Forums

Re: Target board connect/disconnect

Started by Jeff Brower September 23, 2006
Tahir-

> Now i am facing a different problem while connecting
> with my custom board(C6414) and DSK(C6416, 1GHz).
>
> When I connect with DSK, CCS shows the "Target is
> connected" then when i reset CPU, sometime it reset or
> sometime error displays and CCS hang.

Are you using an emulator connected to the JTAG connector on the DSK board, or
built-in onboard emulation via parallel port or USB? If emulator, which one?

> In my custom board when i use Debug->reset cpu
> command, instead of conecting target, same error
> displays.
>
> According to my observation:
> 1- may be it is because of gel file then why DSK
> behaviour is not consistent.
> 2- Boot mode selection is wrong?
> 3- I have check the following things:
> Reset signal (TRST) = high when CCS started.
> TCK and TCK_RET = 10 MHz.
> EMU0 and EMU1 pin = showing 0.3 or 0.4 volts (I don't
> understand it right now, may be some interference).
>
> One thing i figure it out that for C6414 processor the
> emu0 and emu1 pin of the processor is internally pull
> up but during board design we have also applied an
> external pull up on these lines (20K resistor, 3.3v).
> What would be the beheviour of these pull ups and is
> this design creating problem?

20k additional pull-up on EMU0 and EMU1 won't have any effect. What about TRST? You
should have a 4.7k pull-down there, to make sure the chip is not in some
indeterminate JTAG mode after power-on. Also I believe you should have 4.7k pull-up
on TCK and TMS -- or at least wouldn't hurt.

-Jeff
Tahir-

> I am using external JTAG emulator (Spectrum Digital
> XDS510PP plus) on DSK, not the embedded usb emulator.
> I have figure out that this is because of EMIF
> settings in GEL file.
>
> I need a little quick help on GEL file. If you can
> give me some segment of code for this, it would be
> great help.

The best starting point would be dsk6416.gel file that comes with CCS. This is a
text file, so you can edit directly.

> The pull-ups/down are intact and according to
> recomendations. I have checked it again.

I would focus first on making DSK 6416 operation reliable. That's standard stuff
from TI, so it should work with no errors, every time. If you document your exact
CCS action sequence on the C6x group, then other people who have that DSK board can
point out what you're doing wrong.

-Jeff

> --- Jeff Brower wrote:
>
> > Tahir-
> >
> > > Now i am facing a different problem while
> > connecting
> > > with my custom board(C6414) and DSK(C6416, 1GHz).
> > >
> > > When I connect with DSK, CCS shows the "Target is
> > > connected" then when i reset CPU, sometime it
> > reset or
> > > sometime error displays and CCS hang.
> >
> > Are you using an emulator connected to the JTAG
> > connector on the DSK board, or
> > built-in onboard emulation via parallel port or USB?
> > If emulator, which one?
> >
> > > In my custom board when i use Debug->reset cpu
> > > command, instead of conecting target, same error
> > > displays.
> > >
> > > According to my observation:
> > > 1- may be it is because of gel file then why DSK
> > > behaviour is not consistent.
> > > 2- Boot mode selection is wrong?
> > > 3- I have check the following things:
> > > Reset signal (TRST) = high when CCS started.
> > > TCK and TCK_RET = 10 MHz.
> > > EMU0 and EMU1 pin = showing 0.3 or 0.4 volts (I
> > don't
> > > understand it right now, may be some
> > interference).
> > >
> > > One thing i figure it out that for C6414 processor
> > the
> > > emu0 and emu1 pin of the processor is internally
> > pull
> > > up but during board design we have also applied an
> > > external pull up on these lines (20K resistor,
> > 3.3v).
> > > What would be the beheviour of these pull ups and
> > is
> > > this design creating problem?
> >
> > 20k additional pull-up on EMU0 and EMU1 won't have
> > any effect. What about TRST? You
> > should have a 4.7k pull-down there, to make sure the
> > chip is not in some
> > indeterminate JTAG mode after power-on. Also I
> > believe you should have 4.7k pull-up
> > on TCK and TMS -- or at least wouldn't hurt.
> >
> > -Jeff
> >
Tahir-

> Thanks Jeff for replying but its too late. My board is
> up and working. It was just a problem in board design.
>
> I am now working with it and once again thanks a lot
> for your help and guidance.

