Reply by November 19, 20042004-11-19
Interestingly and strange enough, the same design, with the same
microcontroller, gives me good results when reading via IDMA from an
ADSP-2188N, same revision (3.2).

Will I have to redesign the PCB to include the series resistors? Damn!

Thank you,

--
Jaime Andr�s Aranguren Cardona
jaac@nospam.sanjaac.com
SanJaaC Electronics
Soluciones en DSP
www.sanjaac.com

(Remove "nospam" from e-mail address)

> > I went crazy debugging my first 21xx design with IDMA until we put series
resistors
> on the control lines. They seem to be _extremely_ sensitive to noise. > > DH
Reply by November 19, 20042004-11-19
Which values, David?

TIA

--
Jaime Andr�s Aranguren Cardona
jaac@nospam.sanjaac.com
SanJaaC Electronics
Soluciones en DSP
www.sanjaac.com

(Remove "nospam" from e-mail address)

> > > > But when reading the contents of PM or DM I get strange results,
spuriously.
> > The data and code is all in overlay 0, I make sure to do so with an
Overlay
> > Latch cycle when writting (the emulator helps me to make sure that it > > worked). For reading I also set to overlay 0 before writting the
starting
> > address from where I want to start reading. But for some accesses the > > IDDMOVLAY and IDPMOVLAY fields in the IDMA Overlay register change to
0xF or
> > something different than 0x0, so I get wrong results with my readings. > > > > > > I went crazy debugging my first 21xx design with IDMA until we put series
resistors
> on the control lines. They seem to be _extremely_ sensitive to noise. > > DH
Reply by David Horwitt November 19, 20042004-11-19
Jaime Andr�s Aranguren Cardona wrote:
> Hello, > > I am doing IDMA reading and writting on an ADSP-2189N, revision 3.2 When > writting, which I use for boot up from a host PC via a microcontroller, > everything goes fine, which I can verify by means of the 218x-EzICE. > > But when reading the contents of PM or DM I get strange results, spuriously. > The data and code is all in overlay 0, I make sure to do so with an Overlay > Latch cycle when writting (the emulator helps me to make sure that it > worked). For reading I also set to overlay 0 before writting the starting > address from where I want to start reading. But for some accesses the > IDDMOVLAY and IDPMOVLAY fields in the IDMA Overlay register change to 0xF or > something different than 0x0, so I get wrong results with my readings. > >
I went crazy debugging my first 21xx design with IDMA until we put series resistors on the control lines. They seem to be _extremely_ sensitive to noise. DH
Reply by November 18, 20042004-11-18
Hello,

I am doing IDMA reading and writting on an ADSP-2189N, revision 3.2 When
writting, which I use for boot up from a host PC via a microcontroller,
everything goes fine, which I can verify by means of the 218x-EzICE.

But when reading the contents of PM or DM I get strange results, spuriously.
The data and code is all in overlay 0, I make sure to do so with an Overlay
Latch cycle when writting (the emulator helps me to make sure that it
worked). For reading I also set to overlay 0 before writting the starting
address from where I want to start reading. But for some accesses the
IDDMOVLAY and IDPMOVLAY fields in the IDMA Overlay register change to 0xF or
something different than 0x0, so I get wrong results with my readings.

I double-checked and no address or overlay latch cycles are being generated,
which could cause this strange behavior. IAL is kept low, and /RD is being
asserted by the micro as well as /IS, while /WR is clean "1". I reviewed the
anomaly list for rev 3.2 and nothing was stated concerning IDMA.

I say it again: the trouble arises only for accessing some addresses, in a
certain order, but I haven't been able to discover a pattern which I could
reproduce and expose here.

I am confident that many people here are experts with these chips, like Leon
Heller and Al Clark, just to mention two of you. So I hope I can have your
valuable advice on how to make my readings bulletproof.

Kindest regards,

--
Jaime Andr�s Aranguren Cardona
jaac@nospam.sanjaac.com
SanJaaC Electronics
Soluciones en DSP
www.sanjaac.com

(Remove "nospam" from e-mail address)