complex bjt transfer function amplifier

[yefj]

Hello, I have a Q2N2222 on PSPICE simulator.
I need to build a circuit which has a certain gain ,but the main problem is the POLES.
In the AC transfer need a zero to be first order (rise of 20db/dec) but the end needs to be a fall of 40db/dec.
So at start it needs to be a zero at a certain start frequency 20dB/dec ,then we have fall at Fc and then we have another double pole.
Anybody hase an idea of how to desin such a thing?
if with a single bjt its not possible then ill be glad to see some general method.
Thanks.

 

[sutapanaki]

The best way to start designing something is simply to start.

 

[yefj]

Hello Sutapanaki, as you can see bellow i managed to get the needed Gain,
but i dont know how to make a double pole in the end of transmission and how to controll its position?
Thanks.

 

[sutapanaki]

Maybe it is a good idea to make a drawing of what you want to achieve.

 

[yefj]

Hello Sutapanaki,the transfer function shown bellow is what i am trying to achieve.
I can see that we need serios controll of poles and zeros.
it doesnt have to be one bjt transistor,could you give me a direction of a structure which has similar unique transfer function?

Thanks.

 

[FvM]

1. Learn to draw bode diagrams to scale. In your diagram, the so called +20 dB slope is steeper than the -40 dB slope. Also use logarithmic gain scale for the simulation plot.

2. Learn to identify poles and zeros in circuits. The BJT stage has two zeros related to C3 and C1 and a zero/pole combination (lead/lag) related to C2. Plus additional poles and zeros caused by the transistor.

In so far the complete transfer function is more complex than your specification. To make a circuit with flat gain up to 170 MHz you need transistors with GHz transit frequency instead of 2N2222.

More poles can be added as passive RC or LC low passes.

 

[BradtheRad]

Not offering a solution readymade...
In my collection of bandpass simulations is this concept for altering the slope of a filter. Notice the lopsided appearance of the slopes. The high end drops off more steeply.

There may be a way to adapt this to your topology. Perhaps you only need to make a few changes. It does not absolutely need an inductor. I suppose the key is in the capacitive divider.

 

[yefj]

Hello FVM,I need the first stage to be bjt amplifier,I can add to it other mosfet stages or feedback.
Could you recommend me a structure which could have such transfer function?
Where i could play with its properties to achieve such transfer function of 20dB/dec rise and -40dB/dec fall.
Thanks.

--- Updated Jan 23, 2021 ---

Hello FVM,another big problem is the bandwidth,i dont know how many stages to add to my bjt stage in order to get to my results?I can use any MOSFET LTSPICE available.
COuld you recommend me a structure i could try analize to get the desired transfer function?
Thanks.

--- Updated Jan 23, 2021 ---

Hello FVM,another big problem is the bandwidth,i dont know how many stages to add to my bjt stage in order to get to my results?I can use any MOSFET LTSPICE available.
COuld you recommend me a structure i could try analize to get the desired transfer function?
Thanks.

 

[yefj]

Hello FVM,I have LTSPICE i understand that i need to add to my 2N2222 anothe stage could you recomand a mosfet that has a high transit frequency?

"In so far the complete transfer function is more complex than your specification. To make a circuit with flat gain up to 170 MHz you need transistors with GHz transit frequency instead of 2N2222. "

 

[sutapanaki]

Sinse you have a zero at pretty low frequencies, you can try to achieve it with AC decoupling at the input, as you've done. The zero will be defined by the decoupling capacitor and your total input resistance of the stage. Then you can split your gain in two stages (you don't mention how much gain you need) and load each stage capacitively to get your two poles.

 

[yefj]

Hello Sutapanki,I need 40dB gain which i succseed,You can see bellow the CE i have done so far i added a decoupling capacitor to the output but the slope in the end is 20dB/dec as you can see bellow,how can i turn it into 40dB/dec?
Thanks

 

[sutapanaki]

If you want to rip the benefits of your design, start by analyzing this stage. It is actually not so difficult. Try to find where the zeroes and poles are and then you will know what to do about them.

 

[yefj]

Hello sutapanaki,the two poles formula are shown bellow.
i cant see how do i make them one double pole.
is there a way?
Thanks

 

[FvM]

By matching the frequency of both poles.

 

[BradtheRad]

The series capacitor is a high-pass filter, first order. Rolloff 20 dB/dec.

Two capacitor integrators cascaded make a 2nd order low-pass. Rolloff 40 dB/dec. (Each stage needs its own input resistor.)

Place all one after another to create your bandpass.

You can install the one-transistor amplifier most anywhere.

 

[yefj]

Hello BradtheRad, In the first photo bellow we have 20dB/Dec,i added another RC pair and still i have 20db/dec slope in the end.
Where did i go wrong?

 

[FvM]

You are adding zeros (high passes) where you should add poles (low passes).

 

[sutapanaki]

In your original schematic you have two poles and 3 zeroes, which is not going to give you -40dB/dec roll off. As was mentioned before, add an extra stage. The second stage does not have to be DC decoupled from the first one.

 

[yefj]

Hello Sutapanaki,suppose i want a 40dB gain and 100Mhz bandwidth,You said that mosfet is more suitable for this task.
could you please recommend me and LTSPICE mosfet transistor i can use?
Thanks.

 

[sutapanaki]

Hello Sutapanaki,suppose i want a 40dB gain and 100Mhz bandwidth,You said that mosfet is more suitable for this task.
could you please recommend me and LTSPICE mosfet transistor i can use?
Thanks.

I don't think I have ever said that a mosfet stage is more suitable. You can do it with bjt as well.