FreeCalypso status update

Mychaela Falconia mychaela.falconia at gmail.com
Sat Jun 30 23:55:07 UTC 2018


Hi DS!

> My personal preference goes to approach 1, the creation of a handset.

Yes, that is my own personal interest too, but see below regarding the
difficulties.

> Right now I have more use of a phone that does voice calls/texts than
> a modem (although I agree a modem might be useful too!)

It is obvious: a complete phone is something that you can carry in
your pocket or purse and use in everyday situations; OTOH a modem like
our FCDEV3B can make phone calls too if you equip it with a loudspeaker
and a microphone (so you *can* use it to call in an order for a pizza
or whatever), but because it requires a host computer for control and
in the present form also a bunch of additional components (FT2232x
adapter, power supply, speaker) and interconnecting wires, it can only
be used practically from a stationary setup such as home.

> For me the
> least useful option seems to be approach 2 (an SMT module).

I don't have any personal use for one either.  Instead I was trying to
attack the status quo with that idea: right now it seems that everyone
has accepted that the "proper" way to make a "free" phone is to use a
smartphone architecture with separate application and modem processors,
and make it only half-free with the modem being proprietary.  Everyone
is doing it: GolDeliCo, Neo900, Librem 5, ZeroPhone, you name it.  My
idea was to strike at that ecosystem: essentially tell them "you can
keep that two-processor smartphone architecture you love so much, but
why don't you use our published-source modem module instead of that
proprietary one you use now".

But of course those people seem to love their 100% closed and
proprietary modems so much that trying to market a FreeCalypso modem
to them may be casting pearls before swine...

> I realize that approach 1 is the most exensive, however if I may suggest
> a good part of the plastic shell could be made with 3D printing/laser
> cutting instead of the more traditional and expensive approach of a cast.

Actually the plastic case is not the main concern at all - I am not
really worried about it because it will only be needed at a very late
stage in the handset project.  Instead the biggest difficulty is going
to be the part that will be needed at the very beginning: the main
board.  The main issue with the motherboard won't be the cost of
physically making it (I expect it to be about the same as FCDEV3B),
but the difficulty of getting it designed.  There will be two big
issues with creating the PCB design for the handset motherboard:

1) No matter what PCB design software we do it in, doing the layout
for our handset board (adding the LCD, the keypad and other peripherals
to our existing Openmoko/FCDEV3B modem core) will require a large
amount of highly skilled and highly specialized labor.  Such labor
does not come cheap, and I expect that we will need to spend somewhere
around 10 kUSD just on the *desk job* of creating the PCB design.

2) I am sick and tired of being held hostage to Altium, and it is my
goal to do the handset board design in a proper FLOSS tool.  KiCAD is
not compatible with my way of doing things, instead my plan is to
create my own fork based on geda-pcb or pcb-rnd - the latter is
already a fork of the former, thus I am considering a fork of a fork.
A fork will be needed because one of the features I absolutely require
is fundamentally incompatible with the vision of both geda-pcb and
pcb-rnd maintainers.

So basically while I still very much desire a FreeCalypso Libre
Dumbphone handset, it probably won't happen any time soon.  But I am
now thinking of another idea for the shorter term: I wonder if we can
create a kit based around our existing FCDEV3B that can be made
portable.  It would still require an external host (practically
meaning a laptop) for control, but if we can eliminate the requirements
of an external power supply and of having a bunch of little pieces
that have to be interconnected with wires in an ESD-safe environment,
it would become possible to bring the kit to meetings and social events
for demonstrations, as long as it's an event where one can sit down
with a laptop.  I am thinking of two possibilities for power:

Option 1: use an 18650 battery.  In this case I am thinking of an ad
hoc laser-cut plexiglas box that will contain an FCDEV3B modem board,
an off-the-shelf FT2232x adapter board for control host connection,
an 18650 battery holder and an off-the-shelf battery charger board.
The whole setup would need to be enclosed so it can be carried around
and pulled out for demonstration without ESD concerns, so only the two
USB connectors would be exposed: one on the FT2232x adapter board and
the other on the battery charger board.  There would also need to be a
power-on button (a pushbutton switch) on the outside of the enclosure,
connected to the PWON header on the FCDEV3B, so the power-on press can
be given without having to touch the board in an ESD-unsafe manner.

