Tweet

Want a Stomp Box Preamp with Gain to Spare? Step up to a 2SK370 JFET! Our no-fuss, one-transistor circuit might just be your next overdrive project! A Toshiba 2SK370 has a transconductance of up to 22 mS. (Even more as drain current increases to IDSS or 12 mA, the maximum.) That much sensitivity in one JFET can sure work miracles. This is a low-noise device that finds its main use in microphone preamplifiers. It replaces the classic 2SK170 that Toshiba discontinued. By popular demand, Toshiba brought out the newer device. (Chinese sources sell both parts. In this project, either type will work fine. You can interchange them.)

Pinout for 2SK370 (from Toshiba)

No wonder the hi-fi crowd likes 2SK370 JFETS. With nine-volt power, a 2SK370 has ten times the voltage gain of an MPF102 JFET. This gain figure is actually a little less gain than I'd expected. The device's typical 22 mS transconductance figure predicts an unloaded voltage gain of 24. But there's much variation among JFETs. Besides, I didn't operate the 2SK370s strictly according to the datasheet. (The datasheet specifies 10-volt power and 3 mA quiescent current.)

The 2SK370 Preamp Circuit I breadboarded a circuit with a voltage gain of 15. I tested the new circuit with five devices. With each of the five JFETs, the circuit worked splendidly. MPF102 vs. 2SK370. For comparison, my MPF102 circuits require a 560-ohm source Resistor. The comparison is particularly apt, because both circuits operate at a 2-mA current. (I'm running the 2SK370 with 2.2K for the drain resistor. The source resistor is 75 ohms. I plugged in a 3.3M gate resistor. But at the gate, almost any reasonable value would have worked: Say 100K to 10M.) The table below compares my MPF102 and 2SK370 preamplifiers...

Schematic of my 2SK370, 15X JFET preamp

Some batches of the 2SK370 might be hotter than the ones that I tried. If you find one of these "hot" devices, you'll have to reduce the source resistor. Try values like 68, 56, or 47 ohms. Change resistors with the power off. (Otherwise, transients can kill a JFET.) After each resistor change, switch on the power.

With with no signal on the gate, check the drain voltage. The goal is to get as near as possible to half the power voltage. When you do, you've found the right source resistor. Common question: How do I know to increase or reduce the resistor size? Answer: Smaller source resistor values reduce the drain voltage. Larger values increase the drain voltage.

For Tube-O-Philes

Missing data. Tube-o-philes will notice “missing data” in the table above: The table doesn't include plate resistance (R_p) or amplification factor (Mu or µ). In a JFET, we assume that both are infinite. (Anyway, try to find them on a JFET datasheet!) The upshot: Unfortunately with FETs, the famous triode gain equation isn't very useful.

Famous Triode Gain Equation (Seldom useful for JFETs)
G_M= (µ / R_P)

Drain curves. Take a look at the flat-top drain curves nearby (similar to plate curves). These curves show a high drain resistance. The device operates in the flat curve region right of the knee (left side). After the knee, drain voltage has little effect on the JFET drain current.

Plate resistance vs. on resistance. Some FET datasheets list “on resistance” (R_DS(ON)). Take care! This figure isn't equivalent to the old plate resistance. Instead, on resistance is the reciprocal of the FET transconductance, 1/G_M. (Some datasheets call G_M by other names: G_FS or Y_FS.)

Still wondering about “plate resistance”? According to expert Albert Malvino¹, the “plate resistance” (actually drain resistance) of a FET is the reciprocal of the spec G_OS. This spec is the output transconductance of the device. As G_OS is a rather obscure spec, you won't find it on all datasheets. The reciprocal gives pentode-like values of resistance, 10K to about 500K. (For the MPF102, you'll find 20,000 to 66,000.)

Lower Current Limit Reducing current. I also tried the 2SK370 at 1-mA current. The gain isn't really much better. As I suspected, the device acts flaky and runs out of steam. When the drain resistor is 4.7K, the required source resistor is 180 ohms. The device shouldn't really work at all: At 1 mA average drain current, the device is in its current pinch-off region. (Or close by.) With a 150-ohm source resistor, the drain voltage fell from optimum to a little over 2 volts. This quick drop indicates instability. Such instability is typical behavior in the current pinch-off region: Apparently some 2SK370 devices would be unusable at 1 mA.

Drain curves for 2SK370 JFET

I've seen worse: Device impedance can rise excessively high: So high that any load drags the drain down to zero volts. Even a coupling capacitor lead will knock out the drain signal.

