The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> low noise regulator for rf vco
https://designers-guide.org/forum/YaBB.pl?num=1277523905

Message started by mosman on Jun 25^th, 2010, 8:45pm

Title: low noise regulator for rf vco
Post by mosman on Jun 25^th, 2010, 8:45pm

usually most vco are powered by a analog linear regulator.
can vco be supplied by one digital linear regulator? it means the regulator is still linear regulator not switch-type, DCDC. but because it is digital, its output voltage will changed step by step, maybe the step is 10mv, 20mv,... . i dont know if this style was used in vco in any real product. will this worsen the important noise performance? anybody has such experience?

Title: Re: low noise regulator for rf vco
Post by pancho_hideboo on Jun 25^th, 2010, 10:10pm

mosman wrote on Jun 25^th, 2010, 8:45pm:

it means the regulator is still linear regulator not switch-type, DCDC.
but because it is digital, its output voltage will changed step by step, maybe the step is 10mv, 20mv,... .
i dont know if this style was used in vco in any real product.
will this worsen the important noise performance?

No, it does not make noise performance worse as far as it is static change.

Actually it is fairly common to adjust voltage of regulator at room temperature by such digital step adjustment.

In some cases, we aggressively increase output voltage of regulator at high temperature by such digital change.
Vice versa, we aggressively decrease output voltage of regulator at low temperature by such digital change.

Title: Re: low noise regulator for rf vco
Post by mosman on Jun 26^th, 2010, 2:03am

pancho_hideboo wrote on Jun 25^th, 2010, 10:10pm:

i also see your reply in edaborad. thanks.
yes, i am facing such issue. the regulator will decrease its output at high temperature. the noise will be introduced if analog control loop is used. so the digital method is used. to adjust output voltage this digital loop will run once chip work, not only at chip startup. when the digital loop sends bits to adjust supply, the supply of every block will change about, say step=50mv. my application engineer worry about this 50mv step, i also has no experience on this. they also do. you say "it is static change", but at adjusting moment, the supply will be chnged one step, 50mv. i dont very understand this impact on vco's phase noise.

Title: Re: low noise regulator for rf vco
Post by pancho_hideboo on Jun 26^th, 2010, 2:08am

mosman wrote on Jun 26^th, 2010, 2:03am:

the noise will be introduced if analog control loop is used.

Why does it introduce noise ?

mosman wrote on Jun 26^th, 2010, 2:03am:

you say "it is static change", but at adjusting moment, the supply will be chnged one step, 50mv.
i dont very understand this impact on vco's phase noise.

It affects spurious emission rather than phase noise.
It might be unlocked state during a very very very small time, although I don't think it will be unlocked state by such small voltage change.

Anyway it can make purity of PLL worse during very small time.

Is VCO is always active in your system ?

I include voltage change sequence during channel change in PLL synthesizer.
Here I also include VCO frequency calibration sequence in every channel change.

Whether these are possible or not is dependent on required channel change time.

Title: Re: low noise regulator for rf vco
Post by Mayank on Jun 27^th, 2010, 9:14pm

Quote:

the noise will be introduced if analog control loop is used.

Any Analog Circuit will introduce noise.
I think what Pancho was trying to ask, & me too, is that why will the Digitally Controlled Supply NOT introduce any noise ??

Digitally Controlled Supply

What do you mean by Digital Control ?

1 -- You are calibrating the Reference Voltage of the Regulator Loop digitally ??
Without an Analog Loop, How will you remove Power Supply Noise & High & Low Frequency ripples ??

Quote:

you say "it is static change", but at adjusting moment, the supply will be chnged one step, 50mv.
i dont very understand this impact on vco's phase noise.

Pancho is right, i guess such small change of supply would instantaneously but small frequency shift in the VCO o/p Freq. depending on your Kpushing Factor, causing PLL to be in an unlocked state for a v small time.

yes, VCO output, will show spurious components every time you change your calibration code. If your VCO response to the Supply variation is Overdamped, they will die out soon enough.

--
Mayank.

Title: Re: low noise regulator for rf vco
Post by pancho_hideboo on Jun 28^th, 2010, 7:07am

Mayank wrote on Jun 27^th, 2010, 9:14pm:

I think what Pancho was trying to ask, & me too, is that why will the Digitally Controlled Supply NOT introduce any noise ??

