Consider external antenna connector (U.FL)

[zsup]

Currently the Core only has an on-board antenna, but there may be cases where people need greater performance. Let's consider adding a U.FL connector.

[zsup]

This should be designed in as a stuffing option, using an IPEX 20279-001E (3 pad) footprint

[zsup]

Footprint for design: http://www.digikey.com/product-detail/en/RECE.20279.001E.01/931-1107-6-ND/2332758

[zsup]

Ok after a bit of finagling, I think I figured out a pretty decent layout that would allow for both a uFL connector and an antenna. @andyw-lala, would love your thoughts on this.

Basically, here's the thought:

• We should extend the bottom of the Core by 0.1" to give us a bit more breathing room, but knock off the corners.
• We should replace our current antenna with the ACX AT8010, since the CC3000 is already FCC certified with this combination, and it seems that it can be purchased inexpensively in China but has good US support (through Johanson).
• The antenna should be placed as far south as we can put it to give more clearance from the other components. Currently I've drawn it with a 90 degree angle between the trace and the antenna; @andyw-lala, would it be better to round the corner of the trace and move the antenna a bit more to the left?
• An additional 10pF capacitor should be added that shares a pad with C4. This capacitor will jut off to the left and provide a connection to the uFL connector.
• The uFL connector (20279-001E, as mentioned above) should be positioned as pictured in the image.
• As far as stuffing options go, if we're using the antenna, L1, C4, and the antenna should be stuffed; if we're using the uFL connector, the new capacitor and the uFL connector should be stuffed (and L1 and C4 would not be stuffed).

Does all this make sense?

[andyw-lala]

That makes sense. and is pretty vanilla.

As I said the other night, we might be able to get cute and avoid the extra 10pF cap with cute layout tricks.

Check with your board house that producing a non-rectangular PCB doesn't push up the cost, it may cause them to move from abrasive disk singulation to router, that might (or might not) increase the price.

On Thu, Jun 27, 2013 at 2:49 PM, cazzo notifications@github.com wrote:

[image: 2013-06-27 13 57 12]https://f.cloud.github.com/assets/929333/718269/766f79b2-df61-11e2-96d2-c9aed42b5435.jpg

Ok after a bit of finagling, I think I figured out a pretty decent layout that would allow for both a uFL connector and an antenna. @andyw-lalahttps://github.com/andyw-lala, would love your thoughts on this.

Basically, here's the thought:

• We should extend the bottom of the Core by 0.1" to give us a bit more breathing room, but knock off the corners.
• We should replace our current antenna with the ACX AT8010, since the CC3000 is already FCC certified with this combination, and it seems that it can be purchased inexpensively in China but has good US support (through Johanson).
• The antenna should be placed as far south as we can put it to give more clearance from the other components. Currently I've drawn it with a 90 degree angle between the trace and the antenna; @andyw-lalahttps://github.com/andyw-lala, would it be better to round the corner of the trace and move the antenna a bit more to the left?
• An additional 10pF capacitor should be added that shares a pad with C4. This capacitor will jut off to the left and provide a connection to the uFL connector.
• The uFL connector (20279-001E, as mentioned above) should be positioned as pictured in the image.
• As far as stuffing options go, if we're using the antenna, L1, C4, and the antenna should be stuffed; if we're using the uFL connector, the new capacitor and the uFL connector should be stuffed (and L1 and C4 would not be stuffed).

Does all this make sense?

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20150290 .

Andy

[zsup]

Thanks @andyw-lala. Thoughts on rounding the corner of the antenna trace?

Zach Supalla 312.953.3413

On Thursday, June 27, 2013 at 3:01 PM, andyw-lala wrote:

That makes sense. and is pretty vanilla.

As I said the other night, we might be able to get cute and avoid the extra
10pF cap with cute layout tricks.

Check with your board house that producing a non-rectangular PCB doesn't
push up the cost, it may cause them to move from abrasive disk singulation
to router, that might (or might not) increase the price.

On Thu, Jun 27, 2013 at 2:49 PM, cazzo <notifications@github.com (mailto:notifications@github.com)> wrote:

[image: 2013-06-27 13 57 12]https://f.cloud.github.com/assets/929333/718269/766f79b2-df61-11e2-96d2-c9aed42b5435.jpg

Ok after a bit of finagling, I think I figured out a pretty decent layout
that would allow for both a uFL connector and an antenna. @andyw-lalahttps://github.com/andyw-lala,
would love your thoughts on this.

