RF Buffer Amplifier design

[afz23]

I want to design an RF buffer amplifier(@4GHz) using a BJT amplifier device.

I need this to provide a good isolation between two circuits(PLL and PA).

I want to simulate this design
I have an s-parm file, of a BJT device,provided by manufacturer at an operating bias.
I donot have the SPICE model of this device.

How to begin this design? please guide,some reference material will be very helpful.

 

[Bob60]

SPICE you don`t needs. Typically it`s enough with S par and P1 from data sheet. But isolation you can get from one stage is
not more than -20dB and gain 10-15dB, with P1 10-15dBm and current consumption up to 30-100mA. Best SiGe BJTs are Infenion, NEC, NXP.

 

[afz23]

SPICE you don`t needs. Typically it`s enough with S par and P1 from data sheet. But isolation you can get from one stage is
not more than -20dB and gain 10-15dB, with P1 10-15dBm and current consumption up to 30-100mA. Best SiGe BJTs are Infenion, NEC, NXP.

I am looking for more than 30 dB isolation,may be a 2-stage buffer will be required.
Can some member share some related buffer design?

Do you mean ,s-param measured at P1 dB? But manufacturer will provide s-param in linear region.

 

[BigBoss]

SPICE Parameters are of course necessary to simulate the Non-linearity of the buffer.But if you aren't able to find, s-parameters are also used in a limited manner..
A buffer should have high input impedance and low output impedance.For this target, a simple Differential ( CasCode) Amplifier with/without emitter regeneration resistors following a Common Collector Amplifier ( Emitter Follower) will give you a good isolation between PLL and PA.Stages cane be more than two but in general 2 stages will provide at least 30-50dB isolation if you use Cascode Differential Pair.
SPICE model is necessary when Non-linearity is important ( so it's generally important).
Darlington configuration is another choice instead of Emitter Follower if the isolation doesn't become sufficient but it may be more nonlinear.

 

[afz23]

Thanks BigBoss for your suggestion

But I am not very clear,when you say,s-param file will be able to do the job,as shown in figure cascode is CE-CB 2-stage configuration.

Manufacturer provide s-param file ,which is just a two port data extracted in CE configuration,things like putting regeneration resistor at
emitter and connecting BJT in the CB configuration as shown in the figure is not possible with the available s-param file.

I think SPICE mode is must for the purpose.

I am using AT42070,can any member provide SPICE model for it.

 

[vfone]

Here you can find the Spice model:
http://www.hp.woodshot.com/hprfhelp/design/SPICE/atpkgs.htm#at70pkg

At 4GHz on the top of using Spice or S-param files, the design of the PCB layout would play a very important role getting good isolation of the buffer amplifier.

 

[BigBoss]

View attachment 99062
Thanks BigBoss for your suggestion

But I am not very clear,when you say,s-param file will be able to do the job,as shown in figure cascode is CE-CB 2-stage configuration.

Manufacturer provide s-param file ,which is just a two port data extracted in CE configuration,things like putting regeneration resistor at
emitter and connecting BJT in the CB configuration as shown in the figure is not possible with the available s-param file.

I think SPICE mode is must for the purpose.

I am using AT42070,can any member provide SPICE model for it.

AT42070 is obsolete and you'd better find another equivalent for that transistor.
It's expensive too...
However if you still would to use it, please combine AT-420 die model and Package 70 model for full packaged SPICE model.

 

[mtwieg]

You might be better off using an amplifier followed by an attenuator. This will get very good isolation, though it will require much more power in the amp to get low distortion.