How to reduce the spur of OCXO and SRD Multiplier circuit?

I tested OCXO 128.3333M with +12V power, the power cable length is 1.5m. I soldered a 100uF Tan Cap to 12V trace of OCXO power. And I tested the spur at 140K is -84dBc. The seller test result is almost same as me. The seller tested with a special fixture.
My question is:
1. How to reduce the spur? Any other ideas?
2. Any special power supply circuits need for the OCXO to reject any spur?

The following is the output spectrum from the multiplier circuit of the OCXO 128.3333M. The spur is amplified to -62dBc.

My target is -80dBc, what can I do except asking the vendor to reject the spur to -100dbc?

I notice the RF cable circle affect the spur, so I prevent the cable vibrate.

When I use Agilent signal generator to replace the OCXO, the spur diminished. Can I say the spur only caused by OCXO spur, not my multiplier circuit?

hi there,

just wondering what waveform is coming out of your OXCO ?

if its a square wave that would answer a lot of your probs as it would be rich in
harmonics and would result in what you are seeing on the spec an.
Many of the OXCO's I have had anything to do with are all squarewave rather than
sinusoidal.

Have a look at the output of the OXCO on an oscilloscope it will tell all.

As you have seen in your experiment using a clean signal from the signal gene
you didnt have the problem

just a thought :)

Dave

You need to find what is the source of the 140KHz,
It can be your power supply or it can be a really bad OCXO internal regulator design.

Take the OCXO out and test it alone in a fixture, use a linear power supply with a high value electrolytic (>1000uF) close to the fixture.

YY

Hi, Dave,

the OCXO output is sinewave, level +7dBm min, Load=50R

The seller tested the OCXO in the fixture and found the 140KHz spur, and the spur level is about -84dBc, almost same as my test result.

And my multiplier module amplify the spur from -86dBc to -62dBc. Is it because in my module, I used several amplifier in the saturated region? The signal has low gain, but the spur has higher gain, so the spur level is raised. Or is it because something like mixer?

You are multiplying 128.3333MHz x 16 so the theoretical phase noise degradation is 20 x log(16)=24 dB.

You measured the spur before the multiplication = -86dBc
You measured the spur after the multiplication = -62dBc

The difference is 24dB (What a coincidence!) just prove the theory!

I am afraid that the solution is a better OCXO, a decent OCXO should be clean.

My guess is the OCXO vendor in order to obtain the 128.333MHz is using an internal PLL loop multiplication with a comparison frequency set at ~140KHz, that is the spur that you see.
Ask the vendor if this is correct.(let me know I'm curious)
There is also a possibility that the internal OCXO Voltage regulator is producing noise at ~140KHz
YY

Hi, yygomez,
The vendor don't use PLL loop. Because the vendor don't have OCXO sample now, so they can't give a conclusion about the spur.
And the seller in china tested the OCXO again, and they said there are no spur above -94dBc, because their spectrum analysis noise floor is -94dBc.
The spur caused maybe by my power supply, I will test the OCXO again in a few days.
Thank you.

Hi,

The spur diminished greatly when I tested again.
The OCXO spur below -94dBc. And the multiplier output spur is also below -94dBc. The spur was not amplified after X16 multiplier.

One possible explanation is test environment changed, I tested in another room. So maybe some interference changed.

Another possibility is that grounded changed. First time I used a PCB board maybe not good grounded. The DC ground and RF ground seperated about 5cm.
The second time I connect directly on the OCXO PIN. The DC ground and the RF ground link to one point.

Thanks every one who had helped me.