Basically, here's the thought:

• We should extend the bottom of the Core by 0.1" to give us a bit
  more breathing room, but knock off the corners.
• We should replace our current antenna with the ACX AT8010, since the
  CC3000 is already FCC certified with this combination, and it seems that it
  can be purchased inexpensively in China but has good US support (through
  Johanson).
• The antenna should be placed as far south as we can put it to give
  more clearance from the other components. Currently I've drawn it with a 90
  degree angle between the trace and the antenna; @andyw-lalahttps://github.com/andyw-lala,
  would it be better to round the corner of the trace and move the antenna a
  bit more to the left?
• An additional 10pF capacitor should be added that shares a pad with
  C4. This capacitor will jut off to the left and provide a connection to the
  uFL connector.
• The uFL connector (20279-001E, as mentioned above) should be
  positioned as pictured in the image.
• As far as stuffing options go, if we're using the antenna, L1, C4,
  and the antenna should be stuffed; if we're using the uFL connector, the
  new capacitor and the uFL connector should be stuffed (and L1 and C4 would
  not be stuffed).

Does all this make sense?

—
Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20150290
.

Andy

— Reply to this email directly or view it on GitHub (https://github.com/sprk/core/issues/1#issuecomment-20151054).

[andyw-lala]

Sorry - yes always round or at the very least mitre RF traces.

I'll give you details for the 50Ohm trace once I track down the exact parameters of the PrePrep your board house uses. The exact values vary from vendor to vendor, but I'll run the numbers to see how much difference it really makes.

On Thu, Jun 27, 2013 at 3:04 PM, cazzo notifications@github.com wrote:

Thanks @andyw-lala. Thoughts on rounding the corner of the antenna trace?

Zach Supalla 312.953.3413

On Thursday, June 27, 2013 at 3:01 PM, andyw-lala wrote:

That makes sense. and is pretty vanilla.

As I said the other night, we might be able to get cute and avoid the extra 10pF cap with cute layout tricks.

Check with your board house that producing a non-rectangular PCB doesn't push up the cost, it may cause them to move from abrasive disk singulation to router, that might (or might not) increase the price.

On Thu, Jun 27, 2013 at 2:49 PM, cazzo <notifications@github.com(mailto: notifications@github.com)> wrote:

[image: 2013-06-27 13 57 12]< https://f.cloud.github.com/assets/929333/718269/766f79b2-df61-11e2-96d2-c9aed42b5435.jpg>

Ok after a bit of finagling, I think I figured out a pretty decent layout that would allow for both a uFL connector and an antenna. @andyw-lala< https://github.com/andyw-lala>, would love your thoughts on this.

Basically, here's the thought:

• We should extend the bottom of the Core by 0.1" to give us a bit more breathing room, but knock off the corners.
• We should replace our current antenna with the ACX AT8010, since the CC3000 is already FCC certified with this combination, and it seems that it can be purchased inexpensively in China but has good US support (through Johanson).
• The antenna should be placed as far south as we can put it to give more clearance from the other components. Currently I've drawn it with a 90 degree angle between the trace and the antenna; @andyw-lala< https://github.com/andyw-lala>, would it be better to round the corner of the trace and move the antenna a bit more to the left?
• An additional 10pF capacitor should be added that shares a pad with C4. This capacitor will jut off to the left and provide a connection to the uFL connector.
• The uFL connector (20279-001E, as mentioned above) should be positioned as pictured in the image.
• As far as stuffing options go, if we're using the antenna, L1, C4, and the antenna should be stuffed; if we're using the uFL connector, the new capacitor and the uFL connector should be stuffed (and L1 and C4 would not be stuffed).

Does all this make sense?

— Reply to this email directly or view it on GitHub< https://github.com/sprk/core/issues/1#issuecomment-20150290> .

Andy

— Reply to this email directly or view it on GitHub ( https://github.com/sprk/core/issues/1#issuecomment-20151054).

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20151221 .

Andy

[andyw-lala]

Should I expect draft eagle files to be committed soon ?

On Thu, Jun 27, 2013 at 3:10 PM, Andy Warner andyw@pobox.com wrote:

Sorry - yes always round or at the very least mitre RF traces.

I'll give you details for the 50Ohm trace once I track down the exact parameters of the PrePrep your board house uses. The exact values vary from vendor to vendor, but I'll run the numbers to see how much difference it really makes.

