10W LED light with Delrin canister

[ToddK]

Last year I built a Delrin canister LED light for myself, using an SSC P7 LED, and a Taskled HipCC driver. The canister and light head were both made from Delrin, using an inner aluminum head sink to hold the LED, reflector and driver. This first design left the switch boot totally exposed, and used a very small reflector. This year, I wanted to correct some of those short comings, and build a light for my dive buddy as a Christmas present.

I’ve tried to detail all of the dimensions and the part suppliers in this report. When I was making the original light, I worried quite a bit about whether all the details would work out, and would everything fit, so I want to provide this as a potential starting point for others. I’d like to say thanks to Wquiles, Packhorse, 350xfire for their awesome posts, and the book “The Divelight Companion” from AirSpeed Press.

Canister construction

The canister is 7.50” long by 2.5” wide on the outside, and has a 7” by 1.875” hole bored out for the battery. The entire canister diameter is really limited by the Agro gland and switch placement in the lid, so to get a little more room, I overbored the first 0.950” of canister length to a diameter of 2.00”.

I have a fairly small Harbor Freight 8x12 lathe, which is really about 14” between centers. Drilling or boring a hole usually requires 2X the length of the hole to allow for the length of the drill bit or boring bar when first starting out, which rapidly exceeds the length of my lathe for this design. Luckily I found that I could take the drilling & boring in three chunks. After putting a nice finish on the outside of the canister, and parting it off, I began by drilling a 1” hole using a stubby drill bit, to a depth of 2.5”. I then bored this hole out to close to the 1.875” final diameter, which then made room for me to place a 6” long, 1” wide drill bit into the bored out hole, start the lathe, and then drill out another 2-3” and repeat the boring process. I finished with a long 10” brad point drill, with the center spike ground down to be more flush. In using this bit, I run out of length to even chuck the bit, so I have to take the drill holder off the lathe, chuck the bit, and then put everything simultaneously back into the canister hole, and on the bed of the lathe. I have a nice ¾” diameter, 12” boring bar from Victor machining, which was absolutely essential for getting the inside of the canister made. After getting everything to almost the correct dimensions, I did finishing runs with the boring bar, to try and eliminate the sign of the chunks as much as possible.


Drilling with the medium length bit:

Drilling with the long brad point bit. The green tape marks the max extent of the drill into the canister:

Lid construction

 

[ToddK]

The lid of the canister has two grooves for installing barrel seal o-rings, a 15/32” through hole for mounting a toggle switch, and an Agro gland for passing the 18 gauge 2 strand Carol cord. I prefer the DS-116 switch over the ones sold by Salvo / Light Monkey because they are smaller, which enables them fitting into the critical lid diameter, and then have screw terminals for soldering the wires to. In this design, I left the lid thick enough that the entire toggle would be surrounded by Delrin inside a 0.860” diameter whole, so that it wasn’t exposed. OK, I actually missed this and it sticks out a little. I measured the depth required for the toggle hole by measuring the switch boot height. However, the toggle stretches the length of the boot slightly, which I didn’t account for.

Lid construction begins with turning down a piece of Delrin to an outside diameter of 2.50”. I narrowed the last 0.90” down to 2.000”, so that it would fit nicely into the canister top, with minimal extra space. I prefer to then cut the grooves for the barrel o-rings while there is still full support. The O-rings are -224, which require a 0.190” gland width, and a depth of 0.112”. I start under sized, and constantly fit the o-rings and canister, and then stop when the dimensions are pretty close, and everything fits together easily. I found myself constantly consulting these two pages for O-ring sizes and gland dimensions:

Size Cross Reference

O-Ring Gland Design

After the O-ring grooves are cut, I bored a 1.55” diameter hole to a depth of 1.15”. This hole allows mounting the switch and doing the wiring on the inside on the lid, and not intruding into the canister. Now that the side of the lid that goes into the canister is finished, I parted off the lid with enough extra material to allow finishing to a total lid length of 2.30”

Now that the part of the lid that goes into the canister is completely finished, I added one piece of electrical tape to each jaw of the lathe chuck to avoid maring the finished surface, and (I’m sad to say), chucked the piece by the 1.875” diameter that fits into the canister. I then finished the top surface, and moved the work piece offcenter in the 4 jaw chuck, and bored a 15/32” through hole for the switch. I enlarged the first 0.90” of this hole to a diameter of 0.860” to allow for mounting the switch boot inside the recess. I should have bored the hole to a length of 0.95” rather than 0.90” as earlier described to get the switch boot fully recessed. After the switch was taken care of, I moved the piece to a different offcenter location, and bored a 0.550” hole to allow for tapping the M16x1.5 thread for the Agro gland.

