• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 83 subscribers
  • Views 1350 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

28335 Control Card - Test with different crystal (24MHz instead of 20MHz)

Markus Hofinger
Prodigy 90 points

Hi all!

I'm currently building a basebord for the 28335 Control Card.
I'm accessing an ADC, and a DAC throug the McBSP in SPI mode.
As I need to get an exact 6kHz cyclic frequency, I need to change the crystal, because the 6kHz can't be realised with the 20MHz crystal.
As even a 30MHz crystal should be possible with this DSC their shouldn't be a problem with exchanging the crystal.

Unfortunatly the code doesn't work any more after changing the crystal. (It worked with the 20MHz crystal and 2kHz cyclic frequency).
My settings for the PLL are the same as for the 20MHz crystal  x10 /2 (20MHz * 10/2 = 100MHz, 24MHz *10/2 = 120MHz)

I've built my code upon the existing example project for the McBSP inside the sprc530 - C2833xC2823x CC++ Header Files and Peripheral v1.2.
I've adapted every occasion of the 150MHz, and 100MHz define in the code order to fit 120MHz system frequency.

I am using the XDS100 debugger from the experimenters kit, and it even freezes during debugging in step over mode. However  it doesn't freeze in single step mode.

 

Does anyone have an idea what could be wrong?

Is there a special procedure that needs to be taken before setting up the PLL ?
Does the debugger need to be recalibrated or something else?
Are there some settings for the project that need to be changed that I haven't found so far?

I attached the whole project, but the 4 files that I changed compared to the original files are:
McBSP_Test/DSP2833x_common/include/DSP2833x_Examples.h    (This file contains the pll settings)
McBSP_Test/DSP2833x_common/source/DSP2833x_Mcbsp.c           (This file contains the code for the McBSP)
McBSP_Test\DSP2833x_common\source/DSP2833x_SysCtrl.c          (This file contains minimum waitstates required for the flash operating)
McBSP_Test/Main_C_Code/mcbsp_test/Example_2833xMcBSP_DLB.c   (This is the main c file)

Thanks

Markus

McBSP_Test.zip
  • 0
    Genius 5885 points

    Hi Markus!

    Was the 24MHz crystal specified to work at fundamental or third overtone frequency? You should compare what frequency you get on the XCLKOUT pin with 20MHz and 24MHz crystal. I know that you can convince the on-chip oscillator to work with third overtone crystals with additional parallel resistance with the crystal with value around 5-10 kOhm. I don't know what impact this has to long term stability and operation of the oscillator. For additional info in oscillators check Fairchild Semi application note 340.

    Regards, Mitja

  • 0 in reply to Mitja Nemec
    Prodigy 90 points

    Hi Mitja!

    Thanks for your quick reply!
    The crystal that I've ordered (datasheet) is specified to work at fundamental frequency. I've measured the voltages across the crystal. I used 1:10 probes in order not to influence the oscillator to much. I don't know about the specific voltage levels for the DSC, but I assume that this should be enough.

    Unfortunatly I can't measure the XCLKOUT pin, because the control card uses a ball grid array, and the XCLKOUTpin isn't routed to any via, or at least I couldn't find it.
    But thanks for this tip, I will keep it in mind for the next time!

    I've already ordered a second control card with the original crystal, so that I can do a direct comparism...

  • 0 in reply to Markus Hofinger
    Genius 5885 points

    Hi Markus!

    The other thing to check is the core supply voltage. From the symptoms I suspect that the core voltage on control card is set to 1.8 V which supports maximum CPU frequency of 100 MHz. In order to work above 100 MHz you should set core supply voltage to 1.9 V. If you can not do that, set the CPU frequency to 60 MHz and test the code

     

    Regards, Mitja

  • 0 in reply to Mitja Nemec
    Prodigy 90 points

    Hi Mitja!

    You are right, the core voltage is indeed set to 1.8V. It might be possible to simply exchange the feedback resistors of the step down converter on the control card. (Schematic of the Control Card). However i will need to wait till monday to do this...
    I had a look in the datasheet of the F28335 about bout this, but unfortunatly there is no voltage level specified for 120MHz. As the voltage specified for 100MHz is 1.8V and the for 150MHz is 1.9V (on page 110) one could assume that it should be 1.82Volts. This voltage should then also be in the min, max range for both 100MHz and 150MHz. Unfortunatly there is no core voltage vs. frequency graph...
    Anyhow I need to wait for monday to get a chance to test it.....

    Thanks for your fast and good answers!

  • 0 in reply to Markus Hofinger
    Prodigy 90 points

    Hi Mitja!

    It works!
    Today I went back to breadboard testing, and put the control card back into the original experimenters kit and just downloaded the code.
    And surprisingly the code doesn't freeze any more during debugging.
    I therefore exchanged the debugging cord to the baseboard to a shorter one, and everything started to work!

    The old cord  was just about 15cm long and had a connecter in between that was mounted into the case of the housing.
    The new cord is approximatly 5cm long and doesn't have any other connecter in between.

    So to summarize:
    It is possible to use the XDS100 debugger that is built on the experimenters kit directly to debug another system, but only if the additional debugging cable isn't to long!
    If something doesn't fit, you won't get any error message, just some unexplainable behavior  of your hardware.
    However, I'm going to get real debugger.

    Regards Markus

  • 0 in reply to Markus Hofinger
    TI__Intellectual 1615 points

    A great solution for a low cost external USB JTAG emulator is the new XDS100 emulator TI has just released. It is $79 and functions the same as the onboard emulation on the docking station. You can order it and find more info here: http://focus.ti.com/docs/toolsw/folders/print/xds100.html

  • 0 in reply to Charlie3036
    Prodigy 90 points

    Just in order to make this post complete:

    Changing the crystal worked in my case, and I`ve found another reference on the TI E2E community about changing the crystal on the control card of the 28335, for those of you who are also in need of changing it.

    http://e2e.ti.com/forums/p/5893/22174.aspx