On Thu, Jun 27, 2013 at 3:04 PM, cazzo notifications@github.com wrote:

Thanks @andyw-lala. Thoughts on rounding the corner of the antenna trace?

Zach Supalla 312.953.3413

On Thursday, June 27, 2013 at 3:01 PM, andyw-lala wrote:

That makes sense. and is pretty vanilla.

As I said the other night, we might be able to get cute and avoid the extra 10pF cap with cute layout tricks.

Check with your board house that producing a non-rectangular PCB doesn't push up the cost, it may cause them to move from abrasive disk singulation to router, that might (or might not) increase the price.

On Thu, Jun 27, 2013 at 2:49 PM, cazzo <notifications@github.com(mailto: notifications@github.com)> wrote:

[image: 2013-06-27 13 57 12]< https://f.cloud.github.com/assets/929333/718269/766f79b2-df61-11e2-96d2-c9aed42b5435.jpg>

Ok after a bit of finagling, I think I figured out a pretty decent layout that would allow for both a uFL connector and an antenna. @andyw-lala< https://github.com/andyw-lala>, would love your thoughts on this.

Basically, here's the thought:

• We should extend the bottom of the Core by 0.1" to give us a bit more breathing room, but knock off the corners.
• We should replace our current antenna with the ACX AT8010, since the CC3000 is already FCC certified with this combination, and it seems that it can be purchased inexpensively in China but has good US support (through Johanson).
• The antenna should be placed as far south as we can put it to give more clearance from the other components. Currently I've drawn it with a 90 degree angle between the trace and the antenna; @andyw-lala< https://github.com/andyw-lala>, would it be better to round the corner of the trace and move the antenna a bit more to the left?
• An additional 10pF capacitor should be added that shares a pad with C4. This capacitor will jut off to the left and provide a connection to the uFL connector.
• The uFL connector (20279-001E, as mentioned above) should be positioned as pictured in the image.
• As far as stuffing options go, if we're using the antenna, L1, C4, and the antenna should be stuffed; if we're using the uFL connector, the new capacitor and the uFL connector should be stuffed (and L1 and C4 would not be stuffed).

Does all this make sense?

— Reply to this email directly or view it on GitHub< https://github.com/sprk/core/issues/1#issuecomment-20150290> .

Andy

— Reply to this email directly or view it on GitHub ( https://github.com/sprk/core/issues/1#issuecomment-20151054).

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20151221 .

Andy

Andy

[zsup]

@satishgn will put them together tonight

[andyw-lala]

From conversations with the board house Zach intends to use, the key parameters for RF are:

If we specify 7mil (178um) 7628 prepreg, and 1oz (35um) Cu plating, then: we should use 25mil (635um) wide microstrip for 50Ohm RF lines. Spacing to ground on the top layer should be >= 45mil (1.143mm), and the edge of any grounds facing the microstrip shoudl be stitched with vias spaced no more than 4mm apart (lambda/8 ~8mm @ 2.45GHz with this material, 4mm spacing extends the stitching to the 2nd harmonic.)

I suggest we use these as design rules for the RF components, and these are the values I intend to use for the RF test boards.

Questions/corrections -> me.

On Thu, Jun 27, 2013 at 3:14 PM, cazzo notifications@github.com wrote:

@satishgn will put them together tonight

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20151807 .

Andy

[zsup]

Great, thanks @andyw-lala; just made some design tweaks to follow these rules. Would love your feedback when you have a chance to see the board ('satish' branch)

Zach Supalla 312.953.3413

On Friday, June 28, 2013 at 3:04 PM, andyw-lala wrote:

From conversations with the board house Zach intends to use, the key
parameters for RF are:

If we specify 7mil (178um) 7628 prepreg, and 1oz (35um) Cu plating, then:
we should use 25mil (635um) wide microstrip for 50Ohm RF lines.
Spacing to ground on the top layer should be >= 45mil (1.143mm), and the
edge of any grounds facing the microstrip shoudl be stitched with vias
spaced no more than 4mm apart (lambda/8 ~8mm @ 2.45GHz with this material,
4mm spacing extends the stitching to the 2nd harmonic.)

I suggest we use these as design rules for the RF components, and these are
the values I intend to use for the RF test boards.

Questions/corrections -> me.

