The numerical solution using the equivalent circuits is in the Solutions Folder in the Library.
The complete circuit with feedback is given below. Even though the resistor values used in this schematic are for generic transistors, it was assumed that the same values can be used with the transistors available in the P-spice files without changing the transistor parameters.
Once the complete feedback ciruit is simulated, differences in both the drain and collector currents are found (compared to the orinally assumed currents). But, the transistors still operated in their proper regions and therefore the rest of the simulations are carried out using these resistor values and under the simulated bias conditions.
Note: Since this is a DC coupled circuit, you don't expect a low frequency -3dB pole.
In this figure, numbers with 
color are the design values. The numbers with 
color are the results from the P-Spice simulation.
The A-network is easily obtained by using the guidelines for the loading effects of the feedback network. This is done by breaking the connection between the Source of the FET and the collector of the BJT. The source is grounded (R11=0) and a resistance R22=RE is connected from the collector of the BJT to ground. The 
=1.
There is one very important point to remember when this is done. Due to the break in the connection between the source and the collector of the transistors, the bias conditions on the two transistors will change since the source of the FET is directly connected to ground.
In order to avoid changing the bias points of the transistors so that in the simulations of the A-network the same ac conditions will prevail, resistors should be added to both the source terminal of the JFET and the collector terminal of the BJT to maintain the original bias conditions. These resistors are shown in the Figure below. In order to maintain the same ac conditions of the feedback network, large shunt capacitors are used to by-pass these resistors.
The current and voltage values for the A-network are shown in color, The currents and voltages obtained from the Feedback network originally simulated are shown in color. The bias points are almost the same as the feedback network.
(These are read from the Output File of the P-Spice Program)
The Gain vs Frequency plots are shown in the following figures.
The corresponding output voltges and cursor readings for the -3dB points from the two curves are
Note that from this data, the equation Af = A/(1+ A) = 4.79/(1+4.79)=0.827, which is very close to the Af = 0.834.
On the other hand the 
Hf= H(1+ A) = 0.312(1+4.79) = 1.81 MHz. This is off from 1.214 MHz found from the feedback network. This may be due to lack of a single dominant high frequency response pole.
