Dear list,

I have two doubts on 55x DSP.

1) When look at 5509A DSK from Spectrum, I found there was a chip TMS320DA255 besides
TMSVC5509A on the board. It seems TMS320DA255 is also a 55x base audio processor. I also
noticed that the 6713 DSK also posses a TMS320DA250 chip. Why they are put on the DSKs? What do
they do on the DSKs? Why there is no datasheet of TMS320DA25x on TI's website?

2) Not related with the first question. When I move .ssh from DARAM to SARAM in 55x, my code
(written in C in CCS) refused to work. Any reason?

Thanks,

Wei

Li Wei-

> I have two doubts on 55x DSP.
> 
> 1) When look at 5509A DSK from Spectrum, I found there was a chip TMS320DA255
> besides TMSVC5509A on the board. It seems TMS320DA255 is also a 55x base
> audio processor. I also noticed that the 6713 DSK also posses a TMS320DA250
> chip. Why they are put on the DSKs? What do they do on the DSKs? Why there is
> no datasheet of TMS320DA25x on TI's website?

I've noticed this also on the "new generation" of Spectrum Digital DSKs that use
host
USB to connect to JTAG emulation.  For example, I have a DSK 5510 in from of me that
also has the DA250 device.

Here is my take on this -- and I emphasize I don't know this for sure, just
speculating.  Spectrum Digital was faced with three (3) requirements:

  -USB connection was required; someone at TI said "no more
   parallel ports on any DSK cards"

  -TI doesn't like to put competing devices on their DSK cards,
   so that rules out Cypress or FTDI or other simple USB
   controller

  -it must be super difficult to reverse engineer -- Spectrum
   Digital absolutely does not want copy-cat DSK cards to eat
   their business, and they know various outfits around the
   world will spend months trying to reverse engineer anything
   they put out

So their approach was to use a TI chip that a) has an onchip USB port, b) is
programmable and has security protection, c) has onchip ROM program (even more
secure), and d) does not have published documentation without a TI license and NDA.  
The fact the TI DA25x series is normally used in audio applications is irrelevant,
except possibly that it's cost in 10k or higher volume should be relatively low. 
Spectrum Digital seems to have further obscured the host + CCS interface with an
Actel FPGA using nonvolatile "anti-fuse" technology and other security features. 
Now
this combination has got to be unusual and difficult to penetrate.

If you notice, this "chunk of circuitry" seems to be repeated on several recent DSK
cards since 2003.  It looks to have about the same layout each time, and it's always
located near the USB connector.  It works, but there have been costs of this
obfuscation effort, including horribly slow host data transfer rates, lack of full
documentation and accurate schematics for what's supposed to be a "reference
design",
and a mis-direction of engineering effort away from things that should be on DSK
cards, such as Ethernet and disk drive interfaces -- not to mention a user-accessible
USB port.  For example, DSK 6713 and DSK 6416 are slower with MATLAB RTDX than DSK
6711, which came out in 1999 -- did TI expect that to happen?  I wonder if TI Tools
group has thought about these tradeoffs?

-Jeff

Jon-

> I thought about that as well my theory is the DSK board is meant to learn the dsp and code
composer.They would prefer
> in my opinion
> is a person to buy a DSK to learn the software and buy a EVM for development.
> It seems TI with the DSK goes out of the way to prevent development with the dsk board in
general.

Ok so where's the basic reference design for 6713?  5510?  6416?  6455?  If the developers
shouldn't use DSK boards as
reference due to mysterious circuitry that's not documented, then what should they use?

-Jeff

> ----- Original Message ----
> From: Jeff Brower <j...@signalogic.com>
> To: Li Wei <l...@gmail.com>
> Cc: c...@yahoogroups.com
> Sent: Wednesday, January 31, 2007 12:19:36 PM
> Subject: Re: [c55x] two questions on 55x
> Li Wei-
>
>> I have two doubts on 55x DSP.
>>
>> 1) When look at 5509A DSK from Spectrum, I found there was a chip TMS320DA255
>> besides TMSVC5509A on the board. It seems TMS320DA255 is also a 55x base
>> audio processor. I also noticed that the 6713 DSK also posses a TMS320DA250
>> chip. Why they are put on the DSKs? What do they do on the DSKs? Why there is
>> no datasheet of TMS320DA25x on TI's website?
>
> I've noticed this also on the "new generation" of Spectrum Digital DSKs that use
host
> USB to connect to JTAG emulation. For example, I have a DSK 5510 in from of me that
> also has the DA250 device.
>
> Here is my take on this -- and I emphasize I don't know this for sure, just
> speculating. Spectrum Digital was faced with three (3) requirements:
>
> -USB connection was required; someone at TI said "no more
> parallel ports on any DSK cards"
>
> -TI doesn't like to put competing devices on their DSK cards,
> so that rules out Cypress or FTDI or other simple USB
> controller
>
> -it must be super difficult to reverse engineer -- Spectrum
> Digital absolutely does not want copy-cat DSK cards to eat
> their business, and they know various outfits around the
> world will spend months trying to reverse engineer anything
> they put out
>
> So their approach was to use a TI chip that a) has an onchip USB port, b) is
> programmable and has security protection, c) has onchip ROM program (even more
> secure), and d) does not have published documentation without a TI license and NDA.
> The fact the TI DA25x series is normally used in audio applications is irrelevant,
> except possibly that it's cost in 10k or higher volume should be relatively low.
> Spectrum Digital seems to have further obscured the host + CCS interface with an
> Actel FPGA using nonvolatile "anti-fuse" technology and other security features.
Now
> this combination has got to be unusual and difficult to penetrate.
>
> If you notice, this "chunk of circuitry" seems to be repeated on several recent
DSK
> cards since 2003. It looks to have about the same layout each time, and it's always
> located near the USB connector. It works, but there have been costs of this
> obfuscation effort, including horribly slow host data transfer rates, lack of full
> documentation and accurate schematics for what's supposed to be a "reference
design",
> and a mis-direction of engineering effort away from things that should be on DSK
> cards, such as Ethernet and disk drive interfaces -- not to mention a user-accessible
> USB port. For example, DSK 6713 and DSK 6416 are slower with MATLAB RTDX than DSK
> 6711, which came out in 1999 -- did TI expect that to happen? I wonder if TI Tools
> group has thought about these tradeoffs?
>
> -Jeff

