Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Phase noise in PLL- Active vs Passive Filter

Status
Not open for further replies.
K

kommanche

Member level 4
Member level 4
Joined
Mar 18, 2007
Messages
70
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
1,766
Hi,

I am designing a PLL using ADF4157. The VCO I am using requires around 4.5 V tuning voltage at my operating frequency. So I decided to use a passive 10 kHz loop filter. Measured control voltage was around 4.7 V and In-band phase noise was around -70 dBc (@1 kHz, @10kHz). When I switched to active filter I got around 4.3V tuning voltage and -80 dBc in-band phase noise.

Actually I was expecting better phase noise performance in passive filter case but that is what I got. Also measured tuning voltage was changed strangely.

Do you have any idea about the reason or any suggestion to improve in band phase noise for passive filter case?

Thanks,
 

Did you have large resistor values = high thermal noise in the passive loop filter?
 

You have the same Loop Bandwith in both case?
Maybe cange the Loopbandwtih in teh passive case al littel bit.
The Ampilfire in the Aktive LF work like a additional TP.

Greez
 

@volker_muehlhaus : Resistor values are around 20 Ohm, 50 Ohm. Not so big.

@Zottel : Same BW in both case, also I tried changing BW but passive seems to be worse.
 

Have you simulat the PLL with a Sim Tool like the ADisimPLL?
 

Yes, but I am using a VCO with internal doubler. So, in ADISimPLL I used a half freq. version with exactly the same Kv.
 

What are teh Results? For Passive and aktive LF, the same like Measurement?
 

It is strange for the VCO to get a single freqency with 4.3/4.7V vtune. One case should be unlocked. You can try to tune the Charge pump current register to observe and optimize the phase noise. ADF4157 has best performance when VTUner is btween 0.5V and Vp-0.5V. The Vp should be clean.
 
Actually I set the charge pump current to maximum to minimize phase noise, but I may try different values.

Thanks,
 

kayaoo said:
It is strange for the VCO to get a single freqency with 4.3/4.7V vtune. One case should be unlocked. You can try to tune the Charge pump current register to observe and optimize the phase noise. ADF4157 has best performance when VTUner is btween 0.5V and Vp-0.5V. The Vp should be clean.
Click to expand...

I second that opinion. Sometimes it is not clear if a PLL is locked or not unless you do some testing. I like to replace the reference xtal with a lab synthesizer and move the reference frequency a few 100 Hz up and down, and see if the VCO moves N x a few 100 Hz.

Also, you have to make sure your PLL loop filter is stable or it might add peaking noise. A nice way to do that is to add an analog phase shifter to the reference oscillator path, and do a digital step of say 5 degrees while watching the Vtune voltage with a x10 probe/oscillosope. If it is stable, it will look somewhat like a critically damped response. If unstable, it will ring a lot before settling down.

You should also make sure you are using a low noise op amp. It is tempting to use CMOS single supply rail to rail op amps, but they can have horrendous inpur noise, like 25 nV/√Hz. Try to find one with more like 3 nV/√Hz. Also make sure the power supplies are really well decoupled with big capacitors!
 
I am sure PLL is locked and loop filter (50 Degrre margin) is stable, I don't see any ringings.

Thanks
 

The VCO should not have 4.7V in one configuration, and 4.3V in the other if the locked output frequency is the same , the temperature is the same, and the load is the same. Either one or both cases are unlocked, or there is some weird oscillation going on on the Vtune line that you voltmeter is rectifying.
 

Try to add a big Tan capacitor to the Power supply of VCO.
 

Some VCOs' varactor is added between Vcc and Vtune pin rather than between Vtune pin and Ground pin.You can check the VCO VDD between cases.The noise between Vp and Vdd_VCO should be filtered out.
 

@biff44 : Actually I checked the whole band with spectrum analyzer. There is no weird oscillation.
 

The worse phase noise must be in un-lock state.
passive PLL with 4.7V Vt output is nearly the upper limit, i.e. +5V, so that is in un-lock state.
Check your PLL design.
 
did you design the vco or are you purchasing one, if so what is the part no?
 
Here is one thought. I have found that op amps near their rail sometimes stop having gain but still kind of work, so when you are at 4.7 V out, is the op amp near a rail (say at +5V). If so, the gain is reduced from what it should be, the whole loop filter has different gain/phase characteristics, and the VCO may be loosely locked, but not truly locked. Maybe you need to increase the rail voltage to 6 or 7 volts, and the VCO will lock right up with good phase noise?
 
Actually I tried incresing rail voltage but not upto 6 V.
Thanks,
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top