Illum
Flashaholic
EDIT: I think this might be best in homemade and modified instead
I recently dusted off my 2D ROP to check if my ROP low lamp was under the spring, upon disassembly a stock 2D krypton fell to the floor, I picked it up, grinned at it, and wonder how many of its brothers is sitting in my "scrap" pile unused. Then I looked at my recent beta driver using a Zetex driver and I decided to build my own 2 cell "Lambda SMJLED" drop-in.to see how hard it could actually be. :thinking:
Bare lamps, the typical PR based 2 cell krypton lamps you commonly find in a 2D cell flashlight. Warp some masking tape around the globe and make sure you press it snug, it'll save you from a mess of glass shards when you pop it.
Heat the solder iron early, the little green light controls my workbench's overhead fans, see picture 26. These fans keep the air flowing downward, sending the solder smoke traveling horizontally, not at me or at my nose. The little Nidec gamma 27 blowfan [controlled by the toggle switch on the side of the green light] is used for two things, keeping things cool when soldering and curing glue by dehydration
5x120mm Fans salvaged from a PC tower keeps the air moving along, I have a 110V household fan blowing air out the window behind me so the air is one way only.
Make sure you have good lighting! This is taken with the florescent off showing the 3x LuxV driven by a 700ma xitanium.
Place the lamp globe down in a vice, crank it down slowly until it pops, then loosen it, turn the lamp 90 degrees, repeat until you made full circle, by this time the globe should fall off my itself.
This is what it should look like, watch for those sharp edges, remove the glass fragments with some needle-nose pliers.
Use a dremel to take out the potting after you remove the glass. Chuck it in a drill press vice, snug will do, do not overtighten. I'm not sure what tool this is, but its in a standard dremel kit and resembles a barrel cactus, use it on the lowest rpm or you'll have dust blowing everywhere
Extract the center copper contact by heating the bottom contact pad, then using pliers stick it back in the tail end and pull up and down as you remove the solder iron for it to solidify, this will create a hole for your driver's input leads.
This is what you should see after you get the potting out, remember we are not trying to grind it out; we're letting it fall out by itself. The potting is brittle to vibration, take advantage of it.
A 3mm dome Piranha I had kicking around, here it is standing beside the prototype driver [Zetex Zxsc380] and the can it will sit in…soon
Measure the interior dimensions in respect to perf board holes, crop to fit, file it to fit if you have to:
Still a bit tall, more filing is needed. NOTE: do not file it so its flush with the can, it will make it more difficult to work with, leave about 1mm sticking out.
What the end design will resemble
Solder board in hand; let's go to work on the electronics. The inductor shown here is Jameco pt: 208194, 100uH epoxy conformal choke, I tried many epoxy inductors I had available, only this size will fit the can in this fashion. Fit in the inductor along the long side of the PCB. Zetex drivers are in SOT-23, so they do not work well with perf boards. Take advantage of the leads on the inductor. Not only will they provide an effective point of contact but also improve connection and mechanical integrity.
Since the broad side of the driver only touch base with 3 holes instead of 4, compensate it when forming the leads. Pull the leads up when your done
Before soldering, after soldering
It is important to form a notch close [but not over] the center hole, the drivers pins are just outside the hole when placed over it. There the pin will be soldered to the lead. The horizontal lead will go through the bottom of the PR can after you're done.
For the Gnd pin of the driver, I grabbed a fried resistor and used its leads. Bend it in this fashion and solder it in place. NOTE: place a solder bridge across the two contacts, bypassing the resistor so to speak. DO NOT cut off the resistor, you will need it later.
Mousers order numbers on those Zetex ZXSC380s, they are about 60 cents each and comes on a tiny tape roll.
Working with SMTs is important to have tweezers, both for handling them and as "clip-on" heatsinks while you solder. The driver is on the magnifying glass
Fit the driver over the contacts [now dabbed with solder], manipulate the contacts with your tweezers until they are optimum spaced. The "third hand" magnifier is used here to assist the picture taker, NOT for soldering as it wobbles, use the resistor as a handle instead. the resistor also acts a big Gnd contact for tests.
Test your circuit to verify that its working correctly. The PS is set to 1.5V, LED Vin is 3.5V. Yes, the T1-3/4 was from an earlier project that is somehow encased in wood using a spare cree star as a head plate. I can't recall what I used it for.
Insert another lead in to the center hole on the output end and form it in this fashion and solder it in place, this will be the LED+ pin
Solder it to the inductor, cut the resistor off. It doesn't matter if the contacts are protruding, it's a ground line that will eventually be secured to the PR can anyway.
