I built the temperature compensated LM334 current source (Figure 1 below), configured for 5.5 milliamps: R1 = 27R and R2 = 270R and D1 = 1N457A. When I put it on the curve tracer to sweep out its I-vs-V curve, I got quite a pleasant surprise: the LM334's "knee", where it ceases acting at a current source, is at a relatively low voltage: 2 volts. See Figure 2.
Linear Technology's datasheet suggests adding a JFET with large |Vpinchoff| as a cascode (Figure 3). I tried this, using the J105 device (Figure 4). It only moved the "knee" upwards about 200 millivolts or so, maybe to 2.2 volts: Figure 5.
I was pleased.
diyAudio members who have created an LTSPICE model of the LM334 and the J105 and the 1N457A, may wish to simulate these two I-vs-V curves and see whether measured data matches simulation results.
_
Linear Technology's datasheet suggests adding a JFET with large |Vpinchoff| as a cascode (Figure 3). I tried this, using the J105 device (Figure 4). It only moved the "knee" upwards about 200 millivolts or so, maybe to 2.2 volts: Figure 5.
I was pleased.
diyAudio members who have created an LTSPICE model of the LM334 and the J105 and the 1N457A, may wish to simulate these two I-vs-V curves and see whether measured data matches simulation results.
_