On Thu, Jun 27, 2013 at 3:14 PM, cazzo <notifications@github.com (mailto:notifications@github.com)> wrote:

@satishgn will put them together tonight

—
Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20151807
.

Andy

— Reply to this email directly or view it on GitHub (https://github.com/sprk/core/issues/1#issuecomment-20211145).

[technobly]

If you intend on using 1oz. copper this might not be much of a concern, but when relying on the copper weight for trace impedance matching... you should specify the tolerance on the copper weight as well when you place your PCB order. This ensures you won't have "surprises" later.

This is more of an issue if you are plating up to a higher weight, but could still bite you if you are not careful.

[andyw-lala]

The top layer will be plated up from 18um to 35um. Not sure Zach gets to talk tolerances, spec limits or Cpks with the board houses he's using :) On 28 Jun 2013 15:45, "DubbyTT" notifications@github.com wrote:

If you intend on using 1oz. copper this might not be much of a concern, but when relying on the copper weight for trace impedance matching... you should specify the tolerance on the copper weight as well when you place your PCB order. This ensures you won't have "surprises" later.

This is more of an issue if you are plating up to a higher weight, but could still bite you if you are not careful.

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20213282 .

[technobly]

Interesting, never heard of plating up one side of the board before. Tape off the bottom? If you can't specify it... you might as well just say, "gimme 1/2 oz or 1oz or 1-1/2oz. ...whichever". My point is if it matters, don't let them give the you wrong thing. If not, then no worries.

;-)

[andyw-lala]

How else do you get plated vias ? On 28 Jun 2013 16:20, "DubbyTT" notifications@github.com wrote:

Interesting, never heard of plating up one side of the board before. Tape off the bottom? If you can't specify it... you might as well just say, "gimme 1/2 oz or 1oz or 1-1/2oz. ...whichever". My point is if it matters, don't let them give the you wrong thing. If not, then no worries.

;-)

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20215075 .

[technobly]

Plate both sides at the same time plates vias just fine :) You said "top layer" which made me think they were only plating up one side of the board, but taping off the bottom side.

[andyw-lala]

Ah. For RF we only care about traces as deep as the ground plane :) On 29 Jun 2013 22:19, "DubbyTT" notifications@github.com wrote:

Plate both sides at the same time plates vias just fine :) You said "top layer" which made me think they were only plating up one side of the board, but taping off the bottom side.

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20241420 .

[andyw-lala]

OK check out the andyw branch.

Here's what I did:

• added a cap (C12) and overlayed pad with C16. So the two stuffing options are:

u.FL: C16, X2 Built-in: ANT1, C4, L1, C12

I think the impedance bumps will be better this way.

Centred the antenna as best I could

Added RF net class with correct width & specific clearance (had to adjust RF trace clearance down from 45mil, because we're so tightly packed - but it's still > 1 wavelength.)

Added top ground plane & extended under module - removed thermals on that polygon, because it's all SMT/reflow.

Adjusted via stitching (it wasn't stitching to anything before.) Ended up putting via-in-pads on the u.FL - comments welcome. Would like to add more stitching in south east corner of groundplane, but space is tight, and I think extending below the 0.1" header will be useless at RF. Ideas/comments welcome.

I also noted that the u.FL footprint seems to violate the 6/6 DRC rules from the board house, wanted to discuss this before modifying.

Did not run exhaustive DRC.

Am happy designing RF test boards based on this (or agreed upon) design and can test them to evaluate the implementation.

Comments welcome.

On Sun, Jun 30, 2013 at 1:22 AM, Andy Warner andyw@pobox.com wrote:

Ah. For RF we only care about traces as deep as the ground plane :) On 29 Jun 2013 22:19, "DubbyTT" notifications@github.com wrote:

Plate both sides at the same time plates vias just fine :) You said "top layer" which made me think they were only plating up one side of the board, but taping off the bottom side.

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20241420 .

Andy

[technobly]

@andyw-lala did you still need this? https://github.com/sprk/core/pull/16

[andyw-lala]

No - Zach set me up so that I don't for now at least.

However, the final design should be moved to allow people to experiment with the demo version of eagle On 30 Jun 2013 13:07, "DubbyTT" notifications@github.com wrote:

@andyw-lala https://github.com/andyw-lala did you still need this? #16https://github.com/sprk/core/issues/16

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20252391 .

[andyw-lala]

Silk screen will also need work around A0.

