Analog Devices ADAU1761

I am contemplating adding a DSP to a design with very, very little
board space.  The current design uses a small FPGA and a stereo CODEC
to receive or transmit a time code signal that uses a PWM symbol
coding modulated on a 1 kHz carrier.  When I was designing the board
the customer asked me to make it flexible enough to use it for other
work including telephony and "high resolution" audio.  So I used an
AKM 4556 CODEC with 24 bit resolution and an audio bandwidth of 20 kHz
and up.

The problem is that the signal processing design in the FPGA has used
most of the space, about 2/3s and the customer is asking if I can add
to it.  It has occurred to me that an MCU or signal processor would
have been a good addition to the FPGA, but I didn't want to add too
much to the cost or the board space since the board is very, very
small (about 4.5 in x 0.86 in).  As is, there is really *no* space for
an additional chip.  But if I use a processor that includes a CODEC, I
can replace the CODEC with the processor.

So I remembered some of the parts that Analog Devices makes and looked
up their SigmaDSP parts.  The ADAU1761 seems like the best fit with a
low price and a very small footprint.  The ADAU1702 is a bit cheaper,
but is larger than the CODEC it would be replacing.

One thing I can't tell about the ADAU1761 is exactly what the DSP
inside can do.  I have only found references indicating that it does
filtering and such with data from the input to produce an output.  I
need to process ADC data to produce a digital stream and/or process
digital data to produce DAC data.  In essense, these are demod/mod
functions.

The development software is called SigmaStudio and that can likely
tell me what I want to know about the DSP functionality.  But you have
to ask to download the eval version.  So I don't know when that will
happen.

Does anyone know if this part can do what I want?  Is there another
part that will fit in a 6 x 6 mm footprint and do signal processing
with included 16 bit min stereo ADC and DACs?

Rick

I spent the last day or so looking at the part data sheet for the
ADAU1761 and checking out the development software.  The part looks
interesting, but I can't quite figure out if the software is something
I want to use.  I contacted support and they told me that this product
is oriented toward "audio processing" with an emphasis on low cost
including their support.  So this tool is purely graphical with no
ability to write assembly language.  I'm not even sure there is a way
to simulate a design, only testing on eval hardware.

I posted this thread nearly 24 hours ago and no one has replied, so I
am guessing that these parts are not very commonly used.  TI makes
some similar parts and I'll take a look at them.

Rick

On Fri, 20 Feb 2009 20:44:07 -0800 (PST), nemo <gnuarm@gmail.com>
wrote:

>I spent the last day or so looking at the part data sheet for the
>ADAU1761 and checking out the development software.  The part looks
>interesting, but I can't quite figure out if the software is something
>I want to use.  I contacted support and they told me that this product
>is oriented toward "audio processing" with an emphasis on low cost
>including their support.  So this tool is purely graphical with no
>ability to write assembly language.  I'm not even sure there is a way
>to simulate a design, only testing on eval hardware.
>
>I posted this thread nearly 24 hours ago and no one has replied, so I
>am guessing that these parts are not very commonly used.  TI makes
>some similar parts and I'll take a look at them.

   I glanced at Analog's webpage and block diagram for it, and it
looks like what you've presumed, all the "DSP" stuff is totally
internal and only does basic audio processing such as volume and maybe
tone control, and it's nothing you can program on your own. I'll also
guess that TI's similar parts do the same thing.

   The area you said in your original post that you have available is
really small (6x6mm) - while I haven't looked in a couple years, I
doubt you could put a "real" programmable DSP in that area, with or
without an audio A/D/A "codec." Maybe someone else knows something
different.

>
>Rick

"rickman" <gnuarm@gmail.com> wrote in message
news:5e22f64f-4d50-43e9-aab4-584ea18e89c2@j38g2000yqa.googlegroups.com...

> One thing I can't tell about the ADAU1761 is exactly what the DSP
> inside can do.  I have only found references indicating that it does
> filtering and such with data from the input to produce an output.  I
> need to process ADC data to produce a digital stream and/or process
> digital data to produce DAC data.  In essense, these are demod/mod
> functions.
> The development software is called SigmaStudio and that can likely
> tell me what I want to know about the DSP functionality.  But you have
> to ask to download the eval version.  So I don't know when that will
> happen.
>
> Does anyone know if this part can do what I want?  Is there another
> part that will fit in a 6 x 6 mm footprint and do signal processing
> with included 16 bit min stereo ADC and DACs?

The CODECs with the integrated DAPs are available from major vendors: AD,
TI, NXP, ST, AKM, Cirrus Logic, JRC and such. The DAPs are the very
primitive and limited machines programmed in the assembly. The processing is
done at every sample as the flow from the input to the output. There is not
much that you can do with it; it is designed for the simple things like
equalization, tone/volume control, stereo enhancement effects. The
programming is usually done via the GUI where you just drag the blocks
together.