At this point the SOT-23 is electrically connected but mechanically floating. After the circuit is tested drop a drop of wood glue on it, let it flow around on the PCB, prod it around with tweezers if necessary. NOTE: Test the circuit again to verify everything is in order before you "pot" it in place.
This is where the blowfan comes in handy, keeping the fan on the glue dehydrates it very fast and the surface will cure in about a minute.
When it's partially cured, test fit it in the PR can. You'll have to do a little shoving for the inductor to go in, test it after it's in place. Orient the board so that the ground matches the cut-out on the rim of the PR can. Test the assembly again after soldering the input
Form the leads as shown, rest the LED+ on the PCB, see that space between the lead and the can? Aren't you glad you didn't file it flush now?
Form a bent in the negative line and solder it in place, the PR base itself does not contact solder well, I dunno if its steel or…but anyway, it'll make the desired electrical connect, but its very brittle to mechanical leverage, its best to finish forming the lead before soldering.
You should end up with something like this, test the assembly this time by clipping the input Gnd on the can and the led to the top pins. If it works, you are now ready for potting.
Apply a liberal amount of wood glue down its cavity in small drops; take care not to apply glue on the top leads, which has been trimmed to fit. By the time I got here, its past midnight I went ahead to grab a beer and call it a day.
It doesn't look pretty, as this is trial 1, but so far it appears to work properly so that's a definite plus.
The underlying scheme in the driver is the utilization of the least components and those components do not produce heat.
By Zetex ZXSC380's datasheet a 100uH inductor will output 70ma at Vin = 3V, which is very hard on the LEDs, but the good thing is I can maintain 20ma with input of 1.5V eventually dropping go dead at Vdo = 0.8V.
Instead of running this "drop-in" with fresh batteries, I plan to run it on nearly dead Alkalines or CR123As, and keeping the LED soldered in place instead of potting it allows me to swap out the LED if it burns out prematurely.
the total cost of the build is
$0.60 [driver]
$0.011 [inductor, comes in pk of 10]
$1.69 [white Piranha 3mm dome]
2D krypton lamp [free]
A small piece of PCB [essentially free]
About 3 hours of DIY plan, prep, build, and its not done yet:shrug:
I'll start on Part II tomorrow, for now I'm going to sleep
I recently dusted off my 2D ROP to check if my ROP low lamp was under the spring, upon disassembly a stock 2D krypton fell to the floor, I picked it up, grinned at it, and wonder how many of its brothers is sitting in my "scrap" pile unused. Then I looked at my recent beta driver using a Zetex driver and I decided to build my own 2 cell "Lambda SMJLED" drop-in.to see how hard it could actually be. :thinking:
Bare lamps, the typical PR based 2 cell krypton lamps you commonly find in a 2D cell flashlight. Warp some masking tape around the globe and make sure you press it snug, it'll save you from a mess of glass shards when you pop it.
Heat the solder iron early, the little green light controls my workbench's overhead fans, see picture 26. These fans keep the air flowing downward, sending the solder smoke traveling horizontally, not at me or at my nose. The little Nidec gamma 27 blowfan [controlled by the toggle switch on the side of the green light] is used for two things, keeping things cool when soldering and curing glue by dehydration
5x120mm Fans salvaged from a PC tower keeps the air moving along, I have a 110V household fan blowing air out the window behind me so the air is one way only.
Make sure you have good lighting! This is taken with the florescent off showing the 3x LuxV driven by a 700ma xitanium.
Place the lamp globe down in a vice, crank it down slowly until it pops, then loosen it, turn the lamp 90 degrees, repeat until you made full circle, by this time the globe should fall off my itself.
This is what it should look like, watch for those sharp edges, remove the glass fragments with some needle-nose pliers.
Use a dremel to take out the potting after you remove the glass. Chuck it in a drill press vice, snug will do, do not overtighten. I'm not sure what tool this is, but its in a standard dremel kit and resembles a barrel cactus, use it on the lowest rpm or you'll have dust blowing everywhere
Extract the center copper contact by heating the bottom contact pad, then using pliers stick it back in the tail end and pull up and down as you remove the solder iron for it to solidify, this will create a hole for your driver's input leads.
This is what you should see after you get the potting out, remember we are not trying to grind it out; we're letting it fall out by itself. The potting is brittle to vibration, take advantage of it.
A 3mm dome Piranha I had kicking around, here it is standing beside the prototype driver [Zetex Zxsc380] and the can it will sit in…soon
Measure the interior dimensions in respect to perf board holes, crop to fit, file it to fit if you have to:
Still a bit tall, more filing is needed. NOTE: do not file it so its flush with the can, it will make it more difficult to work with, leave about 1mm sticking out.
What the end design will resemble
Solder board in hand; let's go to work on the electronics. The inductor shown here is Jameco pt: 208194, 100uH epoxy conformal choke, I tried many epoxy inductors I had available, only this size will fit the can in this fashion. Fit in the inductor along the long side of the PCB. Zetex drivers are in SOT-23, so they do not work well with perf boards. Take advantage of the leads on the inductor. Not only will they provide an effective point of contact but also improve connection and mechanical integrity.
Since the broad side of the driver only touch base with 3 holes instead of 4, compensate it when forming the leads. Pull the leads up when your done
Before soldering, after soldering
It is important to form a notch close [but not over] the center hole, the drivers pins are just outside the hole when placed over it. There the pin will be soldered to the lead. The horizontal lead will go through the bottom of the PR can after you're done.
For the Gnd pin of the driver, I grabbed a fried resistor and used its leads. Bend it in this fashion and solder it in place. NOTE: place a solder bridge across the two contacts, bypassing the resistor so to speak. DO NOT cut off the resistor, you will need it later.
Mousers order numbers on those Zetex ZXSC380s, they are about 60 cents each and comes on a tiny tape roll.
Working with SMTs is important to have tweezers, both for handling them and as "clip-on" heatsinks while you solder. The driver is on the magnifying glass
Fit the driver over the contacts [now dabbed with solder], manipulate the contacts with your tweezers until they are optimum spaced. The "third hand" magnifier is used here to assist the picture taker, NOT for soldering as it wobbles, use the resistor as a handle instead. the resistor also acts a big Gnd contact for tests.
Test your circuit to verify that its working correctly. The PS is set to 1.5V, LED Vin is 3.5V. Yes, the T1-3/4 was from an earlier project that is somehow encased in wood using a spare cree star as a head plate. I can't recall what I used it for.
Insert another lead in to the center hole on the output end and form it in this fashion and solder it in place, this will be the LED+ pin
Solder it to the inductor, cut the resistor off. It doesn't matter if the contacts are protruding, it's a ground line that will eventually be secured to the PR can anyway.
At this point the SOT-23 is electrically connected but mechanically floating. After the circuit is tested drop a drop of wood glue on it, let it flow around on the PCB, prod it around with tweezers if necessary. NOTE: Test the circuit again to verify everything is in order before you "pot" it in place.
This is where the blowfan comes in handy, keeping the fan on the glue dehydrates it very fast and the surface will cure in about a minute.
When it's partially cured, test fit it in the PR can. You'll have to do a little shoving for the inductor to go in, test it after it's in place. Orient the board so that the ground matches the cut-out on the rim of the PR can. Test the assembly again after soldering the input
Form the leads as shown, rest the LED+ on the PCB, see that space between the lead and the can? Aren't you glad you didn't file it flush now?
Form a bent in the negative line and solder it in place, the PR base itself does not contact solder well, I dunno if its steel or…but anyway, it'll make the desired electrical connect, but its very brittle to mechanical leverage, its best to finish forming the lead before soldering.
You should end up with something like this, test the assembly this time by clipping the input Gnd on the can and the led to the top pins. If it works, you are now ready for potting.
Apply a liberal amount of wood glue down its cavity in small drops; take care not to apply glue on the top leads, which has been trimmed to fit. By the time I got here, its past midnight I went ahead to grab a beer and call it a day.
It doesn't look pretty, as this is trial 1, but so far it appears to work properly so that's a definite plus.
The underlying scheme in the driver is the utilization of the least components and those components do not produce heat.
By Zetex ZXSC380's datasheet a 100uH inductor will output 70ma at Vin = 3V, which is very hard on the LEDs, but the good thing is I can maintain 20ma with input of 1.5V eventually dropping go dead at Vdo = 0.8V.
Instead of running this "drop-in" with fresh batteries, I plan to run it on nearly dead Alkalines or CR123As, and keeping the LED soldered in place instead of potting it allows me to swap out the LED if it burns out prematurely.
the total cost of the build is
$0.60 [driver]
$0.011 [inductor, comes in pk of 10]
$1.69 [white Piranha 3mm dome]
2D krypton lamp [free]
A small piece of PCB [essentially free]
About 3 hours of DIY plan, prep, build, and its not done yet:shrug:
I'll start on Part II tomorrow, for now I'm going to sleep
Last edited:
