JimmyM
Flashlight Enthusiast
I've been posting here and there about an idea I have for a PWM softstarter. So I figured I just consolidate the discussion.
There are plenty of bulbs that can stand higher voltages if the voltage is ramped up instead of just closing a switch. However the choices were pretty limited. You had to buy a regulator that offered soft-starting as a feature. The choices are AWRs Hotdriver and the rare as Hen's teeth PIR by winny. They both have their advantages (actual regulation) and disadvantages (limited power handling and cost).
I really just wanted a way to soft-start a really powreful incan. The hotdriver has ragged edge reliability and the PIR is complex.
So I figured a solution would be a PWM driver that ramps it's duty cycle from 0 to 100% over a period of time. That time needs to be adjustable from almost 0 seconds to about a second. The fact that the MOSFET is either full on or full off, means that the ramp up time isn't limited by the heat dissipation ability of the MOSFET. It will go from fully on to fully off in less than a microsecond.
My design is as follows.
A Micrel MIC1557 oscillator runs at about 1kHz producing a sawtooth wave. The voltage of the sawtooth varies from 1/3 of the operating voltage to 2/3. That signal is sent to a comparator that compares the sawtooth signal to the voltage determined by a voltage divider (2 resistors that go from + to -). If the voltage divider's voltage is 50% of the operating voltage, the comparator will turn on when the sawtooth is below that voltage and off when it's above. if you adjust the voltage divider you can control the percentage of time that the comparator's output is on.
A low voltage at the divider means low duty cycle, high voltage means high duty cycle. Now, remember that whole 1/3-2/3 thing from before? Well, if you adjust the voltage divider so that it's voltage is above 2/3 of the operating voltage, the comparator's output is on all the time.
If you hook that output to a MOSFET you now have a high current/low resistance switch that operates at 1kHz.
A perk of the 1/3-2/3 thing is that if you hook a capacitor to the comparator input, the voltage will ramp up itself when the light is turned on. Well, when you first turn the light in the capacitir starts charging very quickly. That's bad for soft-starting. BUT since the comparator won't even start switching on the MOSFET until the capacitor has charged to 1/3 of the operating voltage, the "ramp up" voltage is less steep, and therefore starts more gently. So instead of ramping up VERY quickly at first and then tapering off almost indefinitely. It's a lot more linear. That is good for soft-starting.
The parts list is as follows.
1 Micrel MIC1557 oscillator (produces the sawtooth wave)
1 Linear Technology LT1716 Comparator (turns sawtooth to square wave)
1 Texas Inst TPS2819 Mosfet driver (ensures the MOSFET switches hard and fast)
1 International Regulator IRLR7843 or IRF1324 MOSFET.
A few resistors and capacitors.
That adds up the ability to control the ramp up time almost indefinitely and cap the duty cycle at something less than 100%. So you can limit the output voltage to a percentage of the input voltage.
Total cost of electronic components is about $10. Believe it or not, the bare etched boards themselves could cost $20 a piece if I were to have them made.
So.... Thoughts?
There are plenty of bulbs that can stand higher voltages if the voltage is ramped up instead of just closing a switch. However the choices were pretty limited. You had to buy a regulator that offered soft-starting as a feature. The choices are AWRs Hotdriver and the rare as Hen's teeth PIR by winny. They both have their advantages (actual regulation) and disadvantages (limited power handling and cost).
I really just wanted a way to soft-start a really powreful incan. The hotdriver has ragged edge reliability and the PIR is complex.
So I figured a solution would be a PWM driver that ramps it's duty cycle from 0 to 100% over a period of time. That time needs to be adjustable from almost 0 seconds to about a second. The fact that the MOSFET is either full on or full off, means that the ramp up time isn't limited by the heat dissipation ability of the MOSFET. It will go from fully on to fully off in less than a microsecond.
My design is as follows.
A Micrel MIC1557 oscillator runs at about 1kHz producing a sawtooth wave. The voltage of the sawtooth varies from 1/3 of the operating voltage to 2/3. That signal is sent to a comparator that compares the sawtooth signal to the voltage determined by a voltage divider (2 resistors that go from + to -). If the voltage divider's voltage is 50% of the operating voltage, the comparator will turn on when the sawtooth is below that voltage and off when it's above. if you adjust the voltage divider you can control the percentage of time that the comparator's output is on.
A low voltage at the divider means low duty cycle, high voltage means high duty cycle. Now, remember that whole 1/3-2/3 thing from before? Well, if you adjust the voltage divider so that it's voltage is above 2/3 of the operating voltage, the comparator's output is on all the time.
If you hook that output to a MOSFET you now have a high current/low resistance switch that operates at 1kHz.
A perk of the 1/3-2/3 thing is that if you hook a capacitor to the comparator input, the voltage will ramp up itself when the light is turned on. Well, when you first turn the light in the capacitir starts charging very quickly. That's bad for soft-starting. BUT since the comparator won't even start switching on the MOSFET until the capacitor has charged to 1/3 of the operating voltage, the "ramp up" voltage is less steep, and therefore starts more gently. So instead of ramping up VERY quickly at first and then tapering off almost indefinitely. It's a lot more linear. That is good for soft-starting.
The parts list is as follows.
1 Micrel MIC1557 oscillator (produces the sawtooth wave)
1 Linear Technology LT1716 Comparator (turns sawtooth to square wave)
1 Texas Inst TPS2819 Mosfet driver (ensures the MOSFET switches hard and fast)
1 International Regulator IRLR7843 or IRF1324 MOSFET.
A few resistors and capacitors.
That adds up the ability to control the ramp up time almost indefinitely and cap the duty cycle at something less than 100%. So you can limit the output voltage to a percentage of the input voltage.
Total cost of electronic components is about $10. Believe it or not, the bare etched boards themselves could cost $20 a piece if I were to have them made.
So.... Thoughts?
