LCD evaluation and other preparations for FC handset

[Mychaela Falconia]

Hi DS!

> A PCB will certainly be very useful for your testing. As a note, you
> may also consider the possibility of directly soldering some Kynar AWG30
> to your LCD's connector. While this is a bit delicate, it's very doable
> (I did this some time ago on a JTAG's flex cable). In this way you can
> directly test the LCD without waiting for your PCB.

Hmm.  I expect to be picking up those HaoRan LCDs from the mailbox
service tomorrow (they arrived today, but I wasn't able to get to the
mailbox service during their open hours today) and that's when I will
be able to examine them, but I am already expecting a lot of difficulty
with the FPC connector.

Background: some older LCD modules like the one in the Pirelli DP-L10
had FPC tails with a stiffened end meant to go into a ZIF connector.
For Pirelli's LCD this arrangement works out well because the white
plastic shell around the LCD is taller on the sides than in the middle,
effectively providing a recessed area in the middle of the back - this
recessed area is where the body of the PCB-mounted connector goes.
But today's LCDs are thinner and have flat backs, so the same approach
cannot be used with them.

When I disassembled a few mainstream dumbphones of the day some time
in late 2013 or early 2014, I saw that they use thin, flat-backed LCD
modules very similar (in mechanical terms) to the ones I am currently
evaluating.  But they don't use connectors: instead the FPC tail that
comes out of the LCD is soldered directly to the motherboard.  The FPC
tail folds under the LCD and the solder connection between this tail
and the motherboard is thus hidden underneath the LCD - but this whole
construction is extremely thin (some small fraction of a millimetre),
and the LCD is pressed flat against the PCB in the final assembly.

When I started searching for an LCD for our FreeCalypso handset, I
expected to get one with the same solder-down tail arrangement.  But
it looks like the industry mainstay has gone back to ZIF connectors,
as both of the candidate LCDs I am evaluating (HaoRan and Startek)
have FPC tails with ends that are meant to go into a connector, not
the kind that are meant to be soldered.

At first I thought "OK, fine, I'll just put the needed FPC connector
on our board", but then I realized that there is a problem: where will
the body of the connector go?  Two solid objects cannot occupy the
same space at the same time...  How can the LCD rest flat against the
PCB surface if there a connector sitting underneath?  The thinnest FPC
connectors I could find (Hirose FH33 series) have 1.2 mm height above
the PCB, and I have a difficult time imagining how anyone can make an
FPC connector thinner than this.

I had no success in getting an answer out of either LCD vendor as to
how one can use their LCD in a cellphone following the traditional
mechanical design.  My strong preference at this point is to get a
semi-custom LCD module based on an existing one, but modified to
connect through a solder-down FPC tail rather than connectorized.
Both of the vendors I am currently evaluating said they can do such a
modification, but the MOQ (minimum order quantity) would be 1000 pcs,
which gets expensive.  Thus I am thinking that we'll need to use an
existing connectorized LCD during the early prototyping stages, and
put some 1.5 mm thick double-sticky tape under it in order to lift it
above the PCB enough to make room for the connector.

But now there is yet another complication.  The specifications for
those thin (1.2 mm height above the PCB) FPC ZIF connectors say that
the FPC end needs to be what they call "tapered", which apparently
means rounded corners.  Yet the FPC tail coming out of Startek's LCD
has a straight end with sharp 90 deg corners; I have yet to see
HaoRan's LCDs (mailbox service closed at this hour, will need to pick
them up during business hours tomorrow), but every indication so far
is that they will be the same in this regard.  But the only FPC ZIF
connectors that "officially" accept straight as opposed to tapered FPC
ends are "traditional" ones like Hirose FH12 series, and their height
above the PCB is a whopping 2.0 mm.  So what, are we going to have to
lift our LCD 2 mm above the PCB with double-sticky tape when that PCB
itself is only 1 mm thick and the LCD module is 2.35 mm thick?

But it gets even stranger: the LCD connector on Pirelli's motherboard
looks very much like Hirose FH33-30S-0.5SH to me (30-pin version of
this 1.2 mm tall FH33 series), yet the FPC tail coming out of that
Giantplus GPM526A0 LCD they are using has a tail that is just as
straight (sharp 90 deg corners) as the one coming out of the shiny new
Startek LCD!  So how did they insert a straight FPC end into a
connector that is supposedly designed to take those "tapered" ends?
My current plan is to try talking to technical support at Hirose (the
connector manufacturer), perhaps going through Digi-Key Applications
Engineering to get to them - a giant hassle.

So in light of all of these difficulties I am having with the FPC
connector, your idea of soldering 30AWG wires to the FPC tail actually
begins to sound attractive.  But I still don't quite picture how it
would work:

* For the data and control lines, the other end of the wire would need
to go to a pin on an FT2232D breakout board just like the ones we use
for FCDEV3B UARTs - how would this connection be made?

* Several pins on the LCD need to be connected to GND, and several LCD
pins need to be connected to the 3.3V power rail.  There are two GND
pins and one 3.3V pin on the FT2232D breakout board.  How would these
"bus bar" connections be accomplished?

* The LCD has a backlight consisting of 3 LEDs; the LCD is transmissive,
hence these backlight LEDs need to be lit in order for any picture to
be visible on the display.  On this particular LCD the anodes of the 3
backlight LEDs are connected together (come out on one FPC pin) and
need to be connected to the USB 5V rail from the FT2232D board; the 3
cathodes are brought out separately, and each needs its own series
resistor - a 120 ohm resistor between each LEDK pin and GND.  Where
would these resistors go in your proposal?

But if you think you can solve all of the above, I would GLADLY send
you one of these HaoRan LCD sample pieces plus an extra FT2232D board
(identical to the one I assume you already have for FCDEV3B UARTs) if
you can then send me back an assembled and working arrangement.  Heck,
I could even send you two LCD pieces (the sales lady from HaoRan sent
me 5) so you could keep another fully assembled setup for yourself. :)
All of the software needed to make it work resides in the
freecalypso-hwlab Hg repository on Bitbucket.

Any help in this LCD testing and selection process would certainly be
appreciated.

Hasta la Victoria, Siempre,
Mychaela aka The Mother