Jeff,

Thanks for the enlightenment.

As 5509, I don't see much difference of 5509A DSK and 5509A EVM Plus, except
the later has a SD memory slot. To connect 5509A DSK to CCS via XDS510, I
have to add 5509A EVM's configuration to CCS setup becasue I cannot find any
for 5509A DSK. And it worked.

Wei
On 1/31/07, Jeff Brower <j...@signalogic.com> wrote:
>
>   Jon-
>
> > I thought about that as well my theory is the DSK board is meant to
> learn the dsp and code composer.They would prefer
> > in my opinion
> > is a person to buy a DSK to learn the software and buy a EVM for
> development.
> > It seems TI with the DSK goes out of the way to prevent development with
> the dsk board in general.
>
> Ok so where's the basic reference design for 6713? 5510? 6416? 6455? If
> the developers shouldn't use DSK boards as
> reference due to mysterious circuitry that's not documented, then what
> should they use?
>
> -Jeff
>
> > ----- Original Message ----
> > From: Jeff Brower <j...@signalogic.com <jbrower%40signalogic.com>>
> > To: Li Wei <l...@gmail.com <liwei2000%40gmail.com>>
> > Cc: c...@yahoogroups.com <c55x%40yahoogroups.com>
> > Sent: Wednesday, January 31, 2007 12:19:36 PM
> > Subject: Re: [c55x] two questions on 55x
> >
> >
> > Li Wei-
> >
> >> I have two doubts on 55x DSP.
> >>
> >> 1) When look at 5509A DSK from Spectrum, I found there was a chip
> TMS320DA255
> >> besides TMSVC5509A on the board. It seems TMS320DA255 is also a 55x
> base
> >> audio processor. I also noticed that the 6713 DSK also posses a
> TMS320DA250
> >> chip. Why they are put on the DSKs? What do they do on the DSKs? Why
> there is
> >> no datasheet of TMS320DA25x on TI's website?
> >
> > I've noticed this also on the "new generation" of Spectrum Digital DSKs
> that use host
> > USB to connect to JTAG emulation. For example, I have a DSK 5510 in from
> of me that
> > also has the DA250 device.
> >
> > Here is my take on this -- and I emphasize I don't know this for sure,
> just
> > speculating. Spectrum Digital was faced with three (3) requirements:
> >
> > -USB connection was required; someone at TI said "no more
> > parallel ports on any DSK cards"
> >
> > -TI doesn't like to put competing devices on their DSK cards,
> > so that rules out Cypress or FTDI or other simple USB
> > controller
> >
> > -it must be super difficult to reverse engineer -- Spectrum
> > Digital absolutely does not want copy-cat DSK cards to eat
> > their business, and they know various outfits around the
> > world will spend months trying to reverse engineer anything
> > they put out
> >
> > So their approach was to use a TI chip that a) has an onchip USB port,
> b) is
> > programmable and has security protection, c) has onchip ROM program
> (even more
> > secure), and d) does not have published documentation without a TI
> license and NDA.
> > The fact the TI DA25x series is normally used in audio applications is
> irrelevant,
> > except possibly that it's cost in 10k or higher volume should be
> relatively low.
> > Spectrum Digital seems to have further obscured the host + CCS interface
> with an
> > Actel FPGA using nonvolatile "anti-fuse" technology and other security
> features. Now
> > this combination has got to be unusual and difficult to penetrate.
> >
> > If you notice, this "chunk of circuitry" seems to be repeated on several
> recent DSK
> > cards since 2003. It looks to have about the same layout each time, and
> it's always
> > located near the USB connector. It works, but there have been costs of
> this
> > obfuscation effort, including horribly slow host data transfer rates,
> lack of full
> > documentation and accurate schematics for what's supposed to be a
> "reference design",
> > and a mis-direction of engineering effort away from things that should
> be on DSK
> > cards, such as Ethernet and disk drive interfaces -- not to mention a
> user-accessible
> > USB port. For example, DSK 6713 and DSK 6416 are slower with MATLAB RTDX
> than DSK
> > 6711, which came out in 1999 -- did TI expect that to happen? I wonder
> if TI Tools
> > group has thought about these tradeoffs?
> >
> > -Jeff
>

