Propeller randomly stops executing Spin Code

Comments

• Mike Green Posts: 23,101

  2014-01-10 07:12 edited 2014-01-10 07:12

  You're missing the stack for the Spin cog for "blink" and the Spin interpreter is using memory starting at location 0. Unfortunately, the first few locations in memory contain crucial information and they're overwritten by the "blink" routine as it executes. Look at the section in the Propeller Manual on the COGNEW statement.
• DarkWarrior Posts: 5

  2014-01-10 07:55 edited 2014-01-10 07:55

  So i try this and its the same problem , i have also problems running the graphic demo´s they dont work. So thats why i try to figure it out with this blink code. I think that the propeller hang up after short time total randomly sometimes the 2 leds with spin code work first then hang up or they dont work from beginning.
• T Chap Posts: 4,260

  2014-01-10 07:58 edited 2014-01-10 07:58

  Can you post the newest code?
• DarkWarrior Posts: 5

  2014-01-10 08:06 edited 2014-01-10 08:06

  CON
  {Object_Title_and_Purpose}
  
  
  CON
          _clkmode = xtal1 + pll16x                                               'Standard clock mode * crystal frequency = 80 MHz
          _xinfreq = 5_000_000
  
  VAR
    long  symbol
    long  stack[10]
     
  PUB public_method_name
    dira[16] := 1
    cognew(@entry,0)
    cognew(blink,@stack)
    repeat
      waitcnt(cnt + 20_000_000)
      !outa[16]  
  
  PUB blink
    dira[18] := 1
    repeat
      waitcnt(cnt + 20_000_000)
      !outa[18]     
  
  DAT
                org
  entry                   mov     r0,#1
                          shl     r0,#17
                          mov     dira,r0
  
  loop                    xor     outa,r0
                          mov     r1,cnt
                          add     r1,wait
                          waitcnt r1,#0
                          jmp     #loop                    
                          
  
  
  r0      long  1
  r1      long  1
  wait    long  20_000_000
          
         
  

  
  and when i deleted the method like this its still the same
  

  CON
  {Object_Title_and_Purpose}
  
  
  CON
          _clkmode = xtal1 + pll16x                                               'Standard clock mode * crystal frequency = 80 MHz
          _xinfreq = 5_000_000
  
  VAR
    long  symbol
     
  PUB public_method_name
    dira[16] := 1
    cognew(@entry,0)
    repeat
      waitcnt(cnt + 20_000_000)
      !outa[16]  
  
  DAT
                org
  entry                   mov     r0,#1
                          shl     r0,#17
                          mov     dira,r0
  
  loop                    xor     outa,r0
                          mov     r1,cnt
                          add     r1,wait
                          waitcnt r1,#0
                          jmp     #loop                    
                          
  
  
  r0      long  1
  r1      long  1
  wait    long  20_000_000
          
  
• JonnyMac Posts: 9,550

  2014-01-10 08:19 edited 2014-01-10 08:19

  There is something else afoot -- did you build your own hardware.
  
  Here's an example of blinking three LEDs: one with PASM, one in a secondary Spin cog, one in the main Spin cog. This is running on a demo board so I know it works.
  

  con  
  
    _clkmode = xtal1 + pll16x
    _xinfreq = 5_000_000                                          ' use 5MHz crystal
  
  
  var
  
    long  spinstack[16]
  
  
  pub main | t
  
    pbmask := 1 << 16                                             ' mask for pin 16
    pbdelaytix := clkfreq / 1000 * 250
    cognew(@pasmblink, 0)                                         ' start pasm cog
  
    
    cognew(spin_blink(17, 250), @spinstack)                       ' start spin cog
  
  
    ' run in main cog
  
    dira[18] := 1                                                 ' make output
  
    t := cnt                                                      ' sync with system
    repeat
      outa[18] := 1                                               ' high
      waitcnt(t += (clkfreq / 1000 * 250))                         ' wait
      outa[18] := 0                                               ' low
      waitcnt(t += (clkfreq / 1000 * 250))                         ' wait
  
  
  
  
  pri spin_blink(pin, ms) | t                                     ' launch with cognew
  
    dira[pin] := 1                                                ' make output
  
    t := cnt                                                      ' sync with system
    repeat
      outa[pin] := 1                                              ' high
      waitcnt(t += (clkfreq / 1000 * ms))                         ' wait
      outa[pin] := 0                                              ' low
      waitcnt(t += (clkfreq / 1000 * ms))                         ' wait
  
  
  dat
  
                          org     0
                          
  pasmblink               mov     tix, pbdelaytix                 ' setup delay
                          add     tix, cnt                        ' sync with system
                          or      dira, pbmask                    ' make output
  
  :loop                   or      outa, pbmask                    ' high
                          waitcnt tix, pbdelaytix                 ' wait
                          andn    outa, pbmask                    ' low
                          waitcnt tix, pbdelaytix                 ' wait
  
                          jmp     #:loop
  
  ' ---------------------------------------------------------------------------------------
  
  pbmask                  long    0-0
  pbdelaytix              long    0-0
  
  tix                     res     1
  
                          fit     496
  
• DarkWarrior Posts: 5

  2014-01-10 08:43 edited 2014-01-10 08:43

  Yeah its an DIL40 Propeller build up like its in the Datasheet with an external EEprom and an 5 Mhz Crystal and the 4 Resistors. I program it with an Usb to Serial adapter with the build up also from the Datasheet.
  
  Your Code seems to work.
• JonnyMac Posts: 9,550

  2014-01-10 09:17 edited 2014-01-10 09:17

  It's a small thing and not likely to be causing trouble, but make sure you have a 0.1uF cap between the Vdd and Vss pins on either side of the chip.
  
  I ran your second program (above my last post) and it seems to work fine.
• DarkWarrior Posts: 5

  2014-01-12 08:12 edited 2014-01-12 08:12

  So today i found more time to Test.
  
  I have now 0,1uF caps there, and it seem all the codes with the Led´s work fine.
  
  But the Problem with the Grafic Demos are still there , they are little bit better but not much.
  When i take the GraficsTest.spin that comes with the Propeller Tool, i get only a Black and White screen that starts moving for short time (0.5 - 1,5 sec sometimes 3-5 but not very often) and then hang up.
  I Try other Grafic demos too but non of them work fine, i checked the Crystal and the Resistors.
  
  So any ideas what´s the Problem ?
• JonnyMac Posts: 9,550

  2014-01-12 16:05 edited 2014-01-12 16:05

  Circuit fabrication? If adding the 0.1 caps fixed other problems that seems to indicate that you circuit construction is creating an issue. Why don't you post a photo of your setup so that others can give you meaningful guidance.