On Sun, Jun 30, 2013 at 12:38 PM, Andy Warner andyw@pobox.com wrote:

OK check out the andyw branch.

Here's what I did:

• added a cap (C12) and overlayed pad with C16. So the two stuffing options are:

u.FL: C16, X2 Built-in: ANT1, C4, L1, C12

I think the impedance bumps will be better this way.

Centred the antenna as best I could

Added RF net class with correct width & specific clearance (had to adjust RF trace clearance down from 45mil, because we're so tightly packed - but it's still > 1 wavelength.)

Added top ground plane & extended under module - removed thermals on that polygon, because it's all SMT/reflow.

Adjusted via stitching (it wasn't stitching to anything before.) Ended up putting via-in-pads on the u.FL - comments welcome. Would like to add more stitching in south east corner of groundplane, but space is tight, and I think extending below the 0.1" header will be useless at RF. Ideas/comments welcome.

I also noted that the u.FL footprint seems to violate the 6/6 DRC rules from the board house, wanted to discuss this before modifying.

Did not run exhaustive DRC.

Am happy designing RF test boards based on this (or agreed upon) design and can test them to evaluate the implementation.

Comments welcome.

On Sun, Jun 30, 2013 at 1:22 AM, Andy Warner andyw@pobox.com wrote:

Ah. For RF we only care about traces as deep as the ground plane :) On 29 Jun 2013 22:19, "DubbyTT" notifications@github.com wrote:

Plate both sides at the same time plates vias just fine :) You said "top layer" which made me think they were only plating up one side of the board, but taping off the bottom side.

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20241420 .

Andy

Andy

[zsup]

Thanks @andyw-lala! At quick glance this looks good; it fails drc due to the 45mil spacing for RF and silkscreen issues but otherwise this looks fine.

I don't see the issue you mentioned on the uFL footprint violating 6/6 DRC; where do you see that?

Tonight @satishgn should be continuing on routing for the rest of the circuit, if he agrees I think this is a good design to base RF test boards on.

Zach Supalla 312.953.3413

On Sunday, June 30, 2013 at 12:38 PM, andyw-lala wrote:

OK check out the andyw branch.

Here's what I did:

• added a cap (C12) and overlayed pad with C16. So the two stuffing options
  are:

u.FL: C16, X2
Built-in: ANT1, C4, L1, C12

I think the impedance bumps will be better this way.

Centred the antenna as best I could

Added RF net class with correct width & specific clearance (had to adjust
RF trace clearance down from 45mil, because we're so tightly packed - but
it's still > 1 wavelength.)

Added top ground plane & extended under module - removed thermals on that
polygon, because it's all SMT/reflow.

Adjusted via stitching (it wasn't stitching to anything before.) Ended up
putting via-in-pads on the u.FL - comments welcome. Would like to add more
stitching in south east corner of groundplane, but space is tight, and I
think extending below the 0.1" header will be useless at RF. Ideas/comments
welcome.

I also noted that the u.FL footprint seems to violate the 6/6 DRC rules
from the board house, wanted to discuss this before modifying.

Did not run exhaustive DRC.

Am happy designing RF test boards based on this (or agreed upon) design and
can test them to evaluate the implementation.

Comments welcome.

On Sun, Jun 30, 2013 at 1:22 AM, Andy Warner <andyw@pobox.com (mailto:andyw@pobox.com)> wrote:

Ah. For RF we only care about traces as deep as the ground plane :)
On 29 Jun 2013 22:19, "DubbyTT" <notifications@github.com (mailto:notifications@github.com)> wrote:

Plate both sides at the same time plates vias just fine :) You said "top
layer" which made me think they were only plating up one side of the board,
but taping off the bottom side.

—
Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20241420
.

Andy

— Reply to this email directly or view it on GitHub (https://github.com/sprk/core/issues/1#issuecomment-20251437).

[technobly]

@andyw-lala ah ok...

[andyw-lala]

I heard no review comments or feedback on the RF test board, please proceed with ordering the entire package of boards, stencil and components. I suggest contacting Johanson about selection criteria for the RF caps and inductor.

I'll work with you on assembly when the kits arrive and schedule testing on the VNA and report results back.

The goal is to get reliable performance data (and be able to make a round of mods & retest, if required) and not delay the timeline for the production PCBs.

On Sun, Jun 30, 2013 at 11:23 PM, Andy Warner andyw@pobox.com wrote:

On the andyw branch, I just checked in eagle files for an RF test board. Please review and comment.

