effort and you shouldn't have to pay for that.

I'll throw in my $10 at the next meeting. Anyone else want to pony up?

[Bill Piepmeyer]

I agree with Jeff. This project is entirely risk based, with a large
dose of trial and error. The boards are still usable even without those,
just not for the full purpose that we want them but for a subset.

I'll throw in my $10 at the next meeting.

(BTW is it to late to run the analog and extra pins out to at least a
pad or plated through pin hole, just in case we want to make use of them.)

--Bill

[Brad Luyster]

I'm sitting down right now to fix the PCBs. I'll bring a corrected
design to the balloon meeting tomorrow for approval, and get the
analog pins brought out to some pins.

[Tim Miller]

These would be good boards to test if they can handle freezing and power spikes

On Sep 15, 2010 10:47 PM, "Bill Piepmeyer" <bi...@solsticeinc.net> wrote:

 I agree with Jeff. This project is entirely risk based, with a large dose of trial and error. The boards are still usable even without those, just not for the full purpose that we want them but for a subset.

I'll throw in my $10 at the next meeting.

(BTW is it to late to run the analog and extra pins out to at least a pad or plated through pin hole, just in case we want to make use of them.)

--Bill

On 9/15/2010 10:34 PM, Jeff Johnson wrote:
>

> Brad, I'm really not part of the balloon group (oth...

[Grace Simrall]

I want to echo my sentiments as well.  The group knew the risks and willingly took it on when it funded the first order.  I was prepared for this outcome, and I'll happily contribute my $10 tomorrow.

[Brad Luyster]

I have uploaded a fixed version to Github. All of the previous errors
were fixed, including the 5V line to the FTDI cable, Pin 1 indicator
for the DS3231, and the layout for the TMP100 and DS3231.
Additionally, all previously unused pins have been brought out to
headers, giving us 3 Analog-to-Digital lines, and 3 GPIO. As a result
of the incredible legibility of the 32 mil Silkscreen, I've changed
all the passive component values to this smaller size, giving more
room for larger silkscreen.

If there are any other suggestions, I'll happily try to get them
integrated by tomorrow's meeting.

[stea...@gmail.com]

Minor issues- the time can be easily gathered from the GPS, in fact we never put an RTC in the air for all 24 flights of spirit of knoxville. Internal temperature can be gathered as easily from a peripheral I2C temp sensor.

I'd suggest we wait until using this board for a bit of software development and integration before ordering an improved version. After some actual development, and further refining of our flight plans, additional unknown improvements will likely become apparent.

:-)
Dan

[D S]

I would just like to point out how amazing the community is which is developing within this group.  Everyone in LVL1 takes on responsibilities with pride and with a sense of communal cooperation that is the most encouraging display of social cohesion I have ever seen.  Situations like this one, where Brad is offering to fix an honest mistake out of his own pocket, have restored a faith in humanity that was all but lost in me before LVL1.  I am SO indescribably proud to be a part of something like this, and I just couldn't help but elucidate how amazing you all are! 

DS

[Brian Wagner]

I will throw some money into the pot also

b