[Gta04-owner] GTA04 hardware suggestion and discussions

Dr. H. Nikolaus Schaller hns at goldelico.com
Sun Jan 23 11:14:39 CET 2011

Hi Glenn,

Am 22.01.2011 um 23:34 schrieb Glenn:

> Hi All!
> Here are some hardware suggestion for discussion:

thank you very much for these deep insights and proposals! Let's see what we can do...
Christoph has already commented some so I will focus on additional aspects.


> 1) Possibility of IR RC5 receiver/transmitter
> 2) Possibility of IRDA transceiver

is planned - unfortunately the 4MBit/s capable IrDA module we have designed in (LiteOn HSDL3602)
is discontniued. We haven't searched deeply for a good replacement yet because it does
not have a high priority. But we have secured a handful of the last stock from DigiKey. So we
can build some boards with this IrDA interface.

It is mounted on the backside of the AUX pushbutton so it should look through the same
transparent button. At least from theory.

> 3) Possibility of a "video-link"? Use touch-screen as a color matrix (2D) barcode video transmitter - and the camera as a receiver (half or full duplex). Let the link automatically adapt to to half duplex, if you need to turn your smartphone to establish the other half duplex connection. The link should also adapt frame rate and if the camera and/or screen is dirty - or the link is through a dirty window. I have no idea how fast the link can be as a function of distance. The link should automatically adapt if the video color matrix (2D) barcode is mirrored in any way - or "played" backwards.

Nice idea! The simplest pattern would be a single pixel (i.e. modulating the display backlight).
Although the bit rate will be quite slow... Nevertheless the DSP of the OMAP3 should
be ideal for image processing.

The other idea from Christoph is also nice to use vibra-motor + accelerometer :)

BTW, the vibra-motor is connected to the TPS65950 so that it should (according
to the data sheets) be capable of being modulated by an audio signal. I.e. vibra-audio...

> 4) Antenna diversity of (preferably pattern, space - or polarisation diversity):

Well, this is a task of squaring a circle. We simply have to find a tradeoff between:

* size of device
* availability of chips (the newest ones are either not yet available or too expensive and
   older ones are not optimal and we risk that they will be discontinued before we have
   optimized everything - it is like driving an old car: while you repair one end, something
   breaks at the other one)
* release date of the device (e.g. we have to decide one day to freeze the PCB layout)
* time/equipment/cost (for very sophisticated experimenting - AKAIF Apple did spend
   150 Million USD for the first iPhone and we don't even have 0.1%)

So we have thought a lot about it but I am sure there are still many improvements possible.
Some within the GTA04 versions (e.g. GTA04A4, A5, A6 etc.) and some have to be
postponed for a GTA05 (which is our dream to have a new case design).

Generally for the antenna design we have very narrow design constraints. One is that
there isn't much room in the case and almost no space on the PCB. Most of the proposals
in the links you have attached assume that you have some 50 x 75 mm space on a PCB.
This is already between one third and half of the whole board.

And the other constraint is that we have a lot of metal by the display frame and the battery.
All this leaves only some mm for antennae.

> * GSM/UMTS? (receive diversity incl. external antenna).

Here, the GTM601 provides a second antenna connector for UMTS diversity but we did not
find a location to add a second antenna.

One is that we have to reuse the existing GSM antenna module of the GTA02 since
we don't have the tools to design and produce a different one (plastics + metal strips).

This is already a severe compromise since the antenna appears to be optimized for the 800/850/900
MHz band and for 1800/1900 MHz (equivalent circuit seems to be two interleaved LC resonators).

We assume that this is quite ok since UMTS uses 1900/2100 MHz and we assume that
1900->2100 isn't too far away.

Nevertheless it is clear that this is not the optimal antenna needing future improvements.
Some ideas for further study exist. I.e. using a flex-pcb instead of metal strips so that we can
much easier get samples at a reasonable price ~50 EUR to run several loops of optimization.

> * 802.11(a)bg(n) (diversity) (W2CBW009Di: "For Dual Antenna application") Almost every laptop has diversity.

Ah, that appears to be a new chip that we were not aware of. But is it already available? If it is
a drop-in replacement for the W2CBW003 we already have two 2.4 GHz antennae - currently
one is for Bluetooth and the other one for WiFi. If a new chip can reuse them for diversity this
could be a simple solution. If it requires major changes I would postpone it until the new chip
becomes better available (e.g. through DigiKey or Mouser).

Placing these antennae on the PCB was already a compromise for the WiFi antenna since
it had to be placed in a way that it is partially shielded by the display metal frame. So we will
see if the performance is acceptable.

> * FM-receiver (receive diversity requires 2*Si4721 - uC selects the best signal (listen to one - let the other turn on and scan when the listen-signal has to low S/N ratio) - or you can listen to one and record another).

Well, this already works with a single Si4705/4721. They have two antenna inputs and can
switch between both. We did follow the reference design and have reused the shield of the
headset cable as the first antenna.