Ways to Increase the Gain

Back to the 2 mA circuit. On my successful two-milliamp circuit, the voltage gain is a hair less than 15. The 220K load resistor actually reduces this gain slightly. (As a load, this resistor appears in parallel with the drain resistor.) The 220-kilohm resistor is optional. I recommend it when the preamp output connects to a tube amp without a grid resistor. Without the resistor, gain will increase toward 14.6.

For much more gain, I could have bypassed the source resistor. But I don’t want to do that, because the resistor value is only 75 ohms! The capacitor would be huge, about 1,000 uF for a 20 kHz bottom frequency. (A gain of 15 should be plenty, and the amp will sound better without the capacitor.)

The third way to increase gain is to add another JFET stage right after the first stage. The preamp circuit is extremely power thrifty. Your battery will demonstrate its thanks by long and faithful service.

Diode for source resistor. I've received a suggestion to replace the source resistor with a diode. The idea is that the diode will drop less voltage than a resistor, leading to “increased gain.” Unfortunately a diode drops 0.5 to 0.7 volt. A Schottky diode drops about 0.4 V. In my 2 mA circuit, the 75-ohm resistor drops less than that, only 0.14 V.

Here's the proof...
I_D= (V / [Rd + 1/G + Rs)]= 4.5 / [2200 + (1/0.013) + 75]= 0.0019
V_S= (75 x 0.0019) = 0.1425 V

Math doesn't work. The question about dropping less voltage than a resistor is interesting. Extra gain can be very desirable. But exchanging proper bias for gain is a unfortunate trade. For operation as an analog amplifier, a JFET requires optimum bias. There just aren't diodes for every necessary source voltage drop. Resistors for nearly every drop are plentiful. Suppose that we had a diode with a 0.14 volt drop. It still wouldn't function any better than a resistor! Besides, the source voltage drop is what sets the drain voltage. A reduced source voltage would just reduce the drain voltage.

Source & drain resistors. Likely with 18-volt power, the JFET would develop even more gain. Such an 18-volt preamplifier requires a different drain resistor. The drain resistor would be about 4.7K. The source resistor would vary from device to device. Yet the drain current of the new circuit is the same with the nine-volt circuit. For that reason, a 75-ohm source resistor might still perform just fine. You can check performance by measuring drain voltage with no signal on the gate. On the drain, you should have 4.5 volts to 5.2 volts. (If you can't tell a drain from a source, how about those Cubs?) RG RD RS VDD 3.3 MΩ 4.7KΩ 75 Ω 18 V

The graph above shows how the 2SK370 transconductance (GM) rises with increasing drain current (ID.) Typical GM= 22 mS.

Finding a Source Resistor Value (R_S)

New Source Resistor Value. Suppose that a 75-ohm source resistor doesn't work out. Then you'll need to work out the right source resistor value. When figuring the source resistor, a good starting point is the reciprocal of the expected transconductance. Device curves predict a transconductance of 22 millisiemens. (Siemens are the inverse of ohms.) Then the source resistor might be (1/0.022), or 47 ohms. The gate resistor would remain the same.

Building & Testing

Build the test circuit on a solderless breadboard. If the no-signal drain voltage exceeds nine volts, reduce the source resistor value. If the no-signal drain voltage is less than 9 volts, increase the source resistor value. A typical resistor range is 22 to 82 ohms.

Engineering Applications

Acclaimed designer Nelson Pass is using the 2SK370 device in preamplifiers and headphone amplifiers. Now I know why: Ease of use, low noise and plenty of gain. The low-noise figure is also intriguing. Maybe I'll need more of these than the few I ordered.

Where to Buy Your 2SK370

Where can I find a 2SK370? Try Ebay Electronic Parts and make your best deal. The prices are all over the map. Some vendors are selling one part for $20. Another vendor might sell you five for three bucks! Sometimes you can even achieve free shipping. Rather than auction prices, most of the parts sell with "Buy It Now!" pricing. Air shipments are almost as fast as with domestic orders. Sea shipments can take three weeks or more.

Footnotes

¹ Albert Paul Malvino, PhD, Electronic Principles, Second Edition. (New York: McGraw-Hill Book Company, 1979), 330.

Copyright © 2016 by James T. Hawes. All rights reserved.

•URL: http://www.hawestv.com/amp_projects/fet_preamp/2sk370-fet-1.htm •Webmaster: James T. Hawes
•Revision—November, 2016 •Page design tools: HTML, Notepad & Explorer