understanding JFET fixed bias method

[fm101]

Hi,
I have the following 2N5459 JFET fixed biased circuit in proteus as shown below. I calculated drain current Id and drain voltage Vd but the circuit simulation does not show the expected values. I am not sure whether my calculation is correct or it is proteus.
I took Vgs(off)=Vp=-2V with Idss=4mA which are minimum values from 2N5459 datasheet. With applied Vgs=-1V, Vdd=5V and using transconductance equation Id=Idss(1-Vgs/Vp)^2 gives Id=1mA and with Vd=Vdd-Id*Rd gives Vd=5V-(1mA)(1kOhm)=4V but the simulation shows Id=0 and Vd=5V.

Is this not the way to bias JFET?

 

[frankrose]

does anybody know the equation for drain characteristics, that is equation relating Id and the Vds for given Vgs?

SPICE-compatible N-Channel JFET - MATLAB

The SPICE NJFET block represents a SPICE-compatible N-channel junction field-effect transistor (NJFET).

www.mathworks.com

check "Drain-Source Current-Voltage Model"

 

[danadakk]

does anybody know the equation for drain characteristics, that is equation relating Id and the Vds for given Vgs?

Regards, Dana.

 

[Audioguru]

@Audioguru, how come the voltage range 1 to 4? I am using 2N5459 JFET

The datasheet says that some of them are turned off when the Vgs is -2V all the way to others with a Vgs of -8V which is a range of 4 times!
The datasheet also says that some of them have a current of 4ma all the way to others with 16ma when Vgs is 0V another range of 4 times!
You will need to buy thousands of them to work properly with one bias voltage or adjust the bias voltage for each one. DC Negative Feedback can be used so that the Jfet picks its own Vgs.

 

[fm101]

It turns out the 2n5459 spice model in proteus does not reflect the datasheet accurately. The drain curve for the 2n5459 spice model plotted in proteus is shown below. And with the knowledge of this I was able to get expected drain current, drain voltage and gate to source voltage.

In my original question I assumed Vgs=-1V and the calculation showed drain current Id=1mA which didn't match with simulation result. According to the graph when Vgs=-1V the drain current is in the order of micro Ampere. I tested with Vgs=-0.6V then the result is as expected with drain current of 1.6mA according to graph. However, the drain current with Vgs=-0.6V and Vp=1.2V gives drain current of 2.5mA which does not match with proteus spice model. The solution is explained in details in the following blog post.

JFET Biasing Worked Out Example Calculation

A worked out example of how to bias JFET with gate shorted, fixed bias and self bias.

ee-diary.blogspot.com

 

[KlausST]

Hi,

It turns out the 2n5459 spice model in proteus does not reflect the datasheet accurately.

This part has a big variance ... thus there are no accurate values one can count on.
I still think that you ignore the variance.

Run 5 simulations and you get 5 times the same result.
Use 5 real parts and you get 5 different results.
--> I'd say the simulation is useless for your application. You gain no information for a real circuit.

The graph of your simulation result is somehow useless without schematic / simulation conditions.

Klaus

 

[FvM]

This part has a big variance ... thus there are no accurate values one can count on.
I still think that you ignore the variance.

JFET parameters have minimal, typical and maximal values. The shown Vp of 0.9 V is far outside 2N5459 specification, a real part with this characteristic has to be sorted out. This was obvious from the start of this thread, see post #9.

In my view, Proteus model parameters are completely intransparent. Proteus has strong points in the simulation of microprocessors, but I won't trust it as analog simulator.

 

[fm101]

Yeah, a spice or other simulation model of a real physical component part can never replace that part. And this is worsened by the fact that same component has different electrical characteristics because of manufacturing process. Besides even more so with 2n5459 as it has wide spread electrical parameters as @Audioguru indicated.
Nonetheless, even if schematic simulation and prototyping is not exact it is useful for education and understanding. I should also correct the fact that Proteus is not to blame as it depends upon the spice model which are usually supplied by other parties. And spice model can also not be blamed entirely since physical object can not always be mathematically represented accurately.