LMK04208: Spurious on output with two ref clock on input

This is my configuration for the LMK04208:

R0 (INIT)    0x00160040
R0    0x00140080
R1    0x80140181
R2    0x80140182
R3    0x80140183
R4    0x00140184
R5    0x80140185
R6    0x01400006
R7    0x04400007
R8    0x04010008
R9    0x55555549
R10    0x9102410A
R11    0x0401100B
R12    0x1B0C006C
R13    0x2302806D
R14    0x0200000E
R15    0x8000800F
R16    0xC1550410
R24    0x00000058
R25    0x02C9C419
R26    0xAFA8001A
R27    0x10001E1B
R28    0x00201E1C
R29    0x0180019D
R30    0x0200019E
R31    0x003F001F

Can you send me a (.tcs) file instead please? 

Best regards, 

Vicente 

LMK04208.tcs

Hi Mathieu, 
I have some thoughts. 
Can you connect LMK40832 OSCout to CLKin1 instead? This should be a square wave. I would expect you still see this spur. 
I believe we are seeing mixing between the 122.88MHz square waves 6th harmonic and the 737.25MHz output. 
This would explain why you don't see it with a signal generator (Sine wave) vs buffer (square wave with harmonics). 
Do you see this on the LMK01801 output?

Best regards, 

Vicente 

Hi Vicente,

I can't connect easily the LMK04832 on CLKin1 easily because CLKin1 is a single ended input and LNK04832 is differential. And must solder some passives on eval board.

But whatever the clk input I use in my LMK04208 configuration, I have spurs on the output. Spurs amplitude and frequency change a little bit when I switch from one clock input to the second one.

How harmonics from the input clock that I don't use can be on the output of the LMK04208? Is it something you can reproduce on your side?

Mathieu, 
My bandwidth is very limited at the moment and I probably won't get a chance to test this till Friday. 
Understood. 
You mentioned you saw spur only when the clock buffer was driving 4208. 

Mathieu Buffard said:

ne thing I noticed: If I disconnect the LMK01801 and connect my signal generator (sine wave) directly to the LMK04208 input, spurs disappear.

This seemed to indicate to me that there is a possibility that the clock generator (Square wave) was having some of the harmonics mixing on the output side (737.25MHz). Especially since when you say you bypass the clock buffer and use a signal generator (Sine Wave) you no longer see the spurs. 

Are you able to see this spur on the output side of the buffer ?

Best regards, 

Vicente 

Hi Vicente,

I looked at the output of the LMK01801 with the spectrum analyzer, and I don't have spurious around my center frequency. I have harmonics at multiple of 122.88MHz, when that's expected...

I am still trying to understand why, with two square waveforms as clock input I have spurs at the output, but with one square and one sine I don't see any spurs...

Hi Mathieu, 
That sounds like what I expected for the output of LMK01801. Due to it being a square wave output I would expect just harmonics of the carrier frequency (122.88MHz in this case). 

One more thing I am confused on - the output frequency should be 737.28MHz if the output clock divider is set to '4' if the VCO frequency is 2949.12MHz. 
Something is kind of off here. 

Are you sure the device is locked? 

Best regards, 

Vicente 

Hi Vicente,

Yes I have 737.28MHz at the output of the LMK04208. The screenshot is on my first message. 122.88MHz is the output of the LMK01801.

Gotcha. 
Okay I was confused since you had 737.25MHz on the block diagram and the spectrum analyzer also measured that as the carrier frequency. 

Hi Vicente,

Were you able to reproduce the issue?

Thanks

Hi Mathieu, 

Vicente is out of office today, please expect a response tomorrow. 

Regards, 

Connor 

Mathieu, 
I have ordered samples to try and test this on a socketed board. 
I will follow up once samples arrive. 

Best regards, 

Vicente