When is JFET used in RF circuits?

[aht2000]

I am trying to understand when to use JFET in an RF design. The band in mind is the FM Band (88-108MHz). I have seen statements like "This oscillator will only work using a JFET not a BJT as its active device" and other statements like "for our low noise front end, we will be using a JFET", ... etc

My basic understanding of JFETs is that their gain is low compared to BJTs, their parameters differ considerably from device to the other even within the same part number, not sure if they operate successfully with low supply voltages like 5V or 3.3V. They have very high input impedance as their gate is isolated, but this is only in low frequency bands (audio ?), once the frequency increases, the device input impedance comes into play and ruin the high input impedance advantage.

In BJTs, there are low power, high power, audio, RF, low noise,... etc type of devices, I assume the same exists for JFETs?

 

[D.A.(Tony)Stewart]

JFETs are depletion mode low power, low-noise figure devices with low capacitance and high impedance making them suitable for constant current sources and fast switches but limited in the range slightly lower than 25 ohms to 100 ohms minimum. They are commonly used inside all elected mics.

SiC JFETs are a new breed with lower resistance capable of driving MOSFET gate capacitance.
They rise in resistance up to the pinch off voltage.

JFET - Wikipedia

en.m.wikipedia.org

Lab 4 - JFET Circuits I | Instrumentation LAB

instrumentationlab.berkeley.edu

 

[KlausST]

Hi,

@aht2000
I'd say it depends. You may use it because of several reasons.
And there are many different JFETs some may be useful for your frequency range, other are not.
So it really depends..

JFETs is that their gain is low compared to BJTs

I'd say this generally is not true.
You need to specify
* which circuit
* which frequency range
* which input impedance
* whether you are talking about voltage gain or current gain.

Voltage gain: mainly deoends on circuit
Current gain: At DC the gate current is negligible, thus the current gain is almost infinite. But the gate mainly acts as capacitance, thus the gate current depends on frequency...and the current gain, too.

Klaus

 

[vfone]

JFETs have better IMD linearity and higher dynamic range than BJTs and MOSFETs, mainly because they have slower rise of IMD3 and IMD5 with output power.
JFETs linearity is that, unlike BJTs and MOSFETs, they have nonsaturating I-V charactcristics.
In any transistor type, the maximum available RF power is function of the load impedance variations. JFETs have wider load impedance requirements (than BJT and MOSFET) to get constant output power.
JFETs works fine at lower supply voltages.
The drawback is, the JFETs require a negative bias and they are limited in frequency range, up to about 1GHz.

 

[aht2000]

Thank you. If we focus more on its use as a low noise front end in a receiver (88-108MHz), would it be a better choice than a BJT taking into consideration that the supply voltage is 5V?

Assuming I am given a specific target gain and noise figure for this front end, can I reach the same target noise figure and gain using BJT instead.

I am sorry about being so generic with no specific figures but I just need to decide which way to focus on. A response like "if you want a NF below x, and gain above y then JFET is your best choice (a recommeded part# would be great), above that a BJT can easily achieve it (a recommended part#)".

 

[betwixt]

At those frequencies and in that application it wouldn't make much difference which type of device you used. For more specialized needs there might be a preference of one over the other but you can use either at the low(ish) frequency and power levels you are talking about. The remainder of the circuit would have far greater influence on performance than the transistor itself.

Brian.

 

[BradtheRad]

can I reach the same target noise figure and gain using BJT instead.

The JFET draws miniscule current at its bias terminal, which makes it useful to detect weak signals. You need not raise bias to a threshold voltage

I purchased a 2N3819 at Radio Shack, and learned how it works as I applied negative voltage from two AA batteries through a high-Mohm potentiometer. (Minus 1.7V pinches off conduction.)

I showed it to a local electronics genius who built an LC circuit with adjustable resonant frequency. We wished to see if we could measure the rf field strength of our 49 MHz walkie-talkies. It worked, so I added an antenna and transistor amplification stage to drive my VOM. It was powered by a 9V battery (in addition to the two AA). So we had a portable meter which read upscale to indicate stronger rf field.

AC comes from the antenna. The fet responds to the positive waveforms. Spec sheets say the 2N3819 works up to 100 MHz.

 

[dick_freebird]

There are "RF" JFETs that were used in LNAs and such, often actually MESFET (Schottky gate) construction.

I once upon a time did a series of element controllers for space radar panels, and the customer liked a NE32S10 FET for the pass device. NJFETs are cute for controlling a supply source that is outside the rails, from a between-rails error amp (depletion mode). But the device application originally was few-hundred MHz signal path.

 

[vfone]

Even they look similar, JFETs and MESFETs are different, mainly how their gates are made (JEFTs have higher gate barrier).
MESFETs usually are built in Gallium Arsenide (max. freq. about 45GHz) when JFETs are silicon-based (less than 1GHz).

http://educypedia.karadimov.info/library/1%20JFETs%20and%20MESFETs.pdf