- Power supply:
This soldering station need to be efficient. I will choose switch mode power supply. The circuit will have 2 power inputs: from 24V AC though a rectifier or 19.5VDC from a spare laptop power supply. I prefer the laptop power supply since it will be safer to use and output voltage will be cleaner than I could make it myself.
The Hakko 907 is rated 50W 24V. I measure the resistant of the heater is around 3-4 Ω when cool and 10 Ω when hottest. When, assume maximum power is the rated power 50W then we have current , Voltage supply need to be . When , we have and voltage supply need to be . So we need a voltage of power supply 24VDC.
If we use a 24VDC and source all the current through the iron when it was cool, let see how much current it has to sink: !!! and power dissipated is 8^2*3=192(W), way over rated power and soon the iron will die. Obviously we need to soft start the current somehow.
To supply 5VDC for the ICs I will use a LM2576S-5V fixed from National (now TI). This is a 3A buck DC-DC converter. It has an output fixed at 5V and required only a few external components.
2. Power control of the heater:
I used a MOSFET to switch current on and off through the heater. Consider maximum current required is around 8A, I chose IRF3205 (rated 110A 55V with Rds_on=8mΩ). Actually I could use a logic level MOSFET with max drain current a little bit over 8A, but to quickly turn it on we need more current than a PIC output pin can supply (25mA). We would end up using a MOSFET driver circuit anyway. So why not spend a little more to get the big guy IRF3205.
To switch on and off IRF3205 efficiently we need to provide a voltage 8-10V at its gate. Moreover, we need to supply large current to the gate to quickly charge the gate capacitance. I tried different circuit on a breadboard to see how the MOSFET response:
First, I used the totem pole transistor configuration below:
The HEAT_CTRL signal is TTL signal coming from PIC pin, through 1.5K Rb. When HEAT_CTRL is logic 1, transistor C1815 conduct and current will charge IRF3205 gate through 47Ω Rg. You would think that charging current would be around 12/47=255mA but it’s not. Because maximum collector current we can get out of C1815 is 150mA only according to datasheet.
When I provided a 20kHz PWM pulse with 50% duty circle to HEAT_CTRL, the iron slowly warmed up and got hot. I then measured voltage across MOSFET and it read 0.4V which is not good. The MOSFET got warm. Maybe 150mA is not enough to quickly turn ON the FET. It would be alright if I had a heat sink on the FET and forget about it. But it just doesn’t feel right. We would lose 0.4V over the FET conducting 4A, dissipating 1.6W, and wasting it.
Maybe I will try again next time with 2N2222A in place of C1815. 2N2222A has max collector current around 1A, good enough for my application. But to quickly turn OFF, I need the same current rated PNP BJT in place of A1015. I currently don’t have any PNP that good. Okay, try different configuration then.
The inverted driver
When HEAT_CTRL is logic 1, BJT C1815 conducts and pulse FET gate low to turn it OFF. We also have a problem of not discharge gate quickly enough. When HEAT_CTRL is logic 0, BJT C1815 open and FET gate is charged through Rc 1.2K. Gate charging current is around 10mA which is small. But the FET doesn’t get hot. Voltage drop on it only about 20-30mV. I scratched my head really hard to see how and why? This inverted driver is supposed to be a slow switching driver, not suitable for 20kHz switching.
Something must be wrong here.