Am 07.01.2016 um 08:06 schrieb Ursin Soler:
>>
>> I don't think this is too useful because one had to rewrite every
>> single needed I?C slave driver out there for this crude interface
>> and of course the tools, too. These drivers and tools do exist.
>> Why reinvent the wheel?
>
> I totally agree, as long as you want to teach me or point me to a
> good tutorial on how to use this weel... ;)
>
Load the I²C core and I²C hostadapter kernel module where your hardware
is connected to. In case of the DS28E17, it would be the proposed (yet
not existing) i2c-w1 and w1-ds28e17 modules. It gives you another I²C
bus e.g i2c-6.
Then load the kernel module for the I²C slave chip you like to use. E.g
if you have an I²C portexpander like the pca95xx series and compatibles,
load the gpio-pca953x module. Then register the chip on the DS28E17 I²C bus:
# echo pca9554 0x3C >/sys/bus/i2c/devices/i2c-6/new_device
given the chip address is 0x3c on that bus. The kernel then gives you an
additional GPIO chip accessible through /sys/class/gpio, e.g.
/sys/class/gpio/gpiochip248. With
# echo 249 >/sys/class/gpio/export
you export pin 1 of that chip to sysfs (as the new GPIOs may be used by
another kernel functionality if you don't do that). By doing
# echo out >/sys/class/gpio/gpio249/direction
it's an output, and by
# echo 1 >/sys/class/gpio/gpio249/value
you can control the output value. Actually, any GPIO tutorial you find
for the Raspi, Beagleboard or such applies.
For sensor chips it's even simpler, their properties appear in
/sys/class/hwmon as soon as the driver module is loaded and the chip is
registered on the I²C bus.
>
> I think I got th idea - very nice again - but I would need more
> docu/tutorial/background about this. I never used I2C chips this way
> (used them the "arduino" way) - sounds extremely useful! :)
>
Well, see above. The whole magic is registering the chip on the I²C bus
ot is on so the kernel driver can find it. You can then use its
functionality through all the abstractions the kernel gives you, e.g. sysfs.
>
> Ball-grid arrays are quite common today so I have to ask what you
> consider a *professional* company? ;) What was the issue? Did the
> use a reflow oven? What about vapour phase soldering? Given that
> Maxim released this package only I would guess they are quite
> confident about this. BUT I agree it is not optimal. (ps: solder =
> leaded solder ... what else?! ;))
>
Sure. I think it's not possible to half-solder a chip without using an
oven. I've tried half-a-dozen companies which labelled themselves as
"professional" but they don't even get to place chips on prototypes all
in the right direction. What do you learn about that: You need to do all
that stuff yourself if you want to look your your own errors instead of
looking for theirs.
>> Have to read and try a lot there. If you manage to put the DS28E17
>> on a breakout board and send it to me, I could spare the cost for
>> the evaluation kit. Where are you located?
>
> Switzerland. Actually here everything is expensive, but I would be
> willing to spend the money (about 50€) if you give me your address.
>
Nah, I will buy the kit myself then.
>>
>> The most horrifing thing that would happen in real life is they
>> send Matthias and/or me a letter we should stop it. And that's when
>> I, in all kindness, would ask "Are you serious? But you do
>> understand you got the Linux driver for your chip for free just
>> because we did this."
>
> This is intressting... Do you know of anybody that happened to? I
> know the story of dvddecrypter only - but I guess this is not
> comparable since its software...
>
Ahh, that's Hollywood. They are just crazy. Too much cocaine. No
reasonable thoughts coming from such persons.
>> As written above, you get that as a free gift as soon the kernel
>> driver creates an /dev/i2c-... device for the DS28E17 I?C
>> interface.
>
> Honestly I think that would be REALLY cool! Wireless 1wire for short
> distances only would allow to cross glass windows and go outside
> (weather station, etc.) without any issue... ;)
>
I think so. However, as it's onewire tunneled through onewire, it would
be extra-slow, I think.
Kind regards
Jan
------------------------------------------------------------------------------
_______________________________________________
Owfs-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/owfs-developers
