Other Parts Discussed in Thread: LM3S1968, LM3S8962, LMFLASHPROGRAMMER, EK-TM4C123GXL, LAUNCHXL-TMS57004
Hi.
I am using a stellaris LM4F120 LaunchPad Evaluation Kit on a windows 8 PC. When I start a debug session with the latest Keil verion(4.72) its automaticly stops and return to the editor.
I have tried with several stellaris kits, and with example projects. I have unistal and reinstal keil and i can´t debug.
Thanks.
1) It seemed to finish loading the code, switched to debug view then switched back to edit view immediately by itself. I lost three LaunchPads (one EK-LM4F120XL, two EK-TM4C123GXL) after that. They are not responding to Keil anymore, not even in a Windows7 system. When I plugged in these three boards to my Windows7 system, the drivers came up but when I tried to debug in Keil, I got
The DK-TM4C123G survived.
2) I used Keil v4.72.10.0 and v5.0.5.15. Both exhibit the same problem.
3) I am not sure what you like to see. There are so many Properties under Details.
Let me know if you need anything else.
At the time the LaunchPad became unusable, I was testing some code exercising the timer capture functions. With the code like the following. If you like to see the whole project, please give me an email address and I will send you the whole project folder.
// Initialize Timer0A to capture rising edge in edge-time mode.
// The input pin of Timer0A is PB6.
void Timer0A_periodCapture_init(void)
{
SYSCTL->RCGCTIMER |= 1; // enable clock to Timer Block 0
SYSCTL->RCGCGPIO |= 2; // enable clock to PORTB
GPIOB->DIR &= ~0x40; // make PB6 an input pin
GPIOB->DEN |= 0x40; // make PB6 a digital pin
GPIOB->AFSEL |= 0x40; // enable alternate function on PB6
GPIOB->PCTL &= ~0x0F000000; // configure PB6 as T0CCP0 pin
GPIOB->PCTL |= 0x07000000;
TIMER0->CTL &= ~1; // disable timer0A during setup
TIMER0->CFG = 4; // configure as 16-bit timer mode
TIMER0->TAMR = 0x17; // up-count, edge-time, capture mode
TIMER0->TAILR = 0xFFFF; // full load
TIMER0->TAPR = 0xFF;
TIMER0->CTL &= ~0xC; // capture the rising edge
TIMER0->CTL |= 1; // enable timer0A
}
// Initialize Timer3A to capture rising edge in edge-count mode.
// The input pin of Timer3A is PB2.
void Timer3A_countCapture_init(void)
{
SYSCTL->RCGCTIMER |= 8; // enable clock to Timer Block 3
SYSCTL->RCGCGPIO |= 2; // enable clock to PORTB
GPIOB->DIR &= ~0x04; // make PB2 an input pin
GPIOB->DEN |= 0x04; // make PB2 a digital pin
GPIOB->AFSEL |= 0x04; // enable alternate function on PB2
GPIOB->PCTL &= ~0x00000F00; // configure PB2 as T3CCP0 pin
GPIOB->PCTL |= 0x00000700;
TIMER3->CTL &= ~1; // disable TIMER3A during setup
TIMER3->CFG = 4; // configure as 16-bit timer mode
TIMER3->TAMR = 0x13; // up-count, edge-count, capture mode
TIMER3->TAMATCHR = 0xFFFF; // set the count limit
TIMER3->TAPMR = 0xFF;
TIMER3->CTL &= ~0xC; // capture the rising edge
TIMER3->CTL |= 1; // enable timer3A
}
// This function captures two consecutive rising edges of
// a periodic signal from Timer Block 0 Timer A and returns
// the time difference (the period of the signal).
int Timer0A_periodCapture(void)
{
int lastEdge, thisEdge;
// capture the first rising edge
TIMER0->ICR = 4; // clear timer0A capture flag
while((TIMER0->RIS & 4) == 0) ; // wait till captured
lastEdge = TIMER0->TAR; // save the timestamp
// capture the second rising edge
TIMER0->ICR = 4; // clear timer0A capture flag
while((TIMER0->RIS & 4) == 0) ; // wait till captured
thisEdge = TIMER0->TAR; // save the timestamp
return thisEdge - lastEdge; // return the time difference
}
int Timer3A_countCapture(void)
{
return (TIMER3->TAPS << 16) + TIMER3->TAR;
}
The LaunchPad was connected to the Windows 8.1 laptop through the USB debug port, a function generator and an oscilloscope. All of them plugged into the same power strip. After the first LaunchPad became unresponsive, I suspect there was ground loop current. To eliminate that, I modified the program to generate pulses from PWM using WTimer5A and installed a jumper between the PWM output to the Timer input so the LaunchPad is completely floating with the USB as the only connection. The other two LaunchPads became unusable with that configuration.
Since I suspected it was hardware problem, I ordered two more LaunchPads. When I plugged in one of them to the same laptop that failed the first three, it bounced out of Keil debug view within a second. I quickly unplugged the board from the laptop. I have been using my old Windows 7 laptop since then and I do not feel comfortable plug these new LaunchPads back to the Windows 8.1 laptop.
When I plug in the first three LaunchPads to the Windows 8.1 and Windows 7 laptop and tried to erase the entire flash, this is what I got when I click the Erase button.
Since I do not want to lose more LaunchPad, I did not use the LaunchPad with the Windows 8 laptop. But I have done several tests with the DK-TM4C123G and the Windows 8 laptop. I tested them with the example programs from TivaWare_C_Series-1.1 with small program as blinky and large program as qs-logger. For these programs, I simply double click on xxx.uvproj to launch Keil uVision, click Build, then click Start/Stop Debug Session. They seemed to download OK but would not stay in debug view.
I have installed the whole suite of Hercules development tools when I attended the Hercules workshop on September 25. The LAUNCHXL-TMS57004 did not work with my Windows 8 laptop. The symptom was that the driver would not stay up. The driver kept disconnecting and reconnecting. I could not complete the exercises of the workshop. The instructors were well aware of the situation. I have not heard anything after that and my Hercules is staying in my storage. I didn't think the Hercules driver have anything to do with Tiva LaunchPad because there was a six week interval before the Tiva LaunchPad failures but I could be wrong.
Petrei,
This is how I interpret your request:
Relating to your picture: can you repeat the Flash Erase with LM Flash Programmer with the unresponsive Tiva LaunchPad and Windows 8.1 and when got the error message, go to configuration tab and snap that window instead of the starting one. My question: did you selected the right clock source for your microcontroller in that tab?
The clock source is fixed when I select TM4C123G LaunchPad.
The solution Craig provided is a great relieve for those who were stranded with Windows 8 and no Keil but it did not address the root cause. We still do not understand how that happened and I am sure Craig and his team is looking into it.
I would like to believe the events that caused my three LaunchPads to become incommunicado at the same time Keil debug stopped working were just pure coincidental. But back in my mind there is a shadow looming. The fact that I offered to ship them to TI for failure mode analysis several times without a taker does not make me feel comfortable.
When I tried to erase these three LaunchPads with LM Flash Programmer, I am still getting "Unable to initialize target -0!" When I tried to use Keil to debug the board I got
When I click on OK, I got this:
After that, Keil is still working with the new boards.
When you used LM Flash Programmer to erase the part what steps did you use? Did you try using the JTAG Unlock? If not can you try this to recover the boards that are unable to connect and let us know if that works?
