ePC: Mezzanine

ARNAUD · Posted: Tue Oct 11, 2011 8:06 am Post subject: ePC: Mezzanine

Dear Support,

1st point
We need the mechanical drawing of the SBC-ComEx RevF. It is an ePC high performance (with i7)
The customer is manufacturing a mezzanine to access at JN1, JN2, JN3
Mezzanine board will be mounted on the SBC-ComEx.

Please send us this drawing

2nd point
On the JN3 connector, they are signal without explanation:
HEADPHONE_OFF#
AC_BITCLK
AC_RST#
AC_SDOUT
AC_SYNC

Can you give us details?

3rd point:
On the JN1 connector, they are signal without explanation:
DIO38
TRIG4
TRIG5

Can you give us details?

rvanbuskirk · Joined: 20 Jul 2009 Posts: 162

1st Point:
I attached the Fab drawing for the SBC-Comex Rev F.

2nd Point:
AC_BITCLK, AC_RST#, AC_SDOUT, AC_SYNC are control signals from the COM Express Module to an audio device that you can put on your Mezzanine board. We use the STAC9200 which uses these four signals. HEADPHONE_OFF# is an additional signal provided on JN1 but not used on our Mezzanine board.

3rd Point:
I don't see DIO38 on the schematic. Can you let me know where you see it?
Trig4 and Trig5 are signals from the SBC-Comex FPGA. I don't think they are used at this time.

See the following site for information about our Mezzanine board:
http://www.innovative-dsp.com/ftp/SbcComEx/Mezzanine/

ARNAUD · Posted: Wed Oct 12, 2011 7:21 am Post subject:

Thank you for your answer.

You can find enclosed a picture showing this I/O (page 111/138 of the technical manual of ePC)

rvanbuskirk · Joined: 20 Jul 2009 Posts: 162

Thankyou for pointing me to this.
The DIO38 in the manual corresponds to the FDIO_CLKP on the schematic, so it is the clock for the DIO bits.
Best Regards

ARNAUD · Posted: Thu Oct 13, 2011 10:59 pm Post subject:

Thanks for your answer.

Regarding the mechanical drawing, the customer needs also the dimension of eight hole which fixing the SBC on the rack.

In fact, his custom I/O mezzanine will be fix on these 8 holes.

rvanbuskirk · Joined: 20 Jul 2009 Posts: 162

Please find two documents attached.
1. Cool_Hole_Dimensions.png
2. drl001.lst (from Gerber directory)

The dimensions are in mils.
The drill size for the holes are 150 mil.
The drl001.lst shows the X, Y Coordinates with a preceeding "1" which I believe is due to the panel when the board is made. The rest of the numbers in drl001.lst match the coordinates I saw in our layout tool properties window for each hole.

Best Regards,
Ron

ARNAUD · Posted: Mon Oct 24, 2011 8:25 am Post subject:

Dear Support,

I had the customer on the phone and there is conflicting information.
If you checked on the technical manual, the page 58 “Mezzanine Mechanicals”, it is said that JN3 connector is 0.010” shifted (in y axis) compare to JN1/JN2.

The customer checked on files posted on the forum (“A9R52A.tmp.pdf” and “drl001.txt”) and it is never mentioned that Pin1 of JN3 connector is shifted of 0.01” compare to JN1/JN2.

What is the correct information?
Is JN3 is shifted (in y axis)?
Please confirm.

rvanbuskirk · Joined: 20 Jul 2009 Posts: 162

I verified the SBC-Comex.pdf manual on pg58 does show a shift of 0.010". However, the actual pcb design shows no y-axis shift. I attached the pin properties window of each connector JN1-JN3 pin 1. They all show the same y coordinate.

I also checked a SBC-Comex board with a micrometer to verify. I don't see a y-axis shift.

Engineering said it is possible that an early rev of the board may have had this shift but the shift is not in our current design.

ARNAUD · Posted: Wed Oct 26, 2011 8:05 am Post subject:

Thanks for your information.

By reading the "DRL01.TXT" file, in the Drill section: .057 (line 2522).
When we calculate the spacing between the center of holes of the Molex connector, we found that the spacing is "1374.02 mil" whereas in the attached datasheet, it is gave "1354 mil"

How should we interpret this difference of 2.20 mil?

ARNAUD · Posted: Wed Oct 26, 2011 8:12 am Post subject:

I made a mistake, the difference is 20.02 mil

rvanbuskirk · Joined: 20 Jul 2009 Posts: 162

We have not had any complaints about the part not fitting correctly on the board, but however, it appears that you are correct. Our boards have a spacing of 1374 mils which differs from the 1354 recommended by the data sheet.

The hole diameter of our board is 0.057", and the diameter of the plug is 0.046". I can see this is visible on an actual board. The plugs actually reside on the inner side of the holes, so due to the hole being 10 mils larger than the plugs, the part still fit.

I would recommend following the datasheet. Note that the data sheet recommended layout has the diameter of those holes at 0.055", which is slightly smaller than our 0.057".