Post Go back to editing

hmc834 temperature problem?

We are using the HMC834 in a radar app as a second LO. The output frequency is set between 7.6 GHz and 8 GHz in pure integer mode using a 95 MHz reference frequency and a 19 MHz PD frequency. Loop filter bw: 130 kHz. From temperature chamber measurements the PLL goes out of lock under 8°C and seems to get back in lock under 0°C. It should work from -30°C to 75°C. There are many registers and some bitfields without explanation. We init the chip as:

      pll_write_datum(0,1);    
      pll_write_datum(0,0);
      pll_write_datum(1,2);
        pll_write_datum(2,5); 
        wait_on_autocal();
        pll_write_datum(6,0x2303F8);
        pll_write_datum(7,0x00274D);
        pll_write_datum(8,0xC1BEFF);
        pll_write_datum(9,0x003264);
        pll_write_datum(0x0A,0x002007);
        pll_write_datum(0x0B,0x07C061);
        pll_write_datum(0x0f,0x81);
        wait_on_autocal();
        pll_write_vco(1, 0x1F);
        pll_write_vco(2, 0x1C1);
        pll_write_vco(3, 0x0);
        pll_write_vco(4, 0x16D);
        pll_write_vco(5, 0x0FC);
        pll_write_vco(6, 0x0FF);
        
        pll_reset_vco();
        
        pll_write_datum(3,N);
        pll_write_datum(4,0);
        wait_on_autocal();

however I am not sure if autocal runs correctly and if this device goes out of lock by temperature change, and what to do then. What if the PLL goes out of LOCK? just rewrite the N register. (We do this now, but without success). What is the correct sequence for programming the PLL. Are the provided values OK? At the moment we can not find any cause why the PLL goes out of lock. Reference signal is 2Vpp clean sinus, clean spectrum. (All other PLLs run on this clock and are OK) Supply voltages are fine.What can be the cause?

PLL_8GPAGE1.pdf
Parents
  • The test with: Reg 0x06=0x30F4A has finished. Something changed: the temp range not working is now -6°C to + 6°C (there is an AD7415 temp-sensor on the board). All PLLs off except the 95 MHz crystal oscillator showed that it's not interference. At least I know now: It's not the power supply, it's not the reference signal, it's not interference. I put an absorber on the box lid, so its possibly not some strange waveguide mode. I changed the loop filter from 250 kHz to 130 kHz, so it's not the loop filter and it's not the layout, there are two gnd layers on this board and the top layer is just above the gnd layer. The loop filter feedback to vtone is between two ground layers, not signal lines crossing the HMC834s area or coming close on any layer. So I hope it is just software..... Next try with changed Reg 0x6 and 0xA while observing vtune....

Reply
  • The test with: Reg 0x06=0x30F4A has finished. Something changed: the temp range not working is now -6°C to + 6°C (there is an AD7415 temp-sensor on the board). All PLLs off except the 95 MHz crystal oscillator showed that it's not interference. At least I know now: It's not the power supply, it's not the reference signal, it's not interference. I put an absorber on the box lid, so its possibly not some strange waveguide mode. I changed the loop filter from 250 kHz to 130 kHz, so it's not the loop filter and it's not the layout, there are two gnd layers on this board and the top layer is just above the gnd layer. The loop filter feedback to vtone is between two ground layers, not signal lines crossing the HMC834s area or coming close on any layer. So I hope it is just software..... Next try with changed Reg 0x6 and 0xA while observing vtune....

Children
No Data