Unless we make some changes to the RF layout in the next 24 hours due to review, I would like 20 of these boards ordered, along with components (including the SMA) to populate 10 with antennas and 10 with u.FL. Please also order a metal solder stencil (no frame, just a temp one I can tape in place.)

I will also need 4 u.FL -> female SMA ~6" jumper cables of reasonable quality.

When ordering specify your production PCB stackup, with 7628 PrePreg for the outer layers, 1.6mm overall finished thickness, 1 oz Cu all layers after plating.

If you can order them to arrive next week, that would be great and I will perform testing on a VNA, and also see what I can do about measuring the antenna radiation pattern. With this board, instead of trying to find a CW mode for the CC3000, we can instead hook up to a signal generator, or some other more hackable 2.4GHz source, like a nordic, xbee or similar..

Questions/problems -> me.

On Sun, Jun 30, 2013 at 5:58 PM, DubbyTT notifications@github.com wrote:

@andyw-lala https://github.com/andyw-lala ah ok...

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20257241 .

Andy

Andy

[zsup]

Andy - sorry for the radio silence (pun only partially intended); RF test board looks good, will get them spun up. Hopefully we'll have them in our hands next week.

Zach Supalla 312.953.3413

On Tuesday, July 2, 2013 at 10:42 AM, andyw-lala wrote:

I heard no review comments or feedback on the RF test board, please proceed
with ordering the entire package of boards, stencil and components.
I suggest contacting Johanson about selection criteria for the RF caps and
inductor.

I'll work with you on assembly when the kits arrive and schedule testing on
the VNA and report results back.

The goal is to get reliable performance data (and be able to make a round
of mods & retest, if required) and not delay the timeline for the
production PCBs.

On Sun, Jun 30, 2013 at 11:23 PM, Andy Warner <andyw@pobox.com (mailto:andyw@pobox.com)> wrote:

On the andyw branch, I just checked in eagle files for an RF test board.
Please review and comment.

Unless we make some changes to the RF layout in the next 24 hours due to
review, I would like 20 of these boards ordered, along with components
(including the SMA) to populate 10 with antennas and 10 with u.FL. Please
also order a metal solder stencil (no frame, just a temp one I can tape in
place.)

I will also need 4 u.FL -> female SMA ~6" jumper cables of reasonable
quality.

When ordering specify your production PCB stackup, with 7628 PrePreg for
the outer layers, 1.6mm overall finished thickness, 1 oz Cu all layers
after plating.

If you can order them to arrive next week, that would be great and I will
perform testing on a VNA, and also see what I can do about measuring the
antenna radiation pattern. With this board, instead of trying to find a CW
mode for the CC3000, we can instead hook up to a signal generator, or some
other more hackable 2.4GHz source, like a nordic, xbee or similar..

Questions/problems -> me.

On Sun, Jun 30, 2013 at 5:58 PM, DubbyTT <notifications@github.com (mailto:notifications@github.com)> wrote:

@andyw-lala https://github.com/andyw-lala ah ok...

—
Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20257241
.

Andy

Andy

— Reply to this email directly or view it on GitHub (https://github.com/sprk/core/issues/1#issuecomment-20354086).

[zsup]

I know they gave a few options for the PrePreg (sp?), do we need to specify something there?

Zach Supalla 312.953.3413

On Tuesday, July 2, 2013 at 10:48 PM, Zach Supalla wrote:

Andy - sorry for the radio silence (pun only partially intended); RF test board looks good, will get them spun up. Hopefully we'll have them in our hands next week.

Zach Supalla 312.953.3413

On Tuesday, July 2, 2013 at 10:42 AM, andyw-lala wrote:

I heard no review comments or feedback on the RF test board, please proceed
with ordering the entire package of boards, stencil and components.
I suggest contacting Johanson about selection criteria for the RF caps and
inductor.

I'll work with you on assembly when the kits arrive and schedule testing on
the VNA and report results back.

The goal is to get reliable performance data (and be able to make a round
of mods & retest, if required) and not delay the timeline for the
production PCBs.

On Sun, Jun 30, 2013 at 11:23 PM, Andy Warner <andyw@pobox.com (mailto:andyw@pobox.com)> wrote:

On the andyw branch, I just checked in eagle files for an RF test board.
Please review and comment.

