Work toward FreeCalypso handset

Mychaela Falconia mychaela.falconia at
Wed Sep 29 02:18:05 UTC 2021

Hello FreeCalypso community,

I am still slowly trudging toward my dream goal of designing and
building my own FC Libre Dumbphone that would serve as a source-enabled
replacement for the Calypso-based but sans-source Pirelli DP-L10.  The
purpose of this post is to share some updates on current developments
toward this goal.

The biggest update at the moment is that I have finalized the selection
of LCD module that will be used in our Pirelli-replacing FC handset,
and I have bought 100 pcs of this chosen LCD module, a quantity which
I expect to be a lifetime buy, covering all anticipated prototyping
iterations.  The LCD module I settled on is a 176x220 pixel 2.0" TFT
made by LCD manufacturer Formike; here is the datasheet:

In terms of the physical size of the display active area in mm, our
2.0" diagonal LCD is about 72% bigger than Pirelli's 1.5" diagonal:
the AA of our new LCD is well-defined to be 31.68x39.60 mm; for
Pirelli's LCD we have no specs, but measuring the lit-up AA with a
ruler shows about 26x28 mm.  But in terms of information display
capacity, our new 176x220 pixel LCD is 2.36 times larger than
Pirelli's 128x128 pix!  In any case, my choice of 176x220 as the pixel
size for our phone LCD comes from my desire to reuse as much as
possible from TI's D-Sample demo/prototype/PoC UI which is coded for
this LCD pixel size, and the physical size of 2.0" diagonal comes from
the industry standard size selection: most 176x220 pixel LCDs including
the one I ultimately ended up choosing are made in this physical size,
not some other.

I also took a few pictures, although my skills with the camera are
abysmal - I am an engineer, not a photographer:

In these pictures you can see the front and back of a bare LCD module,
and our lunalcd2 test board which I made a couple of months ago.  This
test board has the LCD module mounted on it (key mounting feature: see
how the module's FPC tail is folded under the body of the module in
our mounting arrangement), and the LCD is connected to our Luna
interface from 2020.

The strip of Kapton tape seen in the lunalcd2.jpeg photo is a hack for
holding our LCD module in place.  In a complete phone handset with
plastics, there will be no issue: in the classic mechanical design of
all standard phones, there are screws that hold together the front and
back plastic pieces, and the resulting tight grip structure constrains
the LCD module to staying where it needs to be.  However, in a bare
development board setup without plastics, we get this unexpected
annoyance: there is a certain elastic force (originating in the area
where the tail is folded under the module) that pushes the bottom of
the LCD module up, away from the PCB, this force acts against gravity
and against the adhesive, but it is strong enough to lift the bottom
of the LCD module in a way that at least to me looks aesthetically
displeasing and feels like a significantly weakened system.  On
lunalcd2 I am using a strip of Kapton tape wrapped around the board to
keep the LCD module from lifting up at the bottom end (the undesirable
elastic force is countered by the tension of this tape), but as my
next iterative solution, I am commissioning a local sheet metal shop
to make us a custom metal bracket that will be more robust than the

Right now I am waiting for this hired local sheet metal company to
finish the design and fabrication of our LCD retaining bracket, and I
am also waiting for this back-ordered part from Digi-Key:

This part has been on backorder since July 1; originally it was
supposed to arrive around the time of my surgery in late August, but
now it says October 11.  The purpose of this DIP switch is to select
the backlight LED current from 1 to 15 mA in 1 mA increments, allowing
us to evaluate the readability of the LCD under different ambient
lighting conditions with different backlight LED currents, with the
goal of selecting a subset of just 4 possible LED currents for our
Venus board.  Because of this insanely long backorder issue, I got our
current lunalcd2 board (the one pictured) assembled back in August
using a substitute part, a DIP switch pack with less convenient flush
actuators instead of raised ones.  But now that we are getting a
special retaining metal bracket made, I plan on producing a lunalcd3
test board (like lunalcd2, but with appropriate mounting holes added
for the bracket) before proceeding with FC Venus, and given that
Digi-Key's promised delivery date of Oct 11 is approaching, I am
hoping to use the better DIP switch part on lunalcd3.

At this point you have every right to ask: so what is the big deal,
why am I fussing so much over the LCD and its mounting?  The answer is
that I am planning our next FreeCalypso development board, and this
next FC Venus board will be a major advancement toward our FC Libre
Dumbphone handset goal: it will be just like a complete phone, but in
the physical form factor of a bare board, without any plastics or
mechanical design.  With our current FC development boards so far,
that is FCDEV3B and Caramel2, you need to connect the board to a host
computer and control it with AT commands in order to make it do
anything interesting, as in connecting to GSM networks, making and
receiving calls, sending and receiving SMS.  But FC Venus will be on
the next level: the board will include the just-described LCD and
keypad buttons in addition to the Calypso GSM core, and it will be
designed to be powered either by a fixed supply like our current
boards or by a real Li-ion battery, with charging of that battery
managed by the Calypso chipset and our firmware!  My vision is that I
will be able to mount this Venus board plus its powering 18650 battery
on some sturdy substrate such as a thick sheet of acrylic, and take it
out with me on social outings, demonstrating it to people outside my
cave - and given that I live right next to USA-Mexico border and given
my ability to cross this border freely back and forth without being
subject to any Covidian cult rituals, I hope to be able to demo this
next stage of FreeCalypso in both countries.

