Designer’s Guide Community

mtp3

Junior Member

Offline

Posts: 20

Amplifier design for a 2GHz signal
Jun 23^rd, 2013, 9:06am

Hi!

I have just finished designing a 2GHz PLL using a STM process and I hope to tapeout design soon. Anyways the problem I am facing is a bit ambiguous for me. Looking at the published papers from IEEE it is seen that almost all the PLL papers have a picture of VCO phase noise.

My question, How is this phase noise measured either by driving the VCO output off-chip and then measuring or some other technique. If the VCO output is pushed off chip then what kind of amplifier should be used so as not to worsens the phase noise during this process. The VCO output itself is a 2GHz sine-signal with a 700mV p-p amplitude.

Thanks in advance

Back to top

IP Logged

raja.cedt

Senior Fellow

Offline

Posts: 1516
Germany

Re: Amplifier design for a 2GHz signal
Reply #1 - Jun 23^rd, 2013, 10:07am

Hello,
Several ways....
1. As you said drive VCO through a buffer and measure (extra power due to high speed buffer design).

2. After some frequency division, which is realistic in any PLL(here you have to correct the phase noise by 20*log10(N)).

3. Directly measuring PLL phase noise and get some information about VCO phase noise beyond loop BW .

Thanks,
Raj.

Back to top

IP Logged

mtp3

Junior Member

Offline

Posts: 20

Re: Amplifier design for a 2GHz signal
Reply #2 - Jun 24^th, 2013, 6:12am

Many thanks for the reply Raj, I get what you are saying but have some reservations

1) Driving the output via buffers would degrade the phase noise due to the power supply variation of the buffers (because of driving a high freq signal).

2)Dividing by 2 improves the phase by 6dbc but I have not seen any papers where this is done. Usually in papers one sees a picture of a spectrum analyzer showing the output frequency and phase noise at some offset.

3)How can one directly measure the phase noise, could you explain a bit more.

Regards,
MTP3

Back to top

IP Logged

raja.cedt

Senior Fellow

Offline

Posts: 1516
Germany

Re: Amplifier design for a 2GHz signal
Reply #3 - Jun 24^th, 2013, 6:28am

1. Power supply variation will be there, but you can minimize through decaps and all. Moreover most of the supply noise will be like spur rather random so you can ignore.
2.Here i am not talking about improvement, it simplifies the a measurement. recently razavi made 300GHZ vco, he measured through dividers and many people do this. read some more papers...
3.didn't get what you are asking...

Thanks,
Raj.

Back to top

IP Logged