Unless we make some changes to the RF layout in the next 24 hours due to
review, I would like 20 of these boards ordered, along with components
(including the SMA) to populate 10 with antennas and 10 with u.FL. Please
also order a metal solder stencil (no frame, just a temp one I can tape in
place.)

I will also need 4 u.FL -> female SMA ~6" jumper cables of reasonable
quality.

When ordering specify your production PCB stackup, with 7628 PrePreg for
the outer layers, 1.6mm overall finished thickness, 1 oz Cu all layers
after plating.

If you can order them to arrive next week, that would be great and I will
perform testing on a VNA, and also see what I can do about measuring the
antenna radiation pattern. With this board, instead of trying to find a CW
mode for the CC3000, we can instead hook up to a signal generator, or some
other more hackable 2.4GHz source, like a nordic, xbee or similar..

Questions/problems -> me.

On Sun, Jun 30, 2013 at 5:58 PM, DubbyTT <notifications@github.com (mailto:notifications@github.com)> wrote:

@andyw-lala https://github.com/andyw-lala ah ok...

—
Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20257241
.

Andy

Andy

— Reply to this email directly or view it on GitHub (https://github.com/sprk/core/issues/1#issuecomment-20354086).

[technobly]

Just a quick comment here, one of the U.FL pads is going to wipe out the A0 silk screen. Maybe move the A0 silk down a bit?

Also, would you guys mind if I tried to build one of these as well? The only assistance I would need is some beta firmware. I have a slew of projects planned for the Spark Core and would love to get started sooner than later. I only heard about it after all the beta tester slots were taken.

[andyw-lala]

I think I covered that in an email.

Specify 7628 prepreg, 178um (7mil). Final result should be 1oz copper on all layers. Total finished board thickness 1.6mm.

Yes, the core will require some silk-screen cleanup. Didn't care about that until we had agreed on layout. RF test board should have no silkscreen problems. Please advise if anyone sees something. I know they gave a few options for the PrePreg (sp?), do we need to specify something there?

Zach Supalla 312.953.3413

On Tuesday, July 2, 2013 at 10:48 PM, Zach Supalla wrote:

Andy - sorry for the radio silence (pun only partially intended); RF test board looks good, will get them spun up. Hopefully we'll have them in our hands next week.

Zach Supalla 312.953.3413

On Tuesday, July 2, 2013 at 10:42 AM, andyw-lala wrote:

I heard no review comments or feedback on the RF test board, please proceed with ordering the entire package of boards, stencil and components. I suggest contacting Johanson about selection criteria for the RF caps and inductor.

I'll work with you on assembly when the kits arrive and schedule testing on the VNA and report results back.

The goal is to get reliable performance data (and be able to make a round of mods & retest, if required) and not delay the timeline for the production PCBs.

On Sun, Jun 30, 2013 at 11:23 PM, Andy Warner <andyw@pobox.com (mailto: andyw@pobox.com)> wrote:

On the andyw branch, I just checked in eagle files for an RF test board. Please review and comment.

Unless we make some changes to the RF layout in the next 24 hours due to review, I would like 20 of these boards ordered, along with components (including the SMA) to populate 10 with antennas and 10 with u.FL. Please also order a metal solder stencil (no frame, just a temp one I can tape in place.)

I will also need 4 u.FL -> female SMA ~6" jumper cables of reasonable quality.

When ordering specify your production PCB stackup, with 7628 PrePreg for the outer layers, 1.6mm overall finished thickness, 1 oz Cu all layers after plating.

If you can order them to arrive next week, that would be great and I will perform testing on a VNA, and also see what I can do about measuring the antenna radiation pattern. With this board, instead of trying to find a CW mode for the CC3000, we can instead hook up to a signal generator, or some other more hackable 2.4GHz source, like a nordic, xbee or similar..

Questions/problems -> me.

On Sun, Jun 30, 2013 at 5:58 PM, DubbyTT <notifications@github.com(mailto: notifications@github.com)> wrote:

@andyw-lala https://github.com/andyw-lala ah ok...

— Reply to this email directly or view it on GitHub< https://github.com/sprk/core/issues/1#issuecomment-20257241> .

Andy

Andy

— Reply to this email directly or view it on GitHub ( https://github.com/sprk/core/issues/1#issuecomment-20354086).

— Reply to this email directly or view it on GitHubhttps://github.com/sprk/core/issues/1#issuecomment-20394231 .