Ideas For Bulb Flash Inhibition Circuit

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
A group of us have been making simple mods using a Mag host, an array of 123s, and a high wattage lamp. It's difficult finding just the right bulb for a given array of 123s (or for any voltage source).

Of course we can lower the battery voltage or raise the rated bulb voltage, but that makes things dimmer. We want to walk on the wild side. We have been trying to walk that walk, but we're losing a lot of bulbs.

The SureFire A2 has circuitry to ramp up the voltage relatively slowly so that bulb life is lengthened. I presume it also reduces the risk of the bulb flashing. The circuit also regulates.

What we could use is a simple circuit that addresses only the initial transient problem. Does anyone have any ideas -- maybe something with a transistor and a few resistors and diodes?

Paul

P.S. ( Edit) Recently zmoz was looking for a circuit to cuttoff charging when a certain voltage is reached. Some suggested circuits are found here. It would be nice to find something like what was discussed.
 

Tomas

Banned
Joined
Jun 19, 2002
Messages
2,128
Location
Seattle, WA area
Assuming there's at least some resistance in the overall circuit (other than the bulb itself), wouldn't a fairly large capacitance across the bulb reduce the inrush to the bulb while the capacitor charges? How 'bout a large inductance in series, with a reversed diode acros the switch to help it to survive the kick when the lamp is switched off?

(Obviously the electronicly simulated analogs of these would probably be smaller ... /ubbthreads/images/graemlins/grin.gif )
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
Tomas,

Thanks for your ideas. Yes, analog circuits to "amplify" the values of inductance or capacitance would be neat way to go. I'm hoping to avoid that level of complexity.

An LC circuit would indeed be simple. But it turns out that the size of both the inductor and capacitor would have to be huge, as you have stated. This is because the current is so large -- amps instead of microamps.

Paul
 

Entropy

Enlightened
Joined
Dec 30, 2002
Messages
413
Location
Bridgewater, NJ
[ QUOTE ]
PaulW said:
Tomas,

Thanks for your ideas. Yes, analog circuits to "amplify" the values of inductance or capacitance would be neat way to go. I'm hoping to avoid that level of complexity.

An LC circuit would indeed be simple. But it turns out that the size of both the inductor and capacitor would have to be huge, as you have stated. This is because the current is so large -- amps instead of microamps.

Paul

[/ QUOTE ]
Well, this is by no means the "simple" way to do things, but what you could do is implement a current regulator similar to many of the Luxeon drive circuits.

The main reason for inrush in incandescents is that when the bulb is cold, the resistance is low, and hence the current is much higher than normal until the filament heats up and the resistance increases.

If you use current regulation (Easy but inefficient way - Use an LM317 linear regulator, hard but efficient way is a switching regulator), the current will stay the same and the voltage will gradually ramp up as the resistance increases.

Make sure to have a Zener diode to provide a load when the bulb burns out though! /ubbthreads/images/graemlins/smile.gif
 

Doug Owen

Flashlight Enthusiast
Joined
Jan 30, 2003
Messages
1,992
[ QUOTE ]
PaulW said:

The SureFire A2 has circuitry to ramp up the voltage relatively slowly so that bulb life is lengthened. I presume it also reduces the risk of the bulb flashing. The circuit also regulates.

What we could use is a simple circuit that addresses only the initial transient problem. Does anyone have any ideas -- maybe something with a transistor and a few resistors and diodes?


[/ QUOTE ]

You can probably cobble up a simple "soft start" circuit with a single transistor and a couple of resistors and a capacitor. A high 'beta' (DC current gain) transistor with low Vsat is driven into saturation to light the lamp.

Take say one of the those nifty Zetex NPNs that Mr. Al was using a while back, hook it's emitter to V minus. One side of the lamp to the collector, the other to V plus. From V plus you want a resistor to the base large enough to drive the transistor into saturation (so there's nearly zero Volts emitter to collector, 'full voltage' on the bulb), probably values near 1000 ohms. A largish electrolytic cap (say 100 MFD) across the base and emitter will give a time delay on turn on. You may want to try splitting the resistor into two of half the value in series with the capacitor plus at their junction. A few thousand ohms directly across the capacitor might help in rapid resets of the time delay.

You need to be a bit careful with this sort of thing, make turn on too slow (or have it 'hang up half on') and the transistor is likely to overheat quickly.

Doug Owen
 

PsycoBob[Q2]

Enlightened
Joined
Apr 9, 2002
Messages
518
If you can make a transistor-on circuit that has several requirements:
1. turn on in about half a second.
2. Turn off transistor in a few seconds, after removal of power.
3. Low Loss- both when on and off.

#1 is simple- it has to turn on slow enough to make it safe, but fast enough to avoid annoyance.
#2 is a little more interesting- You want to avoid having the transistor get stuck in the 'on' position for very long, or it'll risk blowing the bulb in a repeat turn on.
#3- Mustn't destroy battery-life, 123's are expensive.

Solution: 5 parts, a transistor, capacitor, and 3 resistors. One resistor to discharge the cap, one to charge it, and one to protect the poor transistor. Now someone poke holes in this, or refine it- your choice. (MOSFET might be an option, too.)
 

