LMX2595: Frequency unstable and jumping

Hi David,

9GHz is an integer channel, the 1kHz frequency jump is not due to the PLL.

1KHz is a very small frequency compared with the VCO frequency, so it is also not likely a VCO locking issue.

This problem is most likely due to the reference clock or your PLL configuration.

Could you provide the PLL configuration in TICS Pro format? For example:

Is the reference clock single-ended or differential? How did you connect it to LMX2595?

Hi Noel

Thank you for looking at this, I'm asking the FPGA designer to provide the PLL configuration in TICS Pro format. 

The external reference clock has a single-ended output which feeds into the OSCINp pin. The OSCINn pin is connected to a series 0.1 uF capacitor and 50 ohm resistor to ground. 

The the LMX RF output passes thru a 4 dB attenuator that drives the input of a HMC8413 amplifier whose output goes to the input a a SA. SA bandwidth was set to 500 Hz with a frequency span of 10 kHz.  The level of the SA was adjusted to prevent overload allowing the SA to provide both amplitude and frequency measurements.  I don't recall the sweep speed but is was fast enough to view the frequency jump.  Frequency jump was about ~1 kHz lower than the 9 GHz signal and then quickly return the 9 GHz all within the millisecond range.  Measurements of this jump are difficult to capture.

Dave

HI David,

Ok, let's check the configuration once it is available. 

Hi David,

Thanks for the info, Noel will get back to you when he is available. If you happen to have the TICS Pro configuration file (.tcs) available, feel free to provide it.

Best,

Evan Su

Solution4.tcs

Hi David,

The configuration looks good, so the only reason that will make the VCO jumping is the reference clock is jumping.

do you have test equipment able to measure frequency vs time?

if you have a frequency counter, you may use it to measure the reference clock, see if it is indeed jumping.