We are trying to debug a temperature problem in our product which uses and LMK03033C. We ordered the evaluation board for this part, and set it up the same way as our product. It has the input reference at 50 MHz, PD running at 2MHz, and the PLL running at 2160 MHz. I am using the same code that our product does, but the EVM will not lock. Here are the register settings:
Current constants for LMK03033C
-------------------------------
Divider #: 0 1 2 3 4 5 6 7
MUX: 1 1 1 1 1 0 0 0
EN: 1 0 0 0 0 0 0 0
DIV: 40 40 2 4 1 1 1 1
DLY: 0 0 0 0 0 0 0 0
DIV4: 0 VCO_C3_C4: 10 VCO_R3_LF: 0 VCO_R4_LF: 0
OSCin_FREQ: 50 PLL_R: 25 PLL_N: 216 VCO_DIV: 5
PLL_MUX: 3 EN_Fout: 1 POWERDOWN: 0 PLL_CP_GAIN: 3
GOE: 1 LD: 0 SYNC: 1
I can cycle the LD pin from the R divider, PD or N divider. The R divider output is exactly as expected, a 1MHz square wave. The LD pin is low. The N divider output is either constant high, constant low, or occasionally 115.7kHz. Looking at Fout, it is either not there or it is somewhere between around 1.6GHz and 2.6GHz, but not locked. The CP_Gain does not make any difference. Sometimes cycling the mux for the LD pin causes Fout to disappear. I have tried other settings for PLL_N, but the results are the same. The C3_C4 setting is as on P 33 of the EVM manual.
Again, this is a standard evaluation board, the only change being to move R7 to R11 to allow an external OSC_IN. The input level is +3dBm, and the output of the R divider is exactly correct.
I have looked through the data sheet, but nothing jumps out at me. The registers are being programmed in the order recommended in the data sheet.
The VCO can clearly cover my frequency, and R divider is working. It seems like the N divider is not working for some reason- and of course then the PLL can't lock. But these same settings in our product give good results.
Any ideas on what can cause this behavior would be greatly appreciated.
