RE:[avr-gcc-list] JTAG ICE Problem (Off Topic)

[daniel boals]

James-

 

      I had almost the exact same problem just a couple of days ago. Although mine was with the atmega128 at 16Mhz.  It seems I had a cold solder somewhere between the clock pins and the Crystal.  A quick way to see/test this is to change the fuses to run the uC off of the internal RC oscillator and see if the problem goes away. If everything works after this, there is a good chance there is something wrong with your external circuit. For the atmega128 there is a fuse CLKOPT that will cause the clock to do rail-to-rail clock oscillations.  This also made my problem go away.  I do not know if the 169 has this same ability.

     

      In my case I believe I was not actually running at a full 16Mhz and, I am guessing here, but I believe this is the manifestation of trying to talk to the JTAG port at a speed greater than 1/4 the uC clock.

 

      I was able to get rid of my problem simply by touching the clock pins on the uC and the pins on the Crystal with a soldering iron and adding a little extra solder in both places.  I may have also gone over the pads on the caps in that circuit as well, I can't remember.

 

Hope this helps,

 

Daniel Boals

 

------------------------------

 

Date: Mon, 18 Aug 2003 15:07:03 -0400

From: James Dabbs <address@hidden>

To: avr <address@hidden>

Subject: [avr-gcc-list] JTAG ICE Problem (Off Topic)

Message-ID: <address@hidden>

Content-Type: text/plain

MIME-Version: 1.0

Precedence: list

Message: 2

 

This has turned out not to be a GDB problem, but I had a note ready to go..

 

I am having trouble with our JTAG ICE.  We are using it with our board, based on a 16MHz ATMega169.  Basically, the ICE "looses touch" with the AVR. When I break, if I look at registers or examine breakpoints, the application code just continues running and all of the variables and registers shown in the debugger appear as 0xFF.  The software runs *properly* in this state, just without any connection to the debugger, which has to be restarted to regain control.  This happens with an assembler project under AVRStudio 4.07.  If I move quickly, I can step through code, but as soon as I pause for more than a couple of seconds, the application code starts running and leaves the ICE behind.  The ICE connection has been verified, the voltage is within 2% of 5V, and the oscillator is 16.0000MHz.  I believe this is either an electrical problem with our board (although it is an infuriatingly simple

PCB) or possibly a problem with 16MHz parts.  Has any one had similar troubles?

 

Thanks!