[Arm-netbook] Price for A10 dev kit

[jonsmirl at gmail.com]

On Sat, Dec 31, 2011 at 9:47 AM, lkcl luke <luke.leighton at gmail.com> wrote:
> On Sat, Dec 31, 2011 at 2:05 PM, jonsmirl at gmail.com <jonsmirl at gmail.com> wrote:
>> On Sat, Dec 31, 2011 at 2:23 AM, lkcl luke <luke.leighton at gmail.com> wrote:
>>> On Sat, Dec 31, 2011 at 1:20 AM, jonsmirl at gmail.com <jonsmirl at gmail.com> wrote:
>>>> On Fri, Dec 30, 2011 at 8:14 PM, Tom Cubie <tangliang at allwinnertech.com> wrote:
>>>>> On 12/31/2011 08:35 AM, jonsmirl at gmail.com wrote:
>>>>>> Does anyone know the price for the A10 dev kit?
>>>>>> http://www.wits-tech.com/pages/board.jsp
>>>>>>
>>>>>
>>>>> I'm sorry to say that the current price is $1000. It's mainly for
>>>>> enterprise. Our market guys said there will be a re-designed dev board
>>>>> of A10 at the price of less than 1000RMB soon for diyer.
>>>>
>>>> Is there a datasheet available for the A10?
>>>
>>>  yes... but it would be, not being funny or anything, of absolutely no
>>> use to you [see discussion last week].  summary is: the source code is
>>> of far more use to you than the datasheet, which is more in the form
>>> of "note-taking and incomplete reminder to engineers of the existence
>>> of feature x y and z".
>>>
>>>  that having been said, a post was kindly made which has a link to an
>>> unauthorised leaked copy of the datasheet [which, i promise you, will
>>> be of no use to you, i'm looking at the official RHT copy right now
>>> and section 21.2 AC97 Signal Description has a table with the signal
>>> names and that's it].
>>>
>>>> We have an audio application that maybe good for the chip and I need
>>>> to know about the I2S channels.
>>>
>>>  in shorthand, the datasheet says "they exist".
>>>
>>>  section 20.1 it's the usual stuff - PCM, 8-bit u-law or A-law, I2S or
>>> PCM, data rate from 8khz to 192, support of 8 channel output and 2
>>> channel input, exactly as (i hope!) you'd expect an I2S interface to
>>> have.
>>>
>>>  ok there is an AC97 interface [it's referred to as AC97_MCLK, BCLK,
>>> SYNC, DO and DI] shared pins with I2S
>>>
>>>  but there is also an on-board embedded 24-bit Audio Codec *including*
>>> headphone, FM, Line and MIC (all of them stereo).  hm, that's weird,
>>> the FM radio, Line and MIC are IN.  section 22.
>>>
>>>  anyway, again: that just says "these things exist" and so for details
>>> of how it works, you really do have to go to the source code.
>>>
>>>  so, that aside: could you let me know a bit about what it is that you
>>> would like to do?  we have about 6 (!) possibly 8 pins spare on the
>>> 44-pin expansion header (unless another one is added, or unless it's
>>> taken up to 50-pin)
>>
>> The company I work for builds embedded audio processors. They do a lot
>> of math on the audio signals so we are always looking for cost
>> effective ways to get more CPU power. They are currently based on the
>> TI AM3503. The A10 provides a lot more CPU horsepower. We have enough
>> volume that we'd use the A10 directly if we decide that it will work.
>>
>> For your product I'd put the LINE and MIC signals onto the connector.
>
>  if you mean the 68-pin one: no.  been over this.  if you analyse
> dozens of SoCs like i have, you find that there's absolutely no
> standard, across the board.
>
>  68 pins is not enough to have audio as well.
>
>  whatever you pick for audio, whether it be LINE and MIC, or PCM, or
> Class D, or AC97, or this, or that, it doesn't matter _what_ it is you
> "lose".
>
>  the pins on the 68-pin connector are lowest common denominator
> interfaces, and, clearly, as far as audio is concerned, the lowest
> common denominator across even as few as dozens of SoCs, is "oeugg" -
> the empty set.
>
>  we therefore decided "sod it" and to initially say "yep, fine, you'll
> just have to put on a USB Audio IC".  however since then i realised
> that the STM32F which is very little money in large volumes esp. the
> 48 or 64 pin one, could be used to do audio and a hell of a lot more
> besides, in software.
>
>  so, audio on the standard is out.
>
>  _however_, putting it on the Expansion Header is perfectly acceptable.
>
>
>> That will allow someone to hook up a mic and speaker by using a simple
>> amp.
>
>  ... or, they could plug it into the headphone socket on the CPU card.
>
>> Does the chip support Intel HD Audio? If not, it is complicated
>> getting 5.1 out of it. If it does, route that to the edge connector
>> too and then use an external CODEC chip.
>
>
> *     I2S/PCM controller for 8-channel output and 2-channel input
> *    AC97 controller compatible with AC97 version 2.3 standard
> *    Internal 24-bits Audio Codec for 2 channel headphone, 2 channel
> microphone, 2 channel FM
>     input and Line input
>                                      L
>
> so, apparently, yes.

AC97 is not the same as Intel HD Audio.
http://en.wikipedia.org/wiki/Intel_HD_Audio

I don't see an easy way to get 5.1 audio out of the chip.
Does it have an S/PDIF output? Linux knows how to send 5.1 audio out
over S/PDIF.



>
> would it help, therefore, to put the PCM/AC97 pins onto the last
> remaining 8 pins of the 44-pin expansion connector?
>
> l.
>
> _______________________________________________
> arm-netbook mailing list arm-netbook at lists.phcomp.co.uk
> http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
> Send large attachments to arm-netbook at files.phcomp.co.uk



-- 
Jon Smirl
jonsmirl at gmail.com