The second antenna (also oused as TX antenna on the Si4721) can be connected to solder
pads so that we can experiment with different antenna designs. The reference proposal would
be to draw a loop around the whole PCB - but this collides with the plastics case and the display.
And we don't have to forget the battery...

So we have to experiment a little with this device.

> * GPS? (receive diversity including external antenna).

This is an area where we may be able to do something. By switching to the GTM601 we
did get an integrated GPS receiver "for free". For the GTM501 version (the one we are
currently experimenting with) doesn't have that. There are two reasons why we have not yet
dropped the separate GPS receiver.

One is that the GTM601 receiver appears to be less sensitive. The other is that the
wiring of the antenna is much more difficult since the GTM601 and the GPS antenna plugs/switch
are on opposite ends of the PCB. Since the antenna signals are very sensitive we would have to
draw a tunnel through the board. This collides with some layout constraints (even on a 8 layer
board). So we currently have only added pads to mount a U.FL connector at the GPS input
for doing experiments.

> For FM-diversity you can use a 180 degree hybrid (two loaded LC-circuits?) to separate common-mode and differential-mode - for diversity - if three space divided poles are available. I know that there is a capacitance between ear-wires and their ground, so I do not know, how well it will work. Note that some impedance matching might be needed.
> For VHF/FM the head-phone cable might be used. The diversity might be implemented isolating ground by a adequate inductor (100MHz parallel LC-circuit) - and two high-pass filter from left and right earphone non-ground wires - and the GTA04 itself as a kind of "ground"/third-wire.

> -
> Links:
> Matrix (2D) barcodes:
> http://en.wikipedia.org/wiki/Barcode#Matrix_.282D.29_barcodes
> PCB-antennas (small):
> http://www.numatechnologies.com/pdf/foilantennas.pdf
> An antenna that can be sized for 100MHz reception around the perimeter of the PCB? Use two loops so ground is not needed:
> http://www.hard-core-dx.com/nordicdx/antenna/loop/k9ay/k9ay_orig.pdf
> http://www.hard-core-dx.com/nordicdx/antenna/loop/k9ay/index.html
> Also for 100MHz:
> (T2FD -- The Forgotten Antenna: http://www.hard-core-dx.com/nordicdx/antenna/wire/t2fd.html )
> 1/3/2003 Antenna Diversity Strengthens Wireless LANs:
> http://www.eetimes.com/design/communications-design/4008941/Antenna-Diversity-Strengthens-Wireless-LANs
> Quote: "...While detailed implementations of dynamic diversity can vary widely, it's possible to summarize the simulated performance of a conceptually simple scheme relative to that of a single-antenna mode, as well as full diversity and selection diversity (Figure 6)..."
> http://en.wikipedia.org/wiki/Antenna_diversity
> Quote: "...
> Spatial diversity employs multiple antennas, usually with the same characteristics, that are physically separated from one another. Depending upon the expected incidence of the incoming signal, sometimes a space on the order of a wavelength is sufficient.
> ...
> Pattern diversity consists of two or more co-located antennas with different radiation patterns.
> ...
> A polarisation difference of 90° will result in an attenuation factor of up to 34dB in signal strength. By pairing two complementary polarizations, this scheme can immunize a system from polarization mismatches that would otherwise cause signal fade. Additionally, such diversity has proven valuable at radio and mobile communication base stations since it is less susceptible to the near random orientations of transmitting antennas.
> ..."
> http://web.archive.org/web/20060427002449/www.tvhandbook.com/support/pdf_files/Chapter17_7.pdf
> Quote: "...
> Diversity is only of value for the relatively rapid fading caused by multipath. This kind of fading is encountered extensively at HF and for tropospheric and iono- spheric scatter propagation at higher frequencies. Similarly, mobile vehicles passing through a static multipath field at VHF and UHF encounter such fading as a result of their motion.
> ..."
> 180 degree hybrid:
> http://upload.wikimedia.org/wikipedia/commons/6/62/Combine2.svg
> "Sum" uses sum of ear-wires - and third-wire.
> "Difference" to some extent uses only ear-wires.
> Pattern Diversity:
> 90 degree hybrid:
> http://upload.wikimedia.org/wikipedia/commons/6/62/Combine2.svg

This appears to be the same link as for 180 degrees.

> "X" uses ear-wires in one direction.
> "Y" uses ear-wires in the other one direction.
> 180 degree hybrid and 90 degree hybrid might be combined. One 90 and two 180 degrees? Then you have 4 receive pattern diversity output ports.
> http://www.option.com/en/products/products/embedded-mobile-broadband/lgamodulegtm601-609/specifications/#start
> Quote: "...
> Simultaneous Equalization and Rx Diversity on all bands (Advance Receiver Type 3i), except for GSM/GPRS/EDGE
> ...
> Primary, diversity and GPS antenna contact pads
> ..."
> regards,
> Glenn
> _______________________________________________
> Gta04-owner mailing list
> Gta04-owner at goldelico.com
> http://lists.goldelico.com/mailman/listinfo/gta04-owner

More information about the Gta04-owner mailing list