2
\$\begingroup\$

The basic addition of a 20MHz external resonator is becoming much more painful that I thought. I've got a breadboard setup with a DIP package of a PIC24HJ128GP502. Nothing fancy, and it has been able to blink an LED with the internal FRC for the clock. My configuration bits are set in code as follows:

// General Segment code-protection configuration: // leave code protection off, GSS_OFF // disable code protection, GCP_OFF // disable write protection, GWRP_OFF _FGS(GSS_OFF & GCP_OFF & GWRP_OFF) // OSCillator SELection: // Primary oscillator (XT, HS, EC) w/ PLL FNOSC_PRIPLL // enable two-speed oscillator startup IESO_ON _FOSCSEL(FNOSC_PRIPLL & IESO_ON) // OSCillator configuration: // enable clock switching but not monitoring, FCKSM_CSECMD // allow the RPn pins to be continually remapped, IOL1WAY_OFF // leave OSC2 pin as clock pin (not digital I/O) OSCIOFNC_OFF // High-Speed (10-32MHz crystal) Oscillator, POSCMD_HS _FOSC(FCKSM_CSECMD & IOL1WAY_OFF & OSCIOFNC_OFF & POSCMD_HS) // WatchDog Timer configuration: // disable the watchdog timer, FWDTEN_OFF _FWDT(FWDTEN_OFF) // Power-On-Reset configuration: // map I2C pins to SDA1/SCL1, ALTI2C_OFF // use the smallest power-on-reset value of 2ms, FPWRT_PWR2 _FPOR(ALTI2C_OFF & FPWRT_PWR2) // Debugger configuration: // JTAG is enabled, JTAGEN_ON // communicate on PGC1/EMUC1 and PGD1/EMUD1, ICS_PGD1 _FICD(JTAGEN_ON & ICS_PGD1) int main(void) { // STAYS HERE FOREVER, NEVER LOCKS! while (!OSCCONbits.LOCK); // wait for the PLL to lock ...

The only knobs (or places I can go wrong that I know of) are in _FOSCSEL() and _FOSC(). I've tried nearly every iteration of those configuration bits that made any bit of sense (even just a wee bit..ok even the ones that made no sense). I've also tried programming it in debug mode as well as release mode.

Environment: MPLab v8.88.00.00, C30 compiler

Relevant Research: - section 39.13 on two-speed startup: http://ww1.microchip.com/downloads/en/DeviceDoc/70308B.pdf - Section 2.7 of the datasheet - the header file in general, p24HJ128GP502.h, to find the correct macros and configuration masks.

Circuit: 3.3V power supply the middle pin of the resonator goes to GND, the outer two pins go to OSC1 and OSC2, there is a 1MOhm resistor connecting each of the outer two pins for the 3-pin resonator (for stability) resonator: ZTT 20.00MX http://www.ecsxtal.com/store/pdf/zttr.pdf http://www.digikey.com/scripts/DkSearch/dksus.dll?lang=en&keywords=ztt-20.00mx&WT.term=ztt-20.00mx&WT.mc_id=Crystals%20and%20Oscillators&WT.medium=cpc&WT.campaign=Crystals%20and%20Oscillators&WT.content=text&WT.srch=1&WT.source=google

\$\endgroup\$
9
  • \$\begingroup\$ I've wasted a week searching for a problem with an MSP430. In the end the oscillator just wasn't working. Can you check that with an oscilloscope? \$\endgroup\$ Commented Feb 5, 2013 at 8:31
  • \$\begingroup\$ Do you intend to use PLL with your external XTAL? Have you tried without the PLL? \$\endgroup\$ Commented Feb 5, 2013 at 8:52
  • 1
    \$\begingroup\$ A three pin crystal? Are you sure? Post the schematic. \$\endgroup\$ Commented Feb 5, 2013 at 10:22
  • \$\begingroup\$ Leon is onto something, it's probably either a resonator or oscillator module both of which need connecting a different way so make sure you include the part number as well. \$\endgroup\$ Commented Feb 5, 2013 at 11:14
  • \$\begingroup\$ Please post the part number and a datasheet link to your oscillator. We cannot definitely solve anything with incomplete information. \$\endgroup\$ Commented Feb 5, 2013 at 13:52

1 Answer 1

Reset to default
1
\$\begingroup\$

Apart from the crystal sounding like it's actually a resonator (part number?), you also need to unlock the OSCCON register to switch clock sources. The compiler provides built in functions to do this, as described in the datasheet.

Here is an example from the datasheet of the code to switch over to Primary Oscillator with PLL (you need to check/set the divider/multiplier bits according to your frequency - this is for 20Mhz crystal and 80Mhz operating freq):

int main() { // Configure PLL prescaler, PLL postscaler, PLL divisor PLLFBD=30; // M = 32 CLKDIVbits.PLLPOST=0; // N1 = 2 CLKDIVbits.PLLPRE=2; // N2 = 4 // Initiate Clock Switch to Primary Oscillator with PLL (NOSC = 0b011) __builtin_write_OSCCONH(0x03); __builtin_write_OSCCONL(0x01); // Wait for Clock switch to occur while (OSCCONbits.COSC != 0b011); // Wait for PLL to lock while(OSCCONbits.LOCK!=1) {}; }
\$\endgroup\$
4
  • \$\begingroup\$ yes, i saw this code in an example as well. i am trying to use the automatic two-speed oscillator startup though which shouldn't require an unlocking step. i have just tested putting the unlock macro as the first line in my main though with no result: __builtin_write_OSCCONL(OSCCON & 0xBF); \$\endgroup\$ Commented Feb 5, 2013 at 19:37
  • \$\begingroup\$ I recall having similar issues with a dsPIC a while back. I would try the full switching code shown anyway, both with the auto two speed startup on and off. If it still doesn't work I might try it here (I think I have a PIC24 lying around somewhere) Your resonator speed seems about right if that's half the period you are measuring - 1/50ns = 20MHz. However it still may not be locking if it's noisy/jittery. A scope shot or two might be useful, both short time scale and long. \$\endgroup\$ Commented Feb 5, 2013 at 21:01
  • \$\begingroup\$ I am continuing to experiment but finally have something working. It turns out my V_cap capacitor was not large enough (I had 0.1uF ceramic cap instead of 10uF). I also didn't realize the internal logic level was 2.5V for this pic and not 3.3V. Thank you, Oli \$\endgroup\$ Commented Feb 9, 2013 at 21:54
  • \$\begingroup\$ @tarabyte - ah-ha I see, that would cause issues. Glad to hear you got it working. \$\endgroup\$ Commented Feb 9, 2013 at 22:17

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.