Why my Power Amplifier is oscillating at 1000MHz?

hi friends,
why my PA is oscillating at 100MHz... those every thing is okay

Re: Power Amplifier

Dear,

please send more information such as working frequency,type of transistor (bipolar,fet,mosfet,ldmos),class of operation (A,AB,...) and any other useful information.

good luck.

Re: Power Amplifier

memarian said:

Dear,

please send more information such as working frequency,type of transistor (bipolar,fet,mosfet,ldmos),class of operation (A,AB,...) and any other useful information.

good luck.

Click to expand...

thanks..yaar
my working Freq. is 3.4-3.6 GHz, PHEMT, Class-A, P1dB=27dBm,Gain=28dB,VDS=3.4,VGS=0.5
Please help me to over come this Oscillations, I am just startup in design...
Plaeaee

Power Amplifier

pls check your phase margin and gain margin

Power Amplifier

Do you have proper input and output match networks for the transistor? Do you have proper power supply decoupling circuitry? Try to add more grounded points on the PCB. You should know the gain of a transistor at low frequency(say 100MHz) is pretty much higher than at high frequency(3.5GHz), so you need to check the datasheet of this transistor and do the stability analysis not only at your working frequency, but lower ones.

Re: Power Amplifier

If you are using cadence, use the K-Factor simulation to ensure stability, for more information about this simulation check spectre manual

I you are using any other simulator, make S-Param simulation, and then you can get the K factor and the Delta factor from them,I dont remember the expressions but you can find them in any PA text, K must be greater than one and delta smaller than one...

PA's generally have to types of instability, low freq due to supply and ground termination problems and high frequency due to improper input and output matching.

Re: Power Amplifier

aomeen said:

If you are using cadence, use the K-Factor simulation to ensure stability, for more information about this simulation check spectre manual

I you are using any other simulator, make S-Param simulation, and then you can get the K factor and the Delta factor from them,I dont remember the expressions but you can find them in any PA text, K must be greater than one and delta smaller than one...

PA's generally have to types of instability, low freq due to supply and ground termination problems and high frequency due to improper input and output matching.

Click to expand...

Thanks buddy ++++++
I am using ADS2005A and i have done S-Param 4 stability which is >1 , I have used //LC tanck CKt for drain & gate biasing which is C=2pf and L=1nh for 3.5GHz centre frequency, in put match is better than 15dB and s22 is 9dB,i did not understand clearly can u pls eleborate me ur this sentance, PA's generally have to types of instability, low freq due to supply and ground termination problems and high frequency due to improper input and output "...what i have used 4 biasing is correct or not....
Pls. help me i do not know much abt PA design........i shall be highly gre8 full 2 u.
Thanks dear........

Added after 6 minutes:

windman said:

Do you have proper input and output match networks for the transistor? Do you have proper power supply decoupling circuitry? Try to add more grounded points on the PCB. You should know the gain of a transistor at low frequency(say 100MHz) is pretty much higher than at high frequency(3.5GHz), so you need to check the datasheet of this transistor and do the stability analysis not only at your working frequency, but lower ones.

Click to expand...

Thanks yaaar+++++++++
yaa i have proper input/output matching but i realy don't know wether i am having proper power suply or not .can u tell me how 2 check power supply decoupling circuitry, i have used //LC tanck ckt whose value is c=2pf and L=1nh at 3.5GHz.
thanks+++++

Re: Power Amplifier

It is osc out of band. Maybe the gain is too high at low frequency. Loss can be added in biasing circuit at the osc frequency.You can also try to add a series of bypass capacitors with different values at gate bias and drain bias.