LMX2615EVM-CVAL: Spurious Observed in Spectrum output

Deepak Attri

Part Number: LMX2615EVM-CVAL

Tool/software:

I using 10MHz input reference for LMX2615EVM-CVAL board. I am generated output frequency around 6.84 GHz and observed the output spectrum. While generating integer output (6.840 GHz), I observed band of spurious about 3.7KH ( and it's multiples) away from center frequency. Same were not seen in case of fractional output (6.8346 GHz), and small spurious was observed around 840 Hz.

What could be reason for spurious observed in case of integer output? Why same are not seen in case of fractional output? What are remedies available (in case of LMX2615) in order to obtain clean spectrum?

I am attaching above spectrum output for your reference.

Kindly note, 10 MHz input reference itself had a clean spectrum and no such spurious were observed.

Regards

Deepak

1 month ago

0 Noel Fung 1 month ago

TI__Guru* 94680 points

Hi Deepak,

Could you provide the TICS Pro configuration?

Is the 10MHz clock sine wave?

0 Deepak Attri 1 month ago in reply to Noel Fung

Prodigy 30 points

Hi Noel,

PFA TICS Pro configurations.

Yes, 10 MHz input clock is sine wave.

0 Noel Fung 1 month ago in reply to Deepak Attri

TI__Guru* 94680 points

Hi Deepak,

10MHz sine wave is not a good clock source, however it should not create the problem you have.

I need to check this out in the lab.

0 Deepak Attri 1 month ago in reply to Noel Fung

Prodigy 30 points

Hi Noel,

Thanks. Do let me know your observations and what are remedies possible to mitigate it at our end.

Also, how similar/ different output spectrum will be (in general) in case of integer and fraction output? Does it make a difference if we have a simple fraction or a large fraction?

0 Noel Fung 1 month ago in reply to Deepak Attri

TI__Guru* 94680 points

Hi Deepak,

I don't see the problem in the lab, using your configuration. In fact, with integer channel, usually the spectrum is clean.

integer channel:

Fraction channel:

With 10MHz reference clock, I suggest use CMOS clock as the slew rate will be much higher, this help improve PLL in-band noise significantly. If CMOS clock is not available, use a buffer to turn 10MHz sine wave to square wave differential clock. Detail in below link.

https://www.ti.com/lit/pdf/snaa411

In addition, with 10MHz input clock, do not use the OSCin doubler. The doubler will create big spurs at 10MHz offset.

Clock & timing

Clock & timing forum

LMX2615EVM-CVAL: Spurious Observed in Spectrum output