17th June 2007

Mosfet being tested out. A big resistor is needed between the gate of the MOSFET and the ground as MOSFET can be considered as a small capacitor, and without the resistor the MOSFET will charge also and store the energy within itself.

While we checked the datasheet of the MOSFET we bought, which actually shows the current 0.26A we used to drive one heating element, (with gate voltage 5V and Vds roughly 0.3V) will not be able to make the MOSFET work in the saturation region, because although Vgs>Vth, but Vds<Vgs-Vth. However, we should use it in saturation region because we want to stable the current even there is any change in the voltage (when the circuit it not stable), while in linear region, with the change of Vds, the Id will also change. As the current we supply is very low, it is difficult for the MOSFET to work in saturation region, at least with the MOSFET we bought.

Power circuit problem: the total current we use to drive the whole shirt is very high, up to 9A which won't be able to supplied by any 9V battery, not even the power generator in the lab. We consider about the total power for the circuit, and think we could raise the voltage, which will cause the current to drop. At the same time, the new problem would be the 24V will be put directly across one heating element (37ohm), and the current will be 0.8A across that single element, which might melt the conductive yarn due to heat generated. Although in the datasheet provided by the conductive yarn, the highest temperature is 199°C, it is difficult for us to convert back to the heat of the conductive yarn due to lack of information. Now, it will be a problem for us to know how can we drive a circuit properly with large power required as low voltage will not be able to sustain while high voltage will burn some element because of the parallel connection.

Thermistor Problem:

Maybe, the we could use the difference of the voltages for both the thermistor and the resistor as the input to the op-amp, and then have unity gain at the output part. And because of the output is the difference between the two voltages, the difference will be twice of that when it is only the feedback from the thermistor alone.

Actually there is something we need to test for the thermistor is to test the starting voltage as when different people using it, the starting (room temperature) might be different.

But for the op-amp working in the differential mode, we have to take note when the case it becomes common input rejection mode..

Re-examine of the thermistor characteristics in the book Practical Design Techniques for Power and Thermal Management:
A 10kohm NTC thermistor, with a typical 25ºC temperature coefficient of -44,000ppm/ºC (the coefficient under that temperature). The temperature coefficient does not decrease linearly with increasing temperature. Because of the unlinear change in the coefficient, it is difficult for us to calculate the resistance of the thermistor under different temperatures. To paralleling the thermistor with another fixed resistance resistor would significantly increase the linearility of the resistance change. However, this would have a decrease in the resistance of the combination set and currently I do not think it is necessary for us to that now. And also we lack of the device to display the system temperature, we still have to use the previous method to estimate the values for the thermistor voltage change, and if later on we found this not able for us to change the color accurately, we might start to think of the paralleling it with a fixed resistor.

As thermistor temperature increases, its resistance and sensitivity to temperature change also decreases. This means the number of ohms per bit decreases and the number of degrees per A/D step increases.

Using of op-Amp:

Single-ended op-amp will be needed for the simplification of having the negative voltage supply.

Proposed solution: use 5V to drive the circuit. inverting-input will be connected to the point between the thermistor and the resistor (voltage across thermistor), and connect Vss/2 to the non-inverting input, the positive voltage supply would have Vss, and with negative power supply side connecting to the ground as these two are referencing from the non-inverting side.

Pull-up and pull-down resistor is for the purpose of input to the chip instead of output for the chip, and also said to limit the output pins (http://www.seattlerobotics.org/encoder/mar97/basics.html)

page revision: 0, last edited: 01 Jul 2007 04:21