Option 2: design and build a special adapter that would allow the modem
to be powered from USB without needing a battery.  It is possible to
feed 5V power directly to our FCDEV3B as-is (all of the components
powered directly from the VBAT rail are tolerant of 5V), but it will
be bad in the long term: because all of the regulators are linear and
not switchers, feeding 5V to the board on a long-term basis would
cause greater heat dissipation and reduce the life expectancy.  Thus a
proper USB-powered solution would need a power front-end in the form
of a switching DC/DC converter to bring the voltage down to somewhere
around 3.5 V, and do it efficiently with a switcher instead of a linear
regulator that takes the difference and dissipates it as heat.

Philosophically the USB-powered option is more proper: if a modem
absolutely requires an external host for control and can't do anything
without such external host control, it does not make a whole lot of
sense for it to have its own battery power supply, instead it should
take power from the same external system which it is serving, right?
But then doing the USB-powered approach would require designing and
building a custom USB front end board (basically an FT2232x board that
also incorporates that DC/DC converter from 5 V to 3.5 V and brings
out that supply on a connector designed for high current), whereas the
battery and charger option can be concocted from off-the-shelf parts.
I can also think of scenarios where the controlling laptop itself runs
on battery power, where it would be advantageous for the modem to have
its own battery and draw less power from USB.

We also need a working loudspeaker, both for the modem demo kit based
on FCDEV3B and for the eventual handset board.  So far the only way I
have been able to exercise voice calls and Melody E1 ringtone
generation on the FCDEV3B is by connecting the loudspeaker and mic
inside TI's D-Sample handset to our board.  This arrangement proves
that the loudspeaker driver and microphone input circuits on our board
are good, but of course the problem with using TI's D-Sample handset
is there is only one of it in the world.  We need a loudspeaker of our
own to replace TI's.  Fortunately I found someone local to me who is
willing to work with me on the loudspeaker issue, so hopefully we will
have some solution somewhat soon.

And of course we need to make the new FCDEV3B V2 boards themselves.
My latest update on this front is that earlier this week I had a
discussion with Technotronix (the people who assemble our boards), and
we came to an agreement as to which package variant we should use for
for the 74AXP1T34 buffer.  I initially chose Nexperia's SOT1202
package (1.0 x 1.0 mm with 0.35 mm pad pitch), but my contact at
Technotronix told me that it is too small: they may be able to do it,
but it would be better to not risk it.  So we are going to move to the
next larger package: Nexperia's SOT886, which is 1.0 x 1.45 mm with
0.5 mm pad pitch.  Earlier this morning I emailed my Altium PCB layout
contractor, telling him that we need to make this package change from
SOT1202 to SOT886, and he told me that he should get back to me in
another couple of days with an estimate for the labor cost.  Hopefully
I will be able to cover that cost with the money from the recent sale
of the last 3 FCDEV3B V1 boards.  (The boards were sold to a commercial
entity working on a specialized non-standard application, not related
to the libre phone community mission of FreeCalypso.)

So this is my game plan for the next few months, in this order:

1) Get the new FCDEV3B V2 boards made;
2) Work with the local gentleman from my local LUG on the loudspeaker
issue;
3) Find some local people (must be local, not across the Internet) to
work with me on the enclosure idea, the portable demonstration kit
based around our FCDEV3B modem.

The last part (the portable kit) is what I hope will be the main
outcome of this plan.  Of course the plan to build a practically
usable handset is still on the agenda, but it is so distant (probably
many more years) that we need something else in the shorter term to
keep FreeCalypso relevant.  I need something that I can rub into
people's faces, so to speak: every time someone says that FreeCalypso
is not free software because it's illegal, I want to be able to whip
out a FreeCalypso phone, make a call from it in front of them, and
tell them: "here I am committing a federal felony in front of you,
using a phone/modem without FCC approval, and also using source code
which I treat as free sw despite you telling me it isn't - so if it's
illegal, why don't you call the police right now and have me arrested
right here on this spot in front of you" - but I don't have a working
FreeCalypso phone that I can carry around with me to do what I just
said, instead all I have is a development board that requires a lab
bench setup...  But if we can turn the FCDEV3B into a portable kit
that still requires a laptop for control, but not a complete lab bench
with a grounded ESD mat and external power, I would be able to bring
it to social meetups, demonstrate it working (making a point to
emphasize its law-breaking nature), and basically engage in protest
acts to make FreeCalypso more visible.

And yes, the people to whom I will make such protest demos will
include the local San Diego office of TI, the place where they used to
work on 3G before everything got shut down and thrown in the trash.
Another good place to make highly visible public protest actions
demonstrating FreeCalypso would be in front of FSF headquarters in
Boston, protesting their camp's refusal to accept FreeCalypso as
free sw.

Hasta la Victoria, Siempre,
Mychaela aka The Mother


More information about the Community mailing list