30th June 2007

Things done in this week:
Choose Mosfet for just for one pixel
Based on criteria in __http://homepages.which.net/~paul.hills/SpeedControl/Mosfets.html__
1. N-channel
2. Vds(max) I chose it to be bigger than 9v as take short circuit into consideration that the MOSFET has to be able to handle that high voltage (no 9V on farnell also). The mosfet results from farnel based on the
3. Id(max) = 0.3A all have very high Vds(max) which is above 60V for the TO-92 style. I am not very sure how will this affect the circuit design, while they Rd(on) all very high, like 1.8ohm. Id(max) I do not know what is the downside if we choose slightly higher Id(max) than it required (0.3A). Based on the website, the proper one i could find with compatible Vds(max) is 0.4A.
4. Rds(on) we want the standard value which is low at 0.02 ohm so it won't consumes much power when it is turned on
5. Pd (maximum power handling capability of the MOSFET) It depends on Max current that would go through the MOSFET is 0.3A and the Rds(on) of the MOSFET, we should choose the value to be larger than I²R
6. VGS(th) (minimum voltage required between the gate and source terminals to turn the MOSFET on) I chose it to be below 5V as the gate voltage is provided from the PIC

We tested again the MOSFET we ordered from farnel (9802380), although the voltage between drain and source is still 0.3V, but the circuit is supposed to be switched on properly, as the current passing through the heating element is close to 0.26A. Previously we were worried about the relationship between the vds and Id, which also divides the mosfet in the two regions - linear and saturation. And when the current is 0.26A, from the graph, the mosfet we have is in the linear region which mean the current increase with the Vds across the MOSFEt, which means the MOSFET still acts as a resistor but with resistance 0.3V/0.26A it is about 1ohm, while in the datasheet it indicates a resistance of 0.35ohm when it is fully turned on. This means probably the MOSFET is overrated for our circuit. However, something confuses me is, no matter the mosfet is fully on or not, the circuit is still functioning properly (meaning heating element could be heated to change color properly, not largely affected)
And for the overrated problem, for my circuit is 10V*0.3A, and plus a 50% buffer, it will be 4.5W, there is no way for me to find a MOSFET that has such low power. As most with small Id max current, will have very large Vds voltage about 60V at least; or otherwise when the voltage is small, the current will be very high. The best one I could find is 0.4A and 60V, which on resistance is about 3ohm. But another problem I found is, their Rds(on) all very high, up to 1.8ohm, which will draw a lot more current than the MOSFET we are using now which has Rds(on) of 0.35ohm.

Linearility of the temperature sensor:
1. Initial use of the thermistor whose charactericstic s and advantages has been talked about in the past few weekly reports. While it is hard to control the system accurately when you have a non-linear change in the feedback. And a table would be needed in the programming for the uc to look up to.
2. There is method that could parallel the thermistor with a certain value of R that could increase the liearity of the resistance of thermistor group. And there is also linearised thermistor package. While in the end, the op-amp might still needed as we are using 5V to drive for safety purpose.
3. Temperature to digital converter available on Analog Devices, with only three pins, the output could be connected to any digital pin of a the microcontroller. And formula is available for calculating the temperature based on the PWM signal(counting the on time and off time).
We decided to use the digital converter in the end

I have redesigned the schematic diagram of the PIC board as the voltage regulator part will be different if we changed to use the temperature to digital converter.

Things to be done:
1. Use Voltage regulator to generate voltage pulse to test the MOSFET, and the output should also be some square pulse if the input is square wave
2. Oven in power lab that could set the accurate temperature in order to test out the thermal sensors.
3. Print the shirt to test out the conductive yarn for changing color of the shirt.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-Share Alike 2.5 License.