FCDEV3B update

[Mychaela Falconia]

Hi Serg,

> I had a chance to speak with some RF experts and the answer was that
> changing material, thickness and etc. may be not a big deal or may render
> the end product unusable. They need to look deeper and asked to prepare a
> write up with both materials specifications, images of the RF part of the
> PCB stackup and schematic diagram of the same part. Mychaela if you could
> draft it, I will be able to pass it for evaluation and possibly have a
> followup call to discuss details.

OK, let's take it point by point.  Material specifications: in the
case of Openmoko's original, the only information we have are the
NAMES of the materials they used, as in:

* "PP 1078 61%" for the L1-L2, L2-L3, L6-L7 and L7-L8 dielectrics;
* Two plies of "PP 1080 65%" for the L3-L4 and L5-L6 dielectrics;
* "Core FR4" for the L4-L5 dielectric.

I am going to ask pcbcart for the spec sheets of the materials they
propose using.

Images of the RF part of the PCB stackup: please elaborate what exactly
you are asking for.  The original PCB stackup which pcbcart say they
can't reproduce is in this PDF:

ftp://ftp.freecalypso.org/pub/GSM/GTA02/GTA02-MB-A6_fab_notes.pdf

It is easily readable without knowing any Chinese.  The altered
stackup proposed by pcbcart is here:

https://www.freecalypso.org/members/falcon/fcdev3b/pcbcart/stackup-20170119.png

Or if by "images of the RF part" you meant the layout of the RF
circuits, the gerber files are here:

ftp://ftp.freecalypso.org/pub/GSM/FreeCalypso/fcdev3b/fcdev3b-gerbers-20161223-official.zip

Schematic diagram of the RF part: it is fully contained on page 17 of
this PDF:

ftp://ftp.freecalypso.org/pub/GSM/GTA02/GTA02_Schematic_MB_A5_1220.pdf

Notice that this schematic page has a stackup drawing in the upper
left corner, but it is a little different from the one in the fab
notes.  I can only assume that the one in the fab notes is what FIC/Om
actually built, and that the different one on the schematic page is
some kind of older draft.

> From the initial conversation I learned that more dense weaving is
> generally better.

The original materials were PP 1078 in the outer layers and PP 1080 in
the inner layers.  I don't actually know what these numbers mean, i.e.,
I don't know how dense or not-so-dense the weave is.  Pcbcart haven't
told me specifically what type of PP they plan on using, although they
gave me the impression that they are going to use PP 1080 throughout.
I will ask them to tell us more specifically.

> There are two important parameters we need to look at when compare two
> materials,
>
> 1. Dielectric constant Er, which should not deviate more than by 0.2 in the
> nominal frequency range compared to the original material.

As you can see in Openmoko's fab notes (GTA02-MB-A6_fab_notes.pdf),
they just say "Er 4.3" for every dielectric layer, be it core or
prepreg.  In the case of pcbcart's altered stackup materials, first
they said "Er 4.2", but I asked them what causes the difference
between 4.2 and 4.3, they actually told me this:

: Our ER is within 4.1--4.6,

Can you please ask your RF people is such variance in the actual Er of
the dielectric is normal, or does it mean that pcbcart are being sloppy
and need to control it tighter.

> 2. Dielectric Loss Tangent Tan # , which should be equal or less than in
> the original material.

This one I know nothing about at all.  We could probably ask pcbcart
what this "tan delta" is for their dielectrics, but how would we find
out what it was for Openmoko's original if we only know the names
"PP 1078 61%" and "PP 1080 65%"?

In any case, I'll ask pcbcart for the specific types of prepreg they
propose using - they at least owe us this much information - and go
from there.

M~