The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> phase noise increase due to FD-to-SE converter
https://designers-guide.org/forum/YaBB.pl?num=1276812346

Message started by OpAmp on Jun 17^th, 2010, 3:05pm

Title: phase noise increase due to FD-to-SE converter
Post by OpAmp on Jun 17^th, 2010, 3:05pm

Hi,

I have designed a fully differential (FD) ring oscillator. It has a fairly decent phase noise. I need to convert the FD output of the oscillator to singe-ended (SE). When I use a simple diff pair to do that the phase noise at high frequencies degrades considerably. In other words, the phase noise at the output of the oscillator (input of the FD-to-SE is still good), but the phase noise at the output of the FD-to-SE converter is not good at high frequencies.

I was wondering is there a good FD-to-SE converter which does not add much phase noise/jitter?

Thanks

Title: Re: phase noise increase due to FD-to-SE converter
Post by Mayank on Jun 17^th, 2010, 9:05pm

Quote:

When I use a simple diff pair to do that the phase noise at high frequencies degrades considerably.

Do not use Tail Current Source. Keep sizes large.

--M

Title: Re: phase noise increase due to FD-to-SE converter
Post by rfcooltools.com on Jun 18^th, 2010, 11:50am

OpAmp,

Use a totem pole circuit sometimes called push-pull output driver.

http://rfcooltools.com

Title: Re: phase noise increase due to FD-to-SE converter
Post by love_analog on Jul 29^th, 2010, 6:34am

Mayank, rfctools
Why does the OP's phase noise degrade. I would have thought since its outside the VCO loop, the noise is just going to raise the thermal noise floor rather than get integrated. I am curious to see if I am right.

BTW, how do your solutions help.

Title: Re: phase noise increase due to FD-to-SE converter
Post by Mayank on Jul 29^th, 2010, 9:15pm

Quote:

Why does the OP's phase noise degrade

Any extra circuitry you add is gonna inject its own noise. Hence, phase noise of the signal coming from VCO is bound to increase.

Quote:

I would have thought since its outside the VCO loop, the noise is just going to raise the thermal noise floor rather than get integrated.

The noise neednt be just thermal. The circuit's flicker noise will also gt injected into the system, which is not getting affected by preceding PLL's XF. Hence, unmodulated noise(both flicker n thermal) get added to the signal, thus degrading phase noise.

If by integration, you mean converting freq. signal to phase, then that you would have to do at FD-SE converter's output too.

Quote:

BTW, how do your solutions help.

Quote:

Do not use Tail Current Source. Keep sizes large.

No TCS means no flicker upconversion. Hence lower phase noise. Large sizes keep flicker from i/p pairs in control.

--M

Title: Re: phase noise increase due to FD-to-SE converter
Post by vp1953 on Jul 30^th, 2010, 9:53am

Hi Mayank,

>No TCS means no flicker upconversion. Hence lower phase noise.
>Large sizes keep flicker from i/p pairs in control.

Is the mechanism for flicker up conversion due to the inherent non linearity of the diff amp?

Title: Re: phase noise increase due to FD-to-SE converter
Post by Mayank on Aug 3^rd, 2010, 4:08am

Quote:

Is the mechanism for flicker up conversion due to the inherent non linearity of the diff amp?

No, the mechanism involved is the differential pulling of Tail current source node at twice the freq. of the applied diff. signal.

Non-linearity of diff. pair, imperfect differential signal, finite rout of TCS, CM fluctuations, gm mismatch b/w i/p pairs etc. contribute to the total noise because of upconversion from above mentioned phenomenon.

Title: Re: phase noise increase due to FD-to-SE converter
Post by sheldon on Aug 3^rd, 2010, 8:29pm

Mayank,

Unless OpAmp is doing Monte Carlo simulations including mismatch,
then the imperfections in the differential signal path will not be
manifested unless they are related to unmatched loading. Are there
other potential sources for flicker noise conversion?

Best Regards,

Sheldon

Title: Re: phase noise increase due to FD-to-SE converter
Post by Mayank on Aug 3^rd, 2010, 8:55pm

Yes there is, the prime source -- Flicker noise from Tail current source.
All the other ones i mentioned were to indicate that the secondary sources like diffamp non-linearity et al. that vp1953 was suggesting are contributors, not phenomenon which leads to upconversion.

Flicker noise of TCS upconverts increases phase noise.

Title: Re: phase noise increase due to FD-to-SE converter
Post by vp1953 on Aug 5^th, 2010, 11:26am

>No, the mechanism involved is the differential pulling of Tail current
>source node at twice the freq. of the applied diff. signal.

Hi Mayank,

Thanks for sharing your great insights.

Obviously if the input signals at the diff amp were small, then the drain of the tail current source is a virtual ground, there would not be any 2f component there.

For the case, that the input signals are large, assume that each FET of the diff acts like a switch. For this case, there would be a 2f component at the drain. But the 2f component is additive to the flicker noise and not multiplicative, so you would still need some nonlinearity for the flicker noise to upconvert to the 2f frequency.

At each single ended output of the diff amp, the flicker noise of the tail source is modulated by f (square wave with amplitude 1,0 assuming the FET of the input acts as a switch). The differential output has the flicker noise modulated by f (square wave with amplitude 1,-1). Short of any non ideality present, the upconversion of the flicker noise is to f, not to 2f.

I look forward to any comments you can provide.