Wei-
> Thanks for the enlightenment. As 5509, I don't see much difference of 5509A DSK and
> 5509A EVM Plus, except the later has a SD memory slot. To connect 5509A DSK to CCS
> via XDS510, I have to add 5509A EVM's configuration to CCS setup becasue I cannot
> find any for 5509A DSK. And it worked.

Ok glad to hear this progress.  Onward and upward :-)

-Jeff
>   On 1/31/07, Jeff Brower <j...@signalogic.com> wrote:
>
>      Jon-
>
>      > I thought about that as well my theory is the DSK board is meant to
>      learn the dsp and code composer.They would prefer
>      > in my opinion
>      > is a person to buy a DSK to learn the software and buy a EVM for
>      development.
>      > It seems TI with the DSK goes out of the way to prevent development
>      with the dsk board in general.
>
>      Ok so where's the basic reference design for 6713? 5510? 6416? 6455? If
>      the developers shouldn't use DSK boards as
>      reference due to mysterious circuitry that's not documented, then what
>      should they use?
>
>      -Jeff
>      > ----- Original Message ----
>      > From: Jeff Brower < j...@signalogic.com>
>      > To: Li Wei <l...@gmail.com>
>      > Cc: c...@yahoogroups.com
>      > Sent: Wednesday, January 31, 2007 12:19:36 PM
>      > Subject: Re: [c55x] two questions on 55x
>      >
>      >
>      > Li Wei-
>      >
>      >> I have two doubts on 55x DSP.
>      >>
>      >> 1) When look at 5509A DSK from Spectrum, I found there was a chip
>      TMS320DA255
>      >> besides TMSVC5509A on the board. It seems TMS320DA255 is also a 55x
>      base
>      >> audio processor. I also noticed that the 6713 DSK also posses a
>      TMS320DA250
>      >> chip. Why they are put on the DSKs? What do they do on the DSKs? Why
>      there is
>      >> no datasheet of TMS320DA25x on TI's website?
>      >
>      > I've noticed this also on the "new generation" of Spectrum Digital
DSKs
>      that use host
>      > USB to connect to JTAG emulation. For example, I have a DSK 5510 in
>      from of me that
>      > also has the DA250 device.
>      >
>      > Here is my take on this -- and I emphasize I don't know this for sure,
>      just
>      > speculating. Spectrum Digital was faced with three (3) requirements:
>      >
>      > -USB connection was required; someone at TI said "no more
>      > parallel ports on any DSK cards"
>      >
>      > -TI doesn't like to put competing devices on their DSK cards,
>      > so that rules out Cypress or FTDI or other simple USB
>      > controller
>      >
>      > -it must be super difficult to reverse engineer -- Spectrum
>      > Digital absolutely does not want copy-cat DSK cards to eat
>      > their business, and they know various outfits around the
>      > world will spend months trying to reverse engineer anything
>      > they put out
>      >
>      > So their approach was to use a TI chip that a) has an onchip USB port,
>      b) is
>      > programmable and has security protection, c) has onchip ROM program
>      (even more
>      > secure), and d) does not have published documentation without a TI
>      license and NDA.
>      > The fact the TI DA25x series is normally used in audio applications is
>      irrelevant,
>      > except possibly that it's cost in 10k or higher volume should be
>      relatively low.
>      > Spectrum Digital seems to have further obscured the host + CCS
>      interface with an
>      > Actel FPGA using nonvolatile "anti-fuse" technology and other
security
>      features. Now
>      > this combination has got to be unusual and difficult to penetrate.
>      >
>      > If you notice, this "chunk of circuitry" seems to be repeated on
>      several recent DSK
>      > cards since 2003. It looks to have about the same layout each time, and
>      it's always
>      > located near the USB connector. It works, but there have been costs of
>      this
>      > obfuscation effort, including horribly slow host data transfer rates,
>      lack of full
>      > documentation and accurate schematics for what's supposed to be a
>      "reference design",
>      > and a mis-direction of engineering effort away from things that should
>      be on DSK
>      > cards, such as Ethernet and disk drive interfaces -- not to mention a
>      user-accessible
>      > USB port. For example, DSK 6713 and DSK 6416 are slower with MATLAB
>      RTDX than DSK
>      > 6711, which came out in 1999 -- did TI expect that to happen? I wonder
>      if TI Tools
>      > group has thought about these tradeoffs?
>      >
>      > -Jeff
>