I don't think so.
Simple digitally controled voltage by temperature never introduce any additional noise as far as its change is static.
Here this control is done without any feedback loop.

Mayank wrote on Jun 27^th, 2010, 9:14pm:

What do you mean by Digital Control ?

I think he assumes digital adjustment of feed back resistor ratio in linear regulator.

Title: Re: low noise regulator for rf vco
Post by rfcooltools.com on Jun 28^th, 2010, 11:55am

Linear regulator considerations for a VCO will include:
1. device type
2. size of device
3. vds
4. bandwidth and gain of feedback loop
5. vco symetry (balanced plus and minus swing)
6. Noise of regulators reference voltage

http://rfcooltools.com

Title: Re: low noise regulator for rf vco
Post by Mayank on Jun 29^th, 2010, 1:37am

Quote:

I don't think so.
Simple digitally controled voltage by temperature never introduce any additional noise as far as its change is static.
Here this control is done without any feedback loop.

Again my question,

Quote:

Without an Analog Loop, How will you remove Power Supply Noise & High & Low Frequency ripples ??

Quote:

Simple digitally controled voltage by temperature never introduce any additional noise as far as its change is static.
Here this control is done without any feedback loop.

Quote:

I think he assumes digital adjustment of feed back resistor ratio in linear regulator.

Seems a bit contradicting. I am confused.

I assume Linear Regulator you are talking about is the one with an Analog Loop as rfcooltool points,
Linear regulator considerations for a VCO will include:
1. device type
2. size of device
3. vds
4. bandwidth and gain of feedback loop
5. vco symetry (balanced plus and minus swing)
6. Noise of regulators reference voltage

Title: Re: low noise regulator for rf vco
Post by pancho_hideboo on Jun 29^th, 2010, 3:46am

Mayank wrote on Jun 29^th, 2010, 1:37am:

pancho_hideboo wrote on Jun 28^th, 2010, 7:07am:

I think he assumes digital adjustment of feed back resistor ratio in linear regulator.

Seems a bit contradicting. I am confused.
I assume Linear Regulator you are talking about is the one with an Analog Loop as rfcooltool points,

Right.

But from practical point of view, there is no remarkable affects for PLL characteristics at all, as far as voltage change is small and static.

Mayank wrote on Jun 29^th, 2010, 1:37am:

Linear regulator considerations for a VCO will include:
4. bandwidth and gain of feedback loop

Actually I change supply voltage from 1450mV to 1650mV by digital adjustment of feed back resistor ratio in linear regulator.
For such small voltage change, there is no remarkable change of bandwidth and gain of feedback loop.
And I can't observe any degradation of PLL characteristics as far as voltage change is static.

Title: Re: low noise regulator for rf vco
Post by Mayank on Jun 29^th, 2010, 5:15am

Quote:

Right.

But from practical point of view, there is no remarkable affects for PLL characteristics at all, as far as voltage change is small and static.

Agreed. But then all you are doing is using an digital Calibration of FB resistor over an Analog Loop.

MOSMAN initially quoted :

Quote:

the noise will be introduced if analog control loop is used. so the digital method is used.

That Analog Loop is very much there.
All you are doing is Calibrating your output voltage, but the loop which regulates your FB voltage is unchanged. Hence, the noise introduced in unchanged.
So, the noise from the Analog Loop is very much there, since the Loop is present.

As far as question of Digital Stepping is concerned,

Quote:

As we both said, it's not gonna unlock the PLL.
You are simply over/under driving the VCO, or in other words, varying oscillator VDD, so as to vary the output Frequency by a factor of Kpushing.

VCO can also be calibrated in other ways, like bands in I_tail_current, Cap load on osc nodes, etc. etc.
you are using VDD stepping, which is perfectly fine, is one of them.

Hence, MOSMAN,
you can use digitally trimmed voltage supply from an Analog Linear Regulator. BUT the noise from the Analog Loop will be present as the Loop itself is there.

--
Mayank.

Title: Re: low noise regulator for rf vco
Post by pancho_hideboo on Jun 29^th, 2010, 5:29am

Mayank wrote on Jun 29^th, 2010, 5:15am:

So, the noise from the Analog Loop is very much there, since the Loop is present.

