EvilLithiumMan
Enlightened
(As it turns out, I've just reinvented the wheel. But my wheel isn't nearly as round or a pretty as the others, so here it is)
My original goal was to see if I could a design cicuit using a power mosfet to replace the NTC (thermister) that many Mag623 users employ to prevent their bulbs from flashing. Although inexpensive and easy to hook up, NTC's run quite hot. They are also affected by ambient temperature, require a cool down period and generally have a tolerance +/-20%. I was hoping a simple FET circuit would be a practical solution. And by simple, I mean 6 or 7 discreet components, readily available, less than $20, and easy to tailor to one's specific needs.
This is my initial attempt, a FET with a basic RC network to control the gate:
It has a terribly large time constant and for good cause: I don't have a scope. So I needed a prolonged, clearly visible soft start to verify functionality. Well, I got what I wanted. The startup delay was nearly 3 seconds, with a soft start period of approx. 1/2 second. Here is the breadboard: (It's nasty looking, you may want to wear safety googles)
Unfortunately, the turnoff cycle is nothing short of comical. 130 seconds to shutoff! From 0 to 110 seconds, the lamp appears fully lit. From 110 to 120 seconds, the lamp slowly turns off. But it takes another 10 seconds before there is no visible glow from the filament. I had to laugh.
My next step was to Google search CPF to determine how short of a soft start period would work. Perhaps if 20ms was adequate, a turnoff period of 10-12 seconds wouldn't seem so bad. But in my searching, I came across this thread (and others as well):
http://www.candlepowerforums.com/vb/showthread.php?t=100286&post#17 (post #17)
Clearly, this has all been done before. But they too had to confront the long shutoff period, which meant more components and complexity. To keep my circuit simple, I had to find a simple solution to discharging the gate capacitor. I then recalled I had an assorted collection of small signal PCB relays. Would a relay work? Would a relay fit? Turns out a relay works just fine! A Magnecraft W172DIP-7, plus a resistor, does the trick. And despite the length of the relay, there is no interference with the Mag's cammed reflector.
I now wonder if I can fit all of this stuff into the ubiquitous KIU switch. To that end, I've ordered a couple of KIU kits and a couple of 64623 bulbs. I don't yet have a tri-bore host or battery pack to complete a 623, and I'm not sure I ever will. While I don't doubt for a second that it must be a Hell of a lot of fun to have an armor piercing 3000-5000 lumens in one's hand, I'm not convinced such a monster is for me. For now, I'm content with the challenge of replacing a NTC with a "simple" FET solution.
My original goal was to see if I could a design cicuit using a power mosfet to replace the NTC (thermister) that many Mag623 users employ to prevent their bulbs from flashing. Although inexpensive and easy to hook up, NTC's run quite hot. They are also affected by ambient temperature, require a cool down period and generally have a tolerance +/-20%. I was hoping a simple FET circuit would be a practical solution. And by simple, I mean 6 or 7 discreet components, readily available, less than $20, and easy to tailor to one's specific needs.
This is my initial attempt, a FET with a basic RC network to control the gate:
It has a terribly large time constant and for good cause: I don't have a scope. So I needed a prolonged, clearly visible soft start to verify functionality. Well, I got what I wanted. The startup delay was nearly 3 seconds, with a soft start period of approx. 1/2 second. Here is the breadboard: (It's nasty looking, you may want to wear safety googles)
Unfortunately, the turnoff cycle is nothing short of comical. 130 seconds to shutoff! From 0 to 110 seconds, the lamp appears fully lit. From 110 to 120 seconds, the lamp slowly turns off. But it takes another 10 seconds before there is no visible glow from the filament. I had to laugh.
My next step was to Google search CPF to determine how short of a soft start period would work. Perhaps if 20ms was adequate, a turnoff period of 10-12 seconds wouldn't seem so bad. But in my searching, I came across this thread (and others as well):
http://www.candlepowerforums.com/vb/showthread.php?t=100286&post#17 (post #17)
Clearly, this has all been done before. But they too had to confront the long shutoff period, which meant more components and complexity. To keep my circuit simple, I had to find a simple solution to discharging the gate capacitor. I then recalled I had an assorted collection of small signal PCB relays. Would a relay work? Would a relay fit? Turns out a relay works just fine! A Magnecraft W172DIP-7, plus a resistor, does the trick. And despite the length of the relay, there is no interference with the Mag's cammed reflector.
I now wonder if I can fit all of this stuff into the ubiquitous KIU switch. To that end, I've ordered a couple of KIU kits and a couple of 64623 bulbs. I don't yet have a tri-bore host or battery pack to complete a 623, and I'm not sure I ever will. While I don't doubt for a second that it must be a Hell of a lot of fun to have an armor piercing 3000-5000 lumens in one's hand, I'm not convinced such a monster is for me. For now, I'm content with the challenge of replacing a NTC with a "simple" FET solution.
Last edited:
