C1000,c3xxx Stereo In, Internal Mic, Wifi, Bt

[speculatrix]

some discussion of hardware hacks came up on the angstrom user's mailing list* so I thought people here would be interested.

It began with an off-the-cuff remark by Stanislav Brebec about stereo audio input, I asked for more information. he explained that the Zaurus x000 series are capable of stereo sound input, it's just that Sharp decided to ground one of the inputs, here's the HOW-TO

http://lists.linuxtogo.org/pipermail/angst...May/001888.html

that message also goes on to discuss internal microphone, wifi and bluetooth - click on the "next message in thread" to see the follow-ons.

the hardware discussion then carries on in this thread:
http://lists.linuxtogo.org/pipermail/angst...May/001892.html


* I am not on the dev's list as I got fed up with the bitching

[utx]

It was me.

The WM8750 audio codec inside is capable of: 3 multiplexed stereo inputs (RINPUT3 is repurposed as headphones detection), two stereo outputs (dependent), two mono outputs.

Zaurus wiring:
LOUTPUT1/ROUTPUT1 -> Jack (additional mute transistors are present in the path)
LOUTPUT2+ROUTPUT2 -> Speaker
OUT3 -> unused, traced to testpad near the codec and near the "+" sign
MONOOUT -> unused, traced to testpad near the previous
LINPUT1 <- Jack
RINPUT1 <- completely wired audio path connected to ground
LINPUT2, RINPUT2, LINPUT3 <- unused, but traced somewhere on PCB
RINPUT3 <- repurposed as headphones detection

Non-audio hwmods in that mail have its own thread here.

Parts list for stereo in:
capacitor 220p/50V (SMD 0402/1005) BK
resistor 47k (SMD 0402/1005) J
a short wire

Part that will be removed:
one coil

See the image for my idea. It is possible, that there is an open position for a zero resistor somewhere for this modification.

Please note, that this hwmod is still only my theory. There is no success story yet.

[pelrun]

Well, first things first, RINPUT1 is not just tied to ground; it's connected to pin 2 on the jack. I don't think it's a "historical" leftover or anything - smartest thing to do is to look at eviljazz's driver source to see exactly what is being done with that signal. (headphone/microphone detection perhaps?)

Easier to target RINPUT2/LINPUT2 (pins 25/26 on the chip) anyway - they are not connected to anything on the schematic, and the pictures of the board only show a couple of stub traces covered by solder mask, no vias. Since the chip pins are accessible and larger than those traces, it's probably easiest just to solder straight to them.

[utx]

Well, first things first, RINPUT1 is not just tied to ground; it's connected to pin 2 on the jack. I don't think it's a "historical" leftover or anything - smartest thing to do is to look at eviljazz's driver source to see exactly what is being done with that signal.

RINPUT1 is not connected to pin 2 directly, but via a coil (which could be removed).

I have been experimenting with this signal a lot. The result is the latest alsa-state package in oe.dev, which uses RINPUT1 for the semi-differential mono input. (It is grounded, so it is not a real differential input.) In theory it could give better signal/noise ratio. In practice, I hear no difference. Z seems to be well wired.

I have no idea which WM8750 driver was written by eviljazz. Could you be more specific?

Easier to target RINPUT2/LINPUT2 (pins 25/26 on the chip) anyway - they are not connected to anything on the schematic, and the pictures of the board only show a couple of stub traces covered by conformal coat, no vias. Since the chip pins are accessible and larger than those traces, it's probably easiest just to solder straight to them.

It would be much better to use RINPUT1 - all needed parts (capacitors, PCB traces) are already present there, and the device will remain compatible with unmodified Zaurus.

RINPUT2 and LINPUT2 are wired on the top side to a metalized holes and then traces continue on the reverse side of PCB. I don't have needed photos to trace them yet.

[utx]

I have no idea which WM8750 driver was written by eviljazz. Could you be more specific?

Guessing from this thread, he improved a 2.4 OSS driver. I studied 2.6 ALSA driver for my experiments, but I guess, that the results will be the same: There should be no problem to use stereo-in setup instead of differential mono setup.

[pelrun]

RINPUT2 and LINPUT2 are wired on the top side to a metalized holes and then traces continue on the reverse side of PCB. I don't have needed photos to trace them yet.

Ok, I was wrong, those points do go to vias, but they only go to the next layer down, not the other side of the board (full depth vias are bigger.) However the only place they eventually go to are test points N501 and N502; the schematic is specific. No, I don't know where those points are, nothing I have gives a physical location of the test points. For all I know they're under the CF connector...  

You're free to pursue your option, of course; I just know what route I'll be taking when I get around to doing this mod myself (I investigated it a long time ago, actually - but only just got brave enough to open my Z a couple days ago  )

[Ragnorok]

- Speaking of of brave, the last time I tried messing with the insides of a low-power device, I fragged it.  Had grounding wires on me, the iron, etc but zatched it none the less.  Since then I've been really gun shy about doing things like, oh, dismantling and soldering away at my irreplaceable Hiroshi.  (owlish stare)
- I would, however, very much like to put at least wifi inside.  Are there any pointers on how one who's familiar with audio-grade soldering would go about messing with the the Z's innards and have some reasonable assurance that the first touch of the iron won't end in tears?
- BTW I'm typing this on Hiro and these letters are tiny...