13.6 GHz output signal of ADF5356 synthesizer chip

[Gerhard Hoffmann]

13.6 GHz output signal of ADF5356 synthesizer chip. The ADF5356 has an
on-chip VCO from 3-6.8 GHz and outputs twice its frequency on a pin.

Looks like the doubler has a serious symmetry problem. The 2*f
output is the strongest component, but just barely.
And the level is also on the weak side.

That requires some serious microstrip filters. :-(


<
https://www.flickr.com/photos/137684711@N07/50403778976/in/album-72157662535945536/
>

<
https://www.flickr.com/photos/137684711@N07/50403077953/in/album-72157662535945536/
>


cheers, Gerhard

[jla...@highlandsniptechnology.com]

On Thu, 1 Oct 2020 09:56:23 +0200, Gerhard Hoffmann <dk...@arcor.de>
wrote:

Sinewave purity isn't the norm at high frequencies. I've seen fancy HP
rf sig gens with visible distortion too. They tend to proudly display
low close-in spurs and ignore the stuff an octave up, or down in your
case.

Can you use a coaxial ceramic resonator to clean it up?





--

John Larkin Highland Technology, Inc

Science teaches us to doubt.

Claude Bernard

[Gerhard Hoffmann]

Am 01.10.20 um 17:21 schrieb jla...@highlandsniptechnology.com:

> On Thu, 1 Oct 2020 09:56:23 +0200, Gerhard Hoffmann <dk...@arcor.de>
> wrote:
>
>>
>>
>> 13.6 GHz output signal of ADF5356 synthesizer chip. The ADF5356 has an
>> on-chip VCO from 3-6.8 GHz and outputs twice its frequency on a pin.
>>
>> Looks like the doubler has a serious symmetry problem. The 2*f
>> output is the strongest component, but just barely.
>> And the level is also on the weak side.
>
>

> Sinewave purity isn't the norm at high frequencies. I've seen fancy HP
> rf sig gens with visible distortion too. They tend to proudly display
> low close-in spurs and ignore the stuff an octave up, or down in your
> case.
>
> Can you use a coaxial ceramic resonator to clean it up?

That would be a dielectric resonator on 10 GHz.
I have one in different oscillator in this box.
I'm still struggling to get valid s-parameters from
EM-simulations of the puck.

The synthesizer is there to provide a tuneable beat frequency.
I've made a ring resonator to notch out the fundamental.

I've made a ring resonator that makes a nice trap for the fundamental.
This one is scaled down to 3.5/7GHz since I'm VNA challenged.

<
https://www.flickr.com/photos/137684711@N07/50404949846/in/album-72157662535945536/
>

<
https://www.flickr.com/photos/137684711@N07/50404249018/in/album-72157662535945536/
>

Board material is Rogers TMM6, 25mil. It has the consistency of
cold wax, or of an eraser. Breaks when one looks hard at it.

The filter is self-etched. Laser-printed to foil and exposure
on light-sensitive laquer. I have peeled away some copper
on the top side so see if that has an effect. It had none.

Cheers, Gerhard

[John Larkin]

On Thu, 1 Oct 2020 18:40:15 +0200, Gerhard Hoffmann <dk...@arcor.de>

Looks like the surface-mount CCRs top out at about 10 GHz.

https://www.imcsd.com/news/ceramic-coaxial-resonators

[Simon S Aysdie]

Yep--and that is really pushing it. The upper limit used to be about 6 GHz not too long ago. This filter should be printed (planer). Gerhard already did the fundamental trap as printed.

You can see elsewhere that IMC has stopped selling individual resonators (https://www.imcsd.com/products/resonators). IMC has traditionally had one of the best (highest Q) resonator plating processes. ComNav (Marty Geesaman, https://comnav-eng.com/index.htm) is the other. IMC's founder, Dennis R. Clark, passed away in 2015, and his wife now runs the company.


CTS does monoblock BPF, but they stop sub 8 GHz. CTS is for higher volume. I don't think muRata is even bothering with mono-block anymore. NTK used to have both monoblock and LTCC stuff, but seem to have dropped out of that business altogether.