Jonathan

Enlightened
Joined
Dec 14, 2001
Messages
565
Location
Portland, OR
The circuit in the Surefire A2 is very likely similar to the 'Willie Hunt LVR-3' circuit, which uses a _very_ inexpensive PIC microcontroller to provide all of the regulation and power control, along with a few external passive components and a MOSFET to actually switch the voltage.

I'd strongly suggest that you look at this sort of circuit. It is not 'simple'; you are using a chip with thousands of transistors. However in our mass production digital is cheap world, this circuit is very cheap and easy to implement, and every component outside of the microcontroller is simple and easy to understand. A PIC based regulator circuit would be less expensive to build than one of the high power lamps that you are blowing.

On the other side of the coin, if you want blog simple, look up NTC Inrush Current Limiters, eg. from Thermometrics at http://www.thermometrics.com/assets/images/cl.pdf

These are temperature sensitive resistors where the resistance drops as they warm up. They start out _cold_ with a high resistance, and limit the current flowing to the lamp. But as resistors they self heat, and as they warm up the resistance drops and more current goes to the lamp. Select the proper combination of battery supply voltage, lamp current rating, and inrush current limiter, and you will get a nice smooth ramp-up. Digikey sells these parts.

-Jon
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
Thank you all for your suggestions. Each was fascinating and gave me food for thought over the past few days. The idea that I think I'll set up and play with is the Zetex NPN circuit which Doug Owen furnished. Doug cleverly described a circuit diagram using only words. It's the one that Tomas hinted at when he suggested an "electronically simulated analog" and which PsycoBob[Q2] called a "transistor-on circuit." It's essentially a low-pass filter used to drive the base of a transistor in the common emitter mode. I guess each person has his own words for the same thing. I am going to experiment with this idea.

Entropy suggested a circuit based on a LM 317 regulator. I may end up going to that, but I'll try simple first. Jonathon suggested ultra simple, the use of a thermistor. But from what I gleaned at the link he gave, the time constant is on the order of a large fraction of a minute, whereas I'm looking for a fraction of a second.

Again, thanks guys. Any additional ideas will also be appreciated.

Paul
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
Wilkey,

Nope. I've let this slide.

I would want to determine the gain needed for the transistor. The circuit that Doug Owen described would have a rise-time of about 100 msec -- about right, I would think. The transistor would need to control a load a few ohms, with the base looking back at 500 to 1000 ohms, thus a gain of a few hundred is required. Hmmm, I probably have some stone-age transistors in a box in the garage that could do that. Ah, I've stumbled onto the problem. I no longer have an oscilloscope or or signal generators to test the thing.

Sorry. That was a lot of rambling that means nothing to anyone else, but I'll leave it here for my own reference.

So, in re-answering your question, yes I've had some further thoughts, but so far no more than that.

I don't think I'll answer the unasked question, because I really don't know the answer right now. Other irons, other fires to keep track of. /ubbthreads/images/graemlins/smile.gif

Paul
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
Paulr,

Take a look at the link that Jonathan posted. The right column of the chart suggests to me that I'd have to wait perhaps a minute for the resistance to reduce to the point that the light would light up. Do you have a different inference from the chart?

Paul
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
Charlie,

Yep, that certainly is a way to do it. When I first hear it, I chuckle a little. And then I say, "Hmmmmm." And then I think of Ginseng's Squash Light . . . and I chuckle again. /ubbthreads/images/graemlins/grin.gif

Paul
 

PsycoBob[Q2]

Enlightened
Joined
Apr 9, 2002
Messages
518
Another thought is a 2-stage switch. First stage runs current through a resistor, second is Direct-Drive.
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
I like that idea, too. And building on that idea there could be a second switch, momentary, with a resistor in parallel and the primary switch in series with both. Then, normal operation would have the resistor in the circuit and the momentary would act as a "boost," shunting out the resistor.

Paul
 

PsycoBob[Q2]

Enlightened
Joined
Apr 9, 2002
Messages
518
Now for the hard part... Putting multiple switches in a aluminum-bodied light. Adding electronics is easy- either up in the bulb end, or by boring out some of the dead-space in the switch body. /ubbthreads/images/graemlins/thinking.gif
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
Yes, it won't be easy. But, thanks to Ginseng's encouragement, I finally bought myself a Dremel. I think I now have most of the tools. I just need the courage and the time.

Paul
 

star882

Enlightened
Joined
Sep 7, 2002
Messages
527
Location
C:\\Program Files\\CPF
"I thought a thermistor was the usual way to do this."
In some switching PSUs, a thermistor is used to reduce surge current.
Otherwise, the primary filter cap(which is virtually a short circuit when discharged) will blow the primary rectifier diodes.
 

PaulW

Flashlight Enthusiast
Joined
Mar 23, 2003
Messages
2,060
Location
Laurel, Maryland
Star882,

Do you know how appropriate a thermistor would be for use in a flashlight? How long does the resistance remain at a high value after powering up? If it's on the order of seconds, or tens of seconds, that's too long for me. Are there thermistors available with time constants of around 100 milliseconds?

Paul
 
Top