Perfect module (D751749ZHH)

[Mychaela Falconia]

Update for the community: I have ordered 3 of these GTM900-B modules
from Songbosi, and will hopefully get them in the next couple of weeks,
with shipping from China to USA and whatnot.

> PIN 31(reset) (as C118's pow key, pull down to GND shortly)

Hmm, I am guessing that this pin 31 is actually Iota nTESTRESET,
whereas the PWON line (the one that is shorted to GND by the power
button on end user phones) is on pin 15.  If you are starting from a
switched-off state (battery power applied to the chipset, but the VRPC
state machine in the Iota chip is in the OFF state), shorting either
pin to GND will cause a switch-on and boot, but the nTESTRESET
mechanism also allows regaining control from runaway code.  There are
also other important differences which I have only recently discovered:
please read my recently-added Calypso-test-reset article in the
freecalypso-docs Hg repository.

I am also thinking of designing and building a slightly fancier
interface board for these GTM900 modules that will be nicer than the
"dumb" breakout board that merely transforms the FPC interface into a
40-pin header.  My idea is to make a board that will connect to the
GTM900 via FPC on one side, and on the other side it will have:

* a power input connector like on our FCDEV3B (the bright orange one,
copied from TI's own historical development boards);

* a 10-pin header for the two UARTs (the GTM900 brings out both Calypso
UARTs) with exactly the same pinout as the dual UART connector on our
FCDEV3B, so the same dual UART adapters and cables will work;

* an on-board SIM socket connected to GTM900 SIM interface pins;

* PWON and RESET pushbuttons like on the FCDEV3B, grounding GTM900
interface pins 15 and 31, respectively;

* simple headers pins for the remaining few signals.

Once I design and produce this interface board here in California and
prove it working in my lab, I will share the design with Songbosi and
help him (I hope I got the right pronoun, otherwise please correct me)
procure the few simple parts like those Weidmuller power connectors so
he can produce these interface boards on his end in China - then
hopefully he will be able to sell GTM900-B modem modules accompanied
by these FreeCalypso interface boards to anyone interested, thereby
providing a much cheaper alternative to our expensive FCDEV3B.

Just to be clear, a GTM900 modem is not exactly as good as our FCDEV3B.
The three additional hardware features on our FCDEV3B which cannot be
replicated by hacking GTM900 modules are the digital voice channel via
MCSI (dedicated Calypso pins most likely not accessible on the GTM900),
huge memory (16 MiB of flash and 8 MiB of RAM) allowing full FC modem
fw images to be run out of RAM without flashing, and triband RF.  But
for those who live in 900&1800 MHz lands and who don't need huge RAM
and flash and digital voice, the much cheaper GTM900 ought to be an
attractive option.

Regarding PCS and GSM850 band support, while it is possible to modify
the GTM900 to remove its default EGSM and DCS bands and replace them
with GSM850 and PCS, please don't hold your breath for it.  The actual
hw surgery should be quite easy indeed, but the much more difficult
part is knowing which SAW filter part numbers and which L/C component
values should be populated.  Changing the low band from EGSM to GSM850
would be a matter of replacing one SAW filter, but changing the high
band from DCS to PCS (and doing it properly so the resulting board
will work with FreeCalypso and not just OBB) would be a little more
involved.  There would be no need to remove any of the original DCS
SAW filter or L/C components, instead there is a series component
(seems to be a series capacitor acting as a high-pass filter) which
would need to be removed and repositioned to reroute the high band Rx
path from the DCS input to the PCS one.  Then the new PCS band SAW
filter and appropriately-valued L/C components would need to be
populated onto the provided empty pads.  The unknowns to be determined
will be the correct part numbers for GSM850 and PCS SAW filters (need
to be footprint and impedance-compatible with whatever Huawei used)
and the correct L/C component values for the new PCS Rx path.  (The
L/C components in the low band Rx path also may or may not need to be
changed when going from EGSM to GSM850, also remains to be determined.)
I will probably look into it a little when I receive the modules from
Songbosi, but once again please don't hold your breath.

> I checked the voltage of RF switch by changing TSPACT(1) TSPACT(2),
> it should be correct.

In your current application, do you only receive, or do you also
transmit?  With your modified OBB running on these modems, are you
able to successfully transmit in both EGSM and DCS bands?

Hasta la Victoria, Siempre,
Mychaela aka The Mother