Sure, glad to help. What was the design problem? Please indicate so
others can know about it and help avoid the same situation. Thanks!

-Jeff

> --- Jeff Brower wrote:
>
>> Tahir-
>>
>> > I am using external JTAG emulator (Spectrum
>> Digital
>> > XDS510PP plus) on DSK, not the embedded usb
>> emulator.
>> > I have figure out that this is because of EMIF
>> > settings in GEL file.
>> >
>> > I need a little quick help on GEL file. If you can
>> > give me some segment of code for this, it would be
>> > great help.
>>
>> The best starting point would be dsk6416.gel file
>> that comes with CCS. This is a
>> text file, so you can edit directly.
>>
>> > The pull-ups/down are intact and according to
>> > recomendations. I have checked it again.
>>
>> I would focus first on making DSK 6416 operation
>> reliable. That's standard stuff
>> from TI, so it should work with no errors, every
>> time. If you document your exact
>> CCS action sequence on the C6x group, then other
>> people who have that DSK board can
>> point out what you're doing wrong.
>>
>> -Jeff
>>
>> > --- Jeff Brower wrote:
>> >
>> > > Tahir-
>> > >
>> > > > Now i am facing a different problem while
>> > > connecting
>> > > > with my custom board(C6414) and DSK(C6416,
>> 1GHz).
>> > > >
>> > > > When I connect with DSK, CCS shows the "Target
>> is
>> > > > connected" then when i reset CPU, sometime it
>> > > reset or
>> > > > sometime error displays and CCS hang.
>> > >
>> > > Are you using an emulator connected to the JTAG
>> > > connector on the DSK board, or
>> > > built-in onboard emulation via parallel port or
>> USB?
>> > > If emulator, which one?
>> > >
>> > > > In my custom board when i use Debug->reset cpu
>> > > > command, instead of conecting target, same
>> error
>> > > > displays.
>> > > >
>> > > > According to my observation:
>> > > > 1- may be it is because of gel file then why
>> DSK
>> > > > behaviour is not consistent.
>> > > > 2- Boot mode selection is wrong?
>> > > > 3- I have check the following things:
>> > > > Reset signal (TRST) = high when CCS started.
>> > > > TCK and TCK_RET = 10 MHz.
>> > > > EMU0 and EMU1 pin = showing 0.3 or 0.4 volts
>> (I
>> > > don't
>> > > > understand it right now, may be some
>> > > interference).
>> > > >
>> > > > One thing i figure it out that for C6414
>> processor
>> > > the
>> > > > emu0 and emu1 pin of the processor is
>> internally
>> > > pull
>> > > > up but during board design we have also
>> applied an
>> > > > external pull up on these lines (20K resistor,
>> > > 3.3v).
>> > > > What would be the beheviour of these pull ups
>> and
>> > > is
>> > > > this design creating problem?
>> > >
>> > > 20k additional pull-up on EMU0 and EMU1 won't
>> have
>> > > any effect. What about TRST? You
>> > > should have a 4.7k pull-down there, to make sure
>> the
>> > > chip is not in some
>> > > indeterminate JTAG mode after power-on. Also I
>> > > believe you should have 4.7k pull-up
>> > > on TCK and TMS -- or at least wouldn't hurt.
>> > >
>> > > -Jeff
>> > >
Tahir-

Thanks for the follow-up; others will appreciate your results.

> Few silly problems like:
> 1- pull-ups applied to jumpers for configuration was
> wrong.

Ok, so a few boot config R issues.

> 2- TCK and TCK_ret pin was short and we use extra
> cable for resolving this problem. Due to this noise
> introduce on the EMU0 and EMU1 pin causing problem.

The TCK_ret signal can sometimes be significant. The more DSPs and the
higher the TCK clock rate, the more likely this can be an issue. Normally
we try to obtain it from the "most distant" place that TCK goes on the
board. For example, if TCK goes through a buffer (or more than one)
before it reaches DSP(s), then we take TCK_ret from there.