Right now we have only one FC hardware platform on which we can run
firmware that is controlled with an LCD+buttons UI, rather than via AT
commands, and this platform is FC Luna.  However, this Luna platform
is very unwieldy - it is not a single board, but a set of 4 boards
interconnected with ribbon cables: a Caramel2 motherboard, a lunalcdX
board, a keypad board and a DUART28 adapter for the host computer
interface.  Here is a fairly recent photo of the whole mess:

As you can surely tell from the above photo, this Luna setup is not
exactly suitable for field demos to audiences with limited attention
spans where first impression is everything, even if I went through the
process of mounting all of these pieces on some solid substrate like
our dear Das Signal did.  Thus the logical next step is clear: we need
to eliminate the maze of ribbon cables and integrate everything on one
board.  Furthermore, instead of using a Tango module for the Calypso
core like we did on Caramel2/Luna, our Venus board will be made from
"raw" Calypso and accessory chips like FCDEV3B - this approach will
make both design and fabrication more expensive, but it will give us
access to a few additional chipset signals which iWOW didn't bring out
on their TR-800 module.  By connecting a magnetic buzzer to Calypso
BU/PWT output like on TI's D-Sample board, we will produce a platform
for making a smooth transition from buzzer ringing to Melody E1, and
by connecting to all 3 Iota audio channels, we will prototype the
complete audio routing subsystem for the Pirelli-replacing handset.

If you would like to read more about this Venus board plan and about
the more distant handset for which it will serve as a prototype, here
is the detailed design specification:

There is just one more complication which I am currently working on:
we need to get a custom wired headset made for us.  Trying to use any
of the commonly available already existing headsets won't work:

* Today's mainstream consumer 3.5 mm headsets for Apple/Android phones
won't work because they are stereo (left and right ear), whereas we
need monaural, as in only one earbud.

* Openmoko headsets won't work for the same reason: while they are
infinitely better morally and philosophically than sheeple ones, they
are still stereo.

* Old historical headsets made for classic historical dumbphones (both
Motorola C1xx and non-Calypso classic phones like Nokia) are still
available in substantial quantities on certain markets, but they all
feature the typical 3-wire TRS arrangement with the earpiece speaker
connected between a dedicated headset output wire and ground.  But for
our own FreeCalypso headset I desire a different arrangement which I
haven't seen used by any historical player other than iWOW: the
arrangement I seek is to use a 4-wire TRRS plug, but still keep it
monaural, with 2 non-ground wires used to connect the single earpiece
speaker as a bridge-tied load.

My idea for the custom headset I'm seeking to get made came from iWOW.
In addition to making a huge production run of TR-800 modules (so huge
that I consider the NOS supply of these modules to be inexhaustible
for the purposes of our community), they also made a small handful of
their development board kits - but these iWOW development kits are
much rarer than bare TR-800 modules, Das Signal and I found only two
of them, and the two kits we have scored may well be the last two in
the world.  On their development board, iWOW came up with the really
neat idea of bringing out the *main* Iota audio channel to a headset
jack - this idea is rather revolutionary, as the main audio channel is
meant to be used for the phone's built-in earpiece and mic, whereas a
typical finished phone's headset jack would normally be wired to the
secondary or tertiary Iota audio channel.  But Iota's main audio
output is differential and the phone's built-in earpiece speaker is
meant to be connected to it as a bridge-tied load - thus a traditional
3-wire TRS headset cannot be connected to this main audio channel in
the native and straightforward manner as pioneered by iWOW.

iWOW's solution was to replace the common 3-wire TRS headset with
4-wire TRRS, even though it is still monaural and NOT stereo.  The
headset which iWOW supplied with their development boards has its
earpiece speaker connected between Tip and Ring2 as a bridge-tied
load, and the microphone is connected between Ring1 (positive) and
Sleeve (ground).  But the problem with those iWOW-supplied headsets is
that there are only two of them in the world: I have one, our dear Das
Signal has the other, and there are no more.  My answer to this
situation is to get more of these special headsets custom-made for us
on a special order basis: I am the kind of person who will NEVER
accept a no for an answer, and if iWOW were somehow able to get such
unusual headsets, then we must do likewise.

Our Caramel2 board has its main audio channel brought out to a TRRS
headset jack in iWOW's pinout just like iWOW's original DSK board,
thus anyone who wishes to play with voice call audio on C2 will need
one of these special headsets which I am now trying to get made.  But
even more importantly, I plan to keep the same arrangement on our
Venus board, i.e., bring out the main or primary Iota audio channel to
a headset jack - thus the special headset will be a required accessory
for FC Venus too.  I expect FC Venus to be a more interesting board
than C2, thus the headset issue is now becoming a big deal.

Right now I am actively working on the headset issue, that is, looking
for a headset manufacturer who would be willing to do what we need.
So far I have reached out to two manufs by direct email; if I don't
get a response from either of them, as the next step I will need to
brave Alibaba's rather difficult web interface and post a buy request
or RFQ or whatever they call it.

Hasta la Victoria, Siempre,
Mychaela aka The Mother

More information about the Community mailing list