My new version of the VBE multiplier requires temperature sensing using an NTC thermistor with a resistance at 25C of 465Ω and a temperature coefficient of -1.48Ω/K. My first impression is this may be impossible to find, which means, additional circuitry is needed to emulate this behaviour using available NTC thermistors. The major problem is the operating voltages are below 1V. A BJT transistor used for emulation will not solve the issue as BJTs exhibit a constant resistance below saturation. A transistor is needed which should work under 1V and provide a resistance that is dependent on an input signal.
Aternatively, a network of different NTC termistors may be created to emulate the required behaviour. I am looking at this solution as this provides an average of temperature and an effective resistance which should be more realistic.
Aternatively, a network of different NTC termistors may be created to emulate the required behaviour. I am looking at this solution as this provides an average of temperature and an effective resistance which should be more realistic.
Last edited:
Another possible solution is to use an optocoupler and connect the part containing the photo-sensitive transistor between the VBE multiplier's base and its positive end. This should allow the transistor to act as a parallel resistor. Input into the photocoupler's diode can be supplied through a separate circuit which would be isolated from the VBE multiplier. Isolation gives the possibility to using various levels of complexity to emulate the required thermal behaviour. Simulating this on LTSpice shows very promising BIAS control with increasing temperature.
Yet another solution is to connect the phototransistor between the emitter and base of the VBE multiplier. The latter would need to provide an underbias as current flowing through the phototransistor would lower the effective resistance increasing the voltage across the VBE multiplier. In this way, the phototransistor will have a voltage of about 0.65V across it. Since, Vbe is almost constant, the linearity of the VBE multiplier is not degraded. Also in this way, the pot can be transferred to the isolated control circuit which may contain as many temperature sensing devices as necessary.
I will continue exploring this last solution as it gives much more freedom in how the VBE multiplier is controlled.
I will continue exploring this last solution as it gives much more freedom in how the VBE multiplier is controlled.
In my last simulation the VBE multiplier's resistance from the base to the emitter was removed and replaced by a transistor. The latter is integrated in an optocoupler. This modification is possible for the reason that Veb of the multiplier is almost independent of operating currents, which implies, that the multiplier is still linear. The multiplier's set voltage depends heavily on the current flowing into the resistance across the base and collector. The algebraic relationship should be of the form:
Control is achieved by supplying a biasing current into the optocoupler's input such that a constant current flows through the phototransistor. This current controls the amplifier's BIAS at 25C. Compensation for temperature deviations from 25C is done by modifying the control current so that the bias remains almost constant irrespective of temperature changes. Under LTSpice the control equation is:
Code:
Vmult = 0.65 + i*Rbc
Where: Vmult is the multiplier's voltage,
i is the current flowing into the base collector resistance,
Rbc is the value of the resistance across the base and collector.Control is achieved by supplying a biasing current into the optocoupler's input such that a constant current flows through the phototransistor. This current controls the amplifier's BIAS at 25C. Compensation for temperature deviations from 25C is done by modifying the control current so that the bias remains almost constant irrespective of temperature changes. Under LTSpice the control equation is:
Code:
control I = 3.06u - 19n*(atemp - 25)
Where atemp is the temperature,
I the equivalent controlling current into the phototransistor's base.A thermistor by Vishay (NTCALUG01A103FL):
B25/85 = 3435
R25 = 10k
The resistance at 25C is 10k and at 85C it is ~1.4k. The equation governing the resistance is:
R = 10000*exp[3435*(1/T - 1/298.15)]
The above is non-linear, which implies, the control circuit has to provide a way to linearise it by making the sensing cicuit with a gain dependent on either current or voltage.
B25/85 = 3435
R25 = 10k
The resistance at 25C is 10k and at 85C it is ~1.4k. The equation governing the resistance is:
R = 10000*exp[3435*(1/T - 1/298.15)]
The above is non-linear, which implies, the control circuit has to provide a way to linearise it by making the sensing cicuit with a gain dependent on either current or voltage.