Halogen lamp protector
Halogen lamps, particularly high wattage ones, tend to draw very high currents when they are cold because they then have a very low resistance: of the order of 0.1 Ω or lower. If such a high is operated from a 24 V battery, a switch -on peak current of over 200 A may flow. This is highly detrimental to the lifespan of the light, which after only a few switch-ones may give up the ghost. That costly situation may be prevented by gradually building up the power supplied to the lamp. Since the operation is from a d.c. source, the only practical way of so doing by pulse-width modulation. With that technique, the voltage to the lamp is switched from zero with increasing pulse width, while the current is smoothed by a coil, so that its average level increases gradually. Switching is carried out with two MOSFETs Type BUZ11. This type is characterized by an extremely low channel resistance, which is typically 0.003 Ω for a gate-source voltage of 15 B; moreover, the device can handle continuous of a current of up to 30 A and pulsed ones of up to 120 A. By connecting the two in parallel, the current is split two ways: not exactly 50/50, of course, but near enough to ensure that on switch-on the devices remain within their limits.
The control circuit for the MOSFETs. T2 and T3 provide nothing new. A regulator, 1Ci, ensures that the supply to the circuit cannot rise too high: it is set for about 18.5 V. The battery voltage cannot be used directly, because, among others, the gate-source voltage of the MOSFETs must not be higher than 20 V.
A sort of triangular waveform is generated by Schmitt trigger IC2;1; R2 and R4 ensure that the operating point o1 the opamp is half the supply voltage. I3ecause of the be feedback via R6-05, the output waveform is rectangular. However, the voltage across C5 is an exponential waveform. that is more or less triangular. That voltage is compared by 1C2b with the terminal potential of C6, which after switch-on, rises gradually.
As long as the voltage across C5 is lower than at the non-inverting input of IC2b, the output of opamp will be high. As soon as the inverting input reaches a potential higher than that at the +ve input, the output changes state, which is accelerated by the positive feedback via R10.
The higher the terminal voltage of C6, the longer the output of 1C2b will remain high. Eventually, it will reach a value higher than the maximum voltage across C5: pin 7 of IC2b, is then permanently high. The lamp is then no longer switched on and off but remains on.
The Type CA3240 dual opamp in the IC2 position has the advantage that its output can become almost zero, but the disadvantage that it cannot sink relatively large currents. Because of that, T1 ensures that the gate capacitances of T2 and T3, together 2-10 nF, are discharged rapidly. Those gate capacitances are charged again (that is, T2 and T3 are driven into conduction) by IC2b via D2. An additional driver is not needed because of the CA3240 can source enough current at high levels. Then the FEls are on, their gate potential I.; about 16 V.
When the lamp is switched off, C6 is discharged immediately, so that the circuit is ready at once to switch it on again. To ensure that on switch-off the induced potential across L1 does not rise above the maximum level of the drain-source voltage (50 V). the coil is shunted by D3. This needs to be a fast type (25 ns or b•rter) that can handle currents of up to 30 A.
Resistors R11 and R1, provide the lamp with some voltage before the MOSFETs switch on.
The speed at which the circuit switches the lamp on is preset with P1: normally, it will be sufficient to set this to the center of its travel.
It is advisable to mount the FETs on a heat sink, although their dissipation is of the order of 1.6 W only.
Capacitors Ci and C2 must be able to handle high-frequency puke currents of up to 30 A.
Inductor L1 must ensure that the lamp current does not exceed a predetermined value: the larger the inductance. the lower the maximum level of the current. However, the physical dimensions of the coil must be acceptable. In the prototype, the maximum lamp current was set at 30 A. At a switching frequency of 7 kHz, an inductance of 30 μH is sufficient. Moreover, to avoid saturation problems. the coil is an air-cored one.
It is made by winding 45 turns of 1.5 mm (1/6 in) dia. enameled copper wire in three layers on a 24 mm (15/16 in) dia. round former. During the winding. apply some glue from time to time to the turns.
The current drawn by the circuit is primarily that through the lamp: with a 250 W lamp and a 24 V battery. the current is some 10 A.