Basic N-JFET amp question

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

stickjam

Well-known member
Joined
Jun 17, 2004
Messages
325
Location
Grand Rapids MI
Forgive me if this is basic stuff, but I'm getting back into this after a long absence...

A lot of N-JFET amplifier schematics I've seen mirror triode design:

- drain resistor (eg. 10K) to Vdd
- gate resistor (usually 1Meg) to ground
- a source resistor (eg. 4.7K) often in parallel with a cap (50uF) to ground

I've also seen an occasional variation where...

- there's a resistor in series with the source-to-ground cap.
- the gate resistor ties to a low bias voltage (Vdd/11) instead of ground.

What would be the design intent of those variations? Are those variations found or similarly relevant in triode applications?

Thanks

--Bob
 
[quote author="stickjam"]

- there's a resistor in series with the source-to-ground cap.[/quote]

To reduce the gain w.r.t. the no-series-resistor-situation and/or to get more grip on the gain. Gain is now more exactly set by the ratio of the drain resistor & the cap-series-R (assuming a Rseries >> 1/gm).

Without the series-R the gain would be determined by gm*Rdrain. And gm on it's turn is determined (among others) by drain-current. Drain current is (among others) set by the threshold voltage and the Rsource.... you'll get the picture, it's way more fuzzy/indirect.

Bye,

Peter
 
> there's a resistor in series with the source-to-ground cap.
> the gate resistor ties to a low bias voltage (Vdd/11) instead of ground.

A JFET is the same as a vacuum tube except they pour silicon into the vacuum, and the grid pushes from the side instead of acting like a mesh fence.

So tranconductance is high (silicon conducts better, gate has more control than grid) but very variable (all vacuum is the same, silicon varies from batch to batch). Gain and bias current vary a lot from JFET to JFET, much more than tubes.

To set DC bias, you can just ground the Source and run at Idss, but that can vary 2mA to 20mA within the same part number. A source resistor will reduce and stabilize current, but by an amount that depends on Vp, which may range from 1V to 5V on the same part number. In many applications, the exact bias current is not critical, and a compromise source resistor will handle any FET in the bag well enough. If precise current is needed, a large source resistor with a gate voltage larger than Vp will set current more exactly. If Vgate is +100V and Rsource is 100K, then source current WILL be 1.01mA to 1.05mA for FETs with Vp from 1V to 5V. In many cases a few volt bias is enough to swamp Vp variation well enough.

To set AC(audio) gain, a source resistor reduces Gm. If this resistor is larger than 1/Gm, gain will mostly be controlled by the +/-5% resistor rather than the +/-50% FET. Even a resistor less than 1/Gm will tame FETs on the high-gain end of the tolerance, and add linearity.

Since in general, the source resistor needed for good DC bias and the source resistor needed for best AC gain tolerance and linearity are never the same, except by accident, we consider tricks. When the AC resistor should be lower than the DC resistor, the source resistor is only partly bypassed. DC bias is set by Rs1+Rs2, AC gain is affected by just Rs1. When the AC resistor must be larger than the zero-gate DC resistor, you use the AC resistor in the source and raise the gate voltage to get your bias current back.

Such tricks are done in vacuum tubes, but very rarely. Tubes are more consistent in DC current, and because they traditionally run with plate supply voltage much-much larger than signal voltage, linearity is not such a problem as in FETs working with supply voltage not much greater than signal voltage.
 
Back
Top