I can't understand what you want to say.

Again there is no remarkable affects for both loop frequency characteristics and also PLL characteristics, even if I change feedback resistor ratio in linear regulator.

Mayank wrote on Jun 29^th, 2010, 5:15am:

VCO can also be calibrated in other ways, like bands in I_tail_current, Cap load on osc nodes, etc. etc.
you are using VDD stepping, which is perfectly fine, is one of them.

Maybe you are misunderstanding.
Purpose of increasing vdd at high temperature is keeping output amplitude of VCO from being small not calibrating frequency of VCO or Kvco.

Mayank wrote on Jun 29^th, 2010, 5:15am:

Hence, MOSMAN,
you can use digitally trimmed voltage supply from an Analog Linear Regulator. BUT the noise from the Analog Loop will be present as the Loop itself is there.

From practical point of view, No.

Title: Re: low noise regulator for rf vco
Post by Mayank on Jun 29^th, 2010, 8:52pm

Quote:

Again there is no remarkable affects for both loop frequency characteristics and also PLL characteristics, even if I change feedback resistor ratio in linear regulator.

Digital Calibration would not affect PLL characterstics. Agreed.
But mosman was saying,

Code:

the noise will be introduced if analog control loop is used

To that, Analog Loop is still present, right. You are changing FB resistor ratio, so as to step up/dn output voltage. Hence, there is no removal of noise in using digital calibration just like there is no excess noise introduction. The noise i am talking about is the one from the opamp, The PMOS DRIVER & the resistors shaped by the Analog Loop.

Quote:

Purpose of increasing vdd at high temperature is keeping output amplitude of VCO from being small not calibrating frequency of VCO or Kvco.

Got it. It can also be employed to vary VCO frequency bands.

Title: Re: low noise regulator for rf vco
Post by pancho_hideboo on Jun 30^th, 2010, 4:53am

Mayank wrote on Jun 29^th, 2010, 8:52pm:

The noise i am talking about is the one from the opamp,
The PMOS DRIVER & the resistors shaped by the Analog Loop.

Contributions of such noises to PLL are almost unchanged, even if I change feedback resistor ratio in linear regulator.

Mayank wrote on Jun 29^th, 2010, 8:52pm:

pancho_hideboo wrote on Jun 29^th, 2010, 5:29am:

Purpose of increasing vdd at high temperature is keeping output amplitude of VCO from being small not calibrating frequency of VCO or Kvco.

Got it. It can also be employed to vary VCO frequency bands.

Again read the following.

pancho_hideboo wrote on Jun 26^th, 2010, 2:08am:

I include voltage change sequence during channel change in PLL synthesizer.
Here I also include VCO frequency calibration sequence in every channel change.

Title: Re: low noise regulator for rf vco
Post by Mayank on Jun 30^th, 2010, 5:10am

I guess we both are stressing the same points.

Quote:

Contributions of such noises to PLL are almost unchanged, even if I change feedback resistor ratio in linear regulator.

Thats what i want to convey. The noise from Analog Loop is present & FB resistor ratio doesnt change this noise much.

Title: Re: low noise regulator for rf vco
Post by aLittleKnowledge on Jul 13^th, 2010, 4:38am

A digitally adjusted linear regulator is just the same as a linear regulator where the selection components are adjusted digitally. If you can leave the regulator setting unchanged during the times that you are interested in the noise levels there should be no noise introduced due to the fact that the regulator is digital. If this is possible it will avoid some design complexities such as resonances in the regulator control loop.
If the digital regulator output needs to be adjusted while you are interested in the signal frequency, there will be a quantifiable jump in the output frequency. Acceptability is a matter for calculation.

Title: Re: low noise regulator for rf vco
Post by pancho_hideboo on Jul 18^th, 2010, 10:43am

aLittleKnowledge wrote on Jul 13^th, 2010, 4:38am:

Acceptability is a matter for calculation.

What do you want to mean ?

Whether a discrete time change in synthesizer is acceptable or not is dependent on system requirements such as required channel change time.

See http://www.designers-guide.org/Forum/YaBB.pl?num=1277523905/3#3

Actually DTCXO(Digitally controlled TCXO) is extensively used in cellular phone system.