-Jeff

> --- Jeff Brower wrote:
>
>> Tahir-
>>
>> > Thanks Jeff for replying but its too late. My
>> board is
>> > up and working. It was just a problem in board
>> design.
>> >
>> > I am now working with it and once again thanks a
>> lot
>> > for your help and guidance.
>>
>> Sure, glad to help. What was the design problem?
>> Please indicate so
>> others can know about it and help avoid the same
>> situation. Thanks!
>>
>> -Jeff
>>
>> > --- Jeff Brower wrote:
>> >
>> >> Tahir-
>> >>
>> >> > I am using external JTAG emulator (Spectrum
>> >> Digital
>> >> > XDS510PP plus) on DSK, not the embedded usb
>> >> emulator.
>> >> > I have figure out that this is because of EMIF
>> >> > settings in GEL file.
>> >> >
>> >> > I need a little quick help on GEL file. If you
>> can
>> >> > give me some segment of code for this, it would
>> be
>> >> > great help.
>> >>
>> >> The best starting point would be dsk6416.gel file
>> >> that comes with CCS. This is a
>> >> text file, so you can edit directly.
>> >>
>> >> > The pull-ups/down are intact and according to
>> >> > recomendations. I have checked it again.
>> >>
>> >> I would focus first on making DSK 6416 operation
>> >> reliable. That's standard stuff
>> >> from TI, so it should work with no errors, every
>> >> time. If you document your exact
>> >> CCS action sequence on the C6x group, then other
>> >> people who have that DSK board can
>> >> point out what you're doing wrong.
>> >>
>> >> -Jeff
>> >>
>> >> > --- Jeff Brower wrote:
>> >> >
>> >> > > Tahir-
>> >> > >
>> >> > > > Now i am facing a different problem while
>> >> > > connecting
>> >> > > > with my custom board(C6414) and DSK(C6416,
>> >> 1GHz).
>> >> > > >
>> >> > > > When I connect with DSK, CCS shows the
>> "Target
>> >> is
>> >> > > > connected" then when i reset CPU, sometime
>> it
>> >> > > reset or
>> >> > > > sometime error displays and CCS hang.
>> >> > >
>> >> > > Are you using an emulator connected to the
>> JTAG
>> >> > > connector on the DSK board, or
>> >> > > built-in onboard emulation via parallel port
>> or
>> >> USB?
>> >> > > If emulator, which one?
>> >> > >
>> >> > > > In my custom board when i use Debug->reset
>> cpu
>> >> > > > command, instead of conecting target, same
>> >> error
>> >> > > > displays.
>> >> > > >
>> >> > > > According to my observation:
>> >> > > > 1- may be it is because of gel file then
>> why
>> >> DSK
>> >> > > > behaviour is not consistent.
>> >> > > > 2- Boot mode selection is wrong?
>> >> > > > 3- I have check the following things:
>> >> > > > Reset signal (TRST) = high when CCS
>> started.
>> >> > > > TCK and TCK_RET = 10 MHz.
>> >> > > > EMU0 and EMU1 pin = showing 0.3 or 0.4
>> volts
>> >> (I
>> >> > > don't
>> >> > > > understand it right now, may be some
>> >> > > interference).
>> >> > > >
>> >> > > > One thing i figure it out that for C6414
>> >> processor
>> >> > > the
>> >> > > > emu0 and emu1 pin of the processor is
>> >> internally
>> >> > > pull
>> >> > > > up but during board design we have also
>> >> applied an
>> >> > > > external pull up on these lines (20K
>> resistor,
>> >> > > 3.3v).
>> >> > > > What would be the beheviour of these pull
>> ups
>> >> and
>> >> > > is
>> >> > > > this design creating problem?
>> >> > >
>> >> > > 20k additional pull-up on EMU0 and EMU1 won't
>> >> have
>> >> > > any effect. What about TRST? You
>> >> > > should have a 4.7k pull-down there, to make
>> sure
>> >> the
>> >> > > chip is not in some
>> >> > > indeterminate JTAG mode after power-on. Also
>> I
>> >> > > believe you should have 4.7k pull-up
>> >> > > on TCK and TMS -- or at least wouldn't hurt.
>> >> > >
>> >> > > -Jeff