Designer’s Guide Community

mohammadreza

Junior Member

Offline

Posts: 19

Oscillation in PA
Aug 16^th, 2010, 2:15am

I have designed a cascode class A RF power amplifier operating in 2.45GHz. In pss analysis I am getting reasonable amplitude in my RF frequency. However, In transient analysis I see ripples which has got pepriod of 20ns.

can anyone help me how I can sort out this problem and make sure that my PA is stabel. I have done some tests that are as below. Right now I think the only way to get rid of it is using very large indictors and capacitors in biasing.

I changed the input amplitude and even my biasing elements, and they just affected it in the primary oscilation amplitudes. I replaced the current mirror, that I used for biasing the PA, with a voltage source and a large inductor to make sure that's not causing the oscilation. I also changed the size of the transistors and they didn't make any change in the period of oscilation. the only elements that cause the oscilation are the RF choke and the ac couple cap I use at the drain of my PA. If I make them very big, the oscilation is not seen in the transient, and that's perhaps beacuse the frequency of oscilation has moved to a lower one than the transient. because when I change them not significantly, the period of that oscilation changes and it's exactly equal to the period of the related resonant frequency made by the cap and ind.

Back to top

IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Oscillation in PA Reply #1 - Aug 16^th, 2010, 4:09am try running the simulation to 2 micro second and check whether you see the oscialltion..
Back to top	IP Logged

mohammadreza Junior Member Offline Posts: 19	Re: Oscillation in PA Reply #2 - Aug 16^th, 2010, 4:18am yes, that exists.
Back to top	IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Oscillation in PA Reply #3 - Aug 16^th, 2010, 4:39am May be you have low frequency stability problem..have u chekced the stability circles and k factor
Back to top	IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Oscillation in PA Reply #4 - Aug 16^th, 2010, 4:41am the period of that oscilation changes and it's exactly equal to the period of the related resonant frequency made by the cap and ind. ------------------------------------------------------------------- put some resistance with inductor in series and check..
Back to top	IP Logged

mohammadreza Junior Member Offline Posts: 19	Re: Oscillation in PA Reply #5 - Aug 16^th, 2010, 6:55am well, in fact in my circuit the inductor is not connected directly to vdd. its connected to a parallel RC, r=50Ohm and C-1uF.
Back to top	IP Logged

mohammadreza

Junior Member

Offline

Posts: 19

Re: Oscillation in PA
Reply #6 - Aug 16^th, 2010, 7:03am

I am not sure of my stability factor measurement. So please confirm me if I am right.

To plot the Kf, I just set my input port to a dc type and run the sp analysis, without doing any other changes. Then I direct plot the kf. Its very high especially at dc. but when I added a marker to see if somewhere it is 1, it was around 20MHz.

So, what should I look for now? there is no feedbacks in my circuit, so how can it be possible? should I reduce the number of fingers in my circuit from 10 to 5 and double the consumption in the biasing mirror?

Back to top

IP Logged

mohammadreza

Junior Member

Offline

Posts: 19

Re: Oscillation in PA
Reply #7 - Aug 16^th, 2010, 7:35am

If I am alright with the way I am measuring my Kf, then I should say that the problem is solved like this.

In order to get rid of the frequency resonance of the BFL and BFC, I should use one of the inductors of the pi matching network connected to vdd instead of the ground, which provides dc connection required as well.

Then, when I did this, I faced that for lower input amplitudes, the output is damped. but I had good results in the pss analysis.
then, there must have been another powerful oscilation as well. checking the effect of the bypass capacitance on the Kf, it shouldn't be very big. so, by decreasing that capacitance, another implicit resonance factor that damped my output for lower inputs is removed.

thanks for notifying me about the Kf. just please let me know if I have to take care of anything in my simulation configurations.

Back to top

IP Logged

purplewolf

Senior Member

Offline

Posts: 103

Re: Oscillation in PA
Reply #8 - Aug 16^th, 2010, 7:54am

mohammadreza wrote on Aug 16^th, 2010, 7:03am:

I am not sure of my stability factor measurement. So please confirm me if I am right.

To plot the Kf, I just set my input port to a dc type and run the sp analysis, without doing any other changes. Then I direct plot the kf. Its very high especially at dc. but when I added a marker to see if somewhere it is 1, it was around 20MHz.

So, what should I look for now? there is no feedbacks in my circuit, so how can it be possible? should I reduce the number of fingers in my circuit from 10 to 5 and double the consumption in the biasing mirror?

Device has lots of gain at low frequency and hence it has stability problems...at very high frequeuncies the 'active' device turns into 'passive' one; i.e. it has not gain hence no stability problems...
check the stability for a all possible bias points uptil frequenc uptil 10 times your operating frequency...use both Kf and circles..
s12 (Cgs) is reponsible for staility problems you might be facing..

Back to top

IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Oscillation in PA
Reply #9 - Aug 16^th, 2010, 8:30am

mohammadreza wrote on Aug 16^th, 2010, 7:03am:

I am not sure of my stability factor measurement. So please confirm me if I am right.

Generally you have to consider stability of the following two states.

(1) Quiescent State(Stability Study based on DC Operating Points)
(2) Dynamic State(Stabilty Study based on Periodical Steady State Oerationg Points)

Conventional K-factor is categorized as (1).
But you have to consider stabilty as (2) where input power levels must be swept in PA.

Some oscillations in (1) could be disappeared in (2) especially when input power level is high.
Vice versa, another osciilation which does not exist in (1) could be appeared in (2) when input level is at specific level or more than specific level or less than specific level.

See the followings.
http://www.designers-guide.org/Forum/YaBB.pl?num=1272344339/3#3
http://www.designers-guide.org/Forum/YaBB.pl?num=1280434484/2#2

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged