wquiles
Flashaholic
I just finished building a custom light for an overseas forum member, and I wanted to share how I built it. It is my first multi emiter build, but I think it came out all right 
Instead of starting with the beginning like I always do, this time I am starting with the end. This is the completed light - it is a custom tri-bored host, with custom grooved and fluted head, with a custom SS bezel, 4x P4 emiters, Al heatsink from the Shoppe, 4x 20mm reflectors from the Shoppe, ultra clear glass, MaxFlex boost converter, FM custom 3x 17670 LiIon battery holder, and custom body ring on tailcap:
and at about 1 Amp or so, it is "bright", and the head does get hot/toasty (this is a little over-exposed, but gives you the idea of how nice the beam is):
OK, so here is the journey ... please note that this is not the only way to make/put this together, and I welcome all ideas/suggestions/tips on this thread :twothumbs
First, given the universal heatsink, there are no locating points for the actual emiters:
Luckily, the reflectors touch each other while in position, and they also lightly touch the scalloped cuts in the heatsink. Also, the reflectors don't touch the emiters, but rest on the scallops, while being kept in place (no rattles) by the glass (more on this later on). So basically, I used the mechanical alignment of the reflectors to line up the emiters.
Of course, since the P4's have to isolated from the heatsink and since they will move/flow while the two-part thermal epoxy sets, how do I align the individual emiters once the reflectors are aligned? Drill holes on a spare plastic lens:
then, using a soft plastic tip, gently align each emiter through the reflector (tip learned here in the forums by no other than McGizmo himself - thanks Don!):
Once aligned, proceed to serially solder all LED's (the MaxFlex is a boost driver, so I am using it to drive 4 LED's in series, with a battery source of a lower voltage, in this case 3x LiIon cells). I am using a piece of tape to help hold each wire while I solder it:
then test to make sure it works:
I then added two-part clear epoxy to all emiters/wires to make sure nothing moves/comes loose, even if accidentally dropped:
I then move to my lathe, and cut the switch flush on both sides, to remove the tower and to allow better electrical contact with flat cells (thanks Mac for the tip on using the lathe for this!):
To provide a heatsink path to the MaxFlex, I used the lathe again to fabricate a small Al cylinder to transfer heat from the MaxFlex to the heatsink:
Here I already soldered the LED wired to the MaxFlex driver, and you can see the small cylinder here already glued to the inside of the heatsink (two-part thermal epoxy again), while I am testing the driver with a bench power supply:
I then applied two-part thermal epoxy between the cylinder and the driver, and used two-part clear epoxy to fix the driver to the heatsink. An observant person will also note the "excess" clear epoxy around the wires - that is on purpose - I will explain further below why:
I then moved to work on the switch. The OEM switch has to be converted to a momentary only switch, in order to be used with the MaxFlex driver (and most drivers from TaskLED www.taskled.com). I have already done a detailed photo-by-photo DIY on the switch mod, so I am only showing some highlights this time:
I used a small file to remove burs/sharp edges on the factory spring:
I then test the switch:
and epoxy the wires to the body of the switch - why? It is coming up ...
install the switch:
Now we finish with the head (since the body and switch are ready). First we apply a thin coat of thermal paste to all heatsink surfaces that will touch the body:
we then "carefully" screw the heatsink into the head, and then wipe/clean up any excess thermal paste (we want the new user to have a neat/clean light, don't we?):
Now comes the fun part. Remember when I said above that the reflectors are held in place by friction between themselves, the glass, and the cuts in the heatsink, well, it is not 100% perfect. It is about 98% perfect. What this means is that even though I centered each emiter perfectly, now during re-assembly the reflectors moved "slightly" off center - the good thing is that the beam is still perfect and you can't tell. The second photo was taken at an angle, so all look off-center, but when looking straight down it looks awesome due to the ultra clear glass:
OK, so now to why I had epoxied the wires on both the switch and the driver/heatsink. Simply because this type of heatsink requires that you twist the wires every time you screw the head in or out - in other words, just putting the light together puts stress on the wires!. Here are the 3 wires soldered (not taped yet), to show the dilemma:
So I used Kapton tape to protect the soldered joint:
and when screwing the head back in, I first go the opposite direction for a couple of turns to wind the wires in the opposite direction, so that when I finally screw the head back in I have fewer turns inside the body, therefore reducing stress on the wires. As you can tell by now, the epoxy simply provides additional insurance from the wires coming apart
And then of course, I tested the complete assembly:
And one more time, that very nice, even beam on the highest level:
I hope this was helpful
Will
Instead of starting with the beginning like I always do, this time I am starting with the end. This is the completed light - it is a custom tri-bored host, with custom grooved and fluted head, with a custom SS bezel, 4x P4 emiters, Al heatsink from the Shoppe, 4x 20mm reflectors from the Shoppe, ultra clear glass, MaxFlex boost converter, FM custom 3x 17670 LiIon battery holder, and custom body ring on tailcap:
and at about 1 Amp or so, it is "bright", and the head does get hot/toasty (this is a little over-exposed, but gives you the idea of how nice the beam is):
OK, so here is the journey ... please note that this is not the only way to make/put this together, and I welcome all ideas/suggestions/tips on this thread :twothumbs
First, given the universal heatsink, there are no locating points for the actual emiters:
Luckily, the reflectors touch each other while in position, and they also lightly touch the scalloped cuts in the heatsink. Also, the reflectors don't touch the emiters, but rest on the scallops, while being kept in place (no rattles) by the glass (more on this later on). So basically, I used the mechanical alignment of the reflectors to line up the emiters.
Of course, since the P4's have to isolated from the heatsink and since they will move/flow while the two-part thermal epoxy sets, how do I align the individual emiters once the reflectors are aligned? Drill holes on a spare plastic lens:
then, using a soft plastic tip, gently align each emiter through the reflector (tip learned here in the forums by no other than McGizmo himself - thanks Don!):
Once aligned, proceed to serially solder all LED's (the MaxFlex is a boost driver, so I am using it to drive 4 LED's in series, with a battery source of a lower voltage, in this case 3x LiIon cells). I am using a piece of tape to help hold each wire while I solder it:
then test to make sure it works:
I then added two-part clear epoxy to all emiters/wires to make sure nothing moves/comes loose, even if accidentally dropped:
I then move to my lathe, and cut the switch flush on both sides, to remove the tower and to allow better electrical contact with flat cells (thanks Mac for the tip on using the lathe for this!):
To provide a heatsink path to the MaxFlex, I used the lathe again to fabricate a small Al cylinder to transfer heat from the MaxFlex to the heatsink:
Here I already soldered the LED wired to the MaxFlex driver, and you can see the small cylinder here already glued to the inside of the heatsink (two-part thermal epoxy again), while I am testing the driver with a bench power supply:
I then applied two-part thermal epoxy between the cylinder and the driver, and used two-part clear epoxy to fix the driver to the heatsink. An observant person will also note the "excess" clear epoxy around the wires - that is on purpose - I will explain further below why:
I then moved to work on the switch. The OEM switch has to be converted to a momentary only switch, in order to be used with the MaxFlex driver (and most drivers from TaskLED www.taskled.com). I have already done a detailed photo-by-photo DIY on the switch mod, so I am only showing some highlights this time:
I used a small file to remove burs/sharp edges on the factory spring:
I then test the switch:
and epoxy the wires to the body of the switch - why? It is coming up ...
install the switch:
Now we finish with the head (since the body and switch are ready). First we apply a thin coat of thermal paste to all heatsink surfaces that will touch the body:
we then "carefully" screw the heatsink into the head, and then wipe/clean up any excess thermal paste (we want the new user to have a neat/clean light, don't we?):
Now comes the fun part. Remember when I said above that the reflectors are held in place by friction between themselves, the glass, and the cuts in the heatsink, well, it is not 100% perfect. It is about 98% perfect. What this means is that even though I centered each emiter perfectly, now during re-assembly the reflectors moved "slightly" off center - the good thing is that the beam is still perfect and you can't tell. The second photo was taken at an angle, so all look off-center, but when looking straight down it looks awesome due to the ultra clear glass:
OK, so now to why I had epoxied the wires on both the switch and the driver/heatsink. Simply because this type of heatsink requires that you twist the wires every time you screw the head in or out - in other words, just putting the light together puts stress on the wires!. Here are the 3 wires soldered (not taped yet), to show the dilemma:
So I used Kapton tape to protect the soldered joint:
and when screwing the head back in, I first go the opposite direction for a couple of turns to wind the wires in the opposite direction, so that when I finally screw the head back in I have fewer turns inside the body, therefore reducing stress on the wires. As you can tell by now, the epoxy simply provides additional insurance from the wires coming apart
And then of course, I tested the complete assembly:
And one more time, that very nice, even beam on the highest level:
I hope this was helpful
Will
Last edited:
:shakehead 