Vladimir Vassilevsky
DSP and Mixed Signal Consultant
www.abvolt.com

On Feb 26, 2:13 am, Ben Bradley <ben_nospam_brad...@frontiernet.net>
wrote:
> On Fri, 20 Feb 2009 20:44:07 -0800 (PST), nemo <gnu...@gmail.com>
> wrote:
>
> >I spent the last day or so looking at the part data sheet for the
> >ADAU1761 and checking out the development software.  The part looks
> >interesting, but I can't quite figure out if the software is something
> >I want to use.  I contacted support and they told me that this product
> >is oriented toward "audio processing" with an emphasis on low cost
> >including their support.  So this tool is purely graphical with no
> >ability to write assembly language.  I'm not even sure there is a way
> >to simulate a design, only testing on eval hardware.
>
> >I posted this thread nearly 24 hours ago and no one has replied, so I
> >am guessing that these parts are not very commonly used.  TI makes
> >some similar parts and I'll take a look at them.
>
>    I glanced at Analog's webpage and block diagram for it, and it
> looks like what you've presumed, all the "DSP" stuff is totally
> internal and only does basic audio processing such as volume and maybe
> tone control, and it's nothing you can program on your own. I'll also
> guess that TI's similar parts do the same thing.
>
>    The area you said in your original post that you have available is
> really small (6x6mm) - while I haven't looked in a couple years, I
> doubt you could put a "real" programmable DSP in that area, with or
> without an audio A/D/A "codec." Maybe someone else knows something
> different.

Thanks for your reply.  The problem with both the TI and and ADI
approach seems to be that they are targeting users who don't want to
program DSPs.  In particular, the ADI software gives you a GUI to work
with and does not allow you to program in assembly.  However, that
said, I got a message from a developer at ADI who is willing to help
me with assembly language programming and tools.  But from all other
fronts, they just don't want to get too involved unless I am willing
to spring for an eval kit.  I guess I understand that, but I am just
trying to eval the parts and tools.  I already have enough unused eval
kits sitting around that I will never get the chance to actually do
anything with.  I don't need to spend another $300-$400 on more.

Actually, a DSP is not the best possible approach.  The limitation I
am trying to work around is the size of the FPGA.  Otherwise I would
put the DSP function there.  Then the data flow would be from the
CODEC (highest speed) through the FPGA logic (highest processing
capability) through the processor which has the lowest performance to
match the lower processing requirements finally through FPGA logic
again to handle the particulars of the bit stream interface.

There are some time code standards which require sampling a 100 kHz
carrier.  Of course this rate is pushing the limit of the CODEC I am
using and likely would not be possible at all with the audio
processors.

They are such nice compact devices.

Rick

>On Feb 26, 2:13 am, Ben Bradley <ben_nospam_brad...@frontiernet.net>
>wrote:
>> On Fri, 20 Feb 2009 20:44:07 -0800 (PST), nemo <gnu...@gmail.com>
>> wrote:
>>
>> >I spent the last day or so looking at the part data sheet for the
>> >ADAU1761 and checking out the development software.  The part looks
>> >interesting, but I can't quite figure out if the software is
something
>> >I want to use.  I contacted support and they told me that this
product
>> >is oriented toward "audio processing" with an emphasis on low cost
>> >including their support.  So this tool is purely graphical with no
>> >ability to write assembly language.  I'm not even sure there is a way
>> >to simulate a design, only testing on eval hardware.
>>
>> >I posted this thread nearly 24 hours ago and no one has replied, so I
>> >am guessing that these parts are not very commonly used.  TI makes
>> >some similar parts and I'll take a look at them.
>>
>>    I glanced at Analog's webpage and block diagram for it, and it
>> looks like what you've presumed, all the "DSP" stuff is totally
>> internal and only does basic audio processing such as volume and maybe
>> tone control, and it's nothing you can program on your own. I'll also
>> guess that TI's similar parts do the same thing.
>>
>>    The area you said in your original post that you have available is
>> really small (6x6mm) - while I haven't looked in a couple years, I
>> doubt you could put a "real" programmable DSP in that area, with or
>> without an audio A/D/A "codec." Maybe someone else knows something
>> different.
>
>Thanks for your reply.  The problem with both the TI and and ADI
>approach seems to be that they are targeting users who don't want to
>program DSPs.  In particular, the ADI software gives you a GUI to work
>with and does not allow you to program in assembly.  However, that
>said, I got a message from a developer at ADI who is willing to help
>me with assembly language programming and tools.  But from all other
>fronts, they just don't want to get too involved unless I am willing
>to spring for an eval kit.  I guess I understand that, but I am just
>trying to eval the parts and tools.  I already have enough unused eval
>kits sitting around that I will never get the chance to actually do
>anything with.  I don't need to spend another $300-$400 on more.
>
>Actually, a DSP is not the best possible approach.  The limitation I
>am trying to work around is the size of the FPGA.  Otherwise I would
>put the DSP function there.  Then the data flow would be from the
>CODEC (highest speed) through the FPGA logic (highest processing
>capability) through the processor which has the lowest performance to
>match the lower processing requirements finally through FPGA logic
>again to handle the particulars of the bit stream interface.
>
>There are some time code standards which require sampling a 100 kHz
>carrier.  Of course this rate is pushing the limit of the CODEC I am
>using and likely would not be possible at all with the audio
>processors.
>
>They are such nice compact devices.
>
>Rick
>

Hi all:

Believe it or not the 1761 is a full blown DSP, and ADI does provide tools
for writing assembly code. It does seem like your application should be
pretty straight forward to code graphically. Being a Dso coder myself, why
would you want to reinvent the wheel when there ar tools available that
already do what you are looking for?. If your algorithm was difficult to
implement, by all means go ahead and do the low level coding, for any other
purposes, use the tools.

I used the 1761 in a camera application and I thought I was going to need
to code assembly code for it, but it just so happened that I didn't and my
development time shrunk by 90%. However as expected the other 10% used 100%
of the time (as it usually happens).

ADI does provide a free trial version of SigmaStudio just in case you
don't want to go ahead and purchase the eval board.

Hope this helps.