Tapping for the gland (this is a shot from an earlier, thinner lid):

The Agro glands are very nice, but do require an M16x1.5 metric tap, which I tapped manually on my lathe. Since the glands only have 2-3 full turns of thread on the outside for screwing into the canister, I used some 3M 4000 UV Marine Sealant (similar to what wquiles suggested). My Agro glands had O-rings that were dry, cracked, and generally looked old and a little thin. I replaced them with M2x14 O-rings (ID 14mm, OD 18mm), which fit quite well. The glands also require removing about 0.020” of diameter on the inside of the stainless steel nipple, and plastic bushing to get the cord to pass easily. I also found that the amount to remove varies with the piece of Carol cord that you get, so measure before you bore it out! I also cut the first ring off the tip of the strain relief to allow the coord to feed through more easily. When putting the cord through the glands, use some silicon spray on the cord to get it to slide easily through the rubber strain relief. Also, tug on the end of the strain relief to ensure that it is fully seated in the stainless steel ferule. If it is properly seated, there are two little rubber dimples that will show past the end of the ferule. If you don’t have it seated fully, it is a bear to get the internal threads of the gland to get close enough to mate!

I ended up using Dean’s Ultra connectors for connecting the battery to the cord. I’ve attached a comparison of the Tamiya, Anderson Powerpole modular connectors, Traxxas connectors, and the Dean’s connectors. The first two really are bulky, and cheap looking. Both the Traxxas and Dean’s connectors are very nice, and I went with the Dean’s due to the small size, and high quality.

When attaching the wires, I found that switching my soldering iron from 20W to 40W made it much easier to do a high quality job. I used the heat shrink that comes with the Dean’s connectors on each wire, and then put some 3/4” heat shrink tubing over the entire connector. When soldering, it is important to keep the heat shrink tubing about an inch away from the soldering iron, or it will start to shrink. I put the male end on the canister lid, so that the female could be on the battery pack.

Dean's connectors post soldering:

Connector with final heat shrink:

Just before installing the switch:

The outside of the completed lid:

The inside of the completed lid:

Final canister & lid assembly notes

Next the spring latches are added to enable locking the lid to the canister. Drilling the holes is somewhat nerve wracking, and the sets of holes need to be completely opposite of each other on the canister, set at the correct length on the canister and lid, and drilled to the correct depth, so that they don’t penetrate the canister. I mark the drill bit with a piece of electrical tape, so that the hole doesn’t get drilled too deep. I used Airspeed press’s tip of a sheet of paper wrapped around the canister and the lid, and then folding it in half, to allow for exact placement of the two sets of holes. Taping with a 6-32 bottoming tap completes the holes. I tried to bend the tabs on the latches inwards so that they wouldn’t get inadvertently snagged. I never hit on a great process for doing this, and found that about half the times, the latches would move as nicely as when I started.

I folded a piece of 1” webbing into thirds, and epoxied the inner third to the outer third, and clamped it in a vice while the epoxy dried. The webbing is then secured with two SS hose clamps. I bought the 1 13/16” to 2 ¾” size, and cut 3-4 increments off of the strap and ground down the resulting edges.

After seeing several posts from Tobin at Deep Sea Supply that amateurs shouldn’t put Li-Ion battery packs together, I purchased two that were custom made by BatterySpace.com/AA Portable Power Corp. Tel: 510-525-2328 - Powerizer Battery Official Site. The packs contain 8 18650 Li-Ion cells, two cells in series for 7.4 volt output, and 4 strings in parallel. At the time, the price was quite reasonable, maybe $20 over what was available off the shelf. Since then, it appears that they won’t do custom packs without UN regulation testing, which is prohibitively expensive. Looking at other online battery places, I don’t see the same sort of restrictions. The charger also from batteryspace, and is 1.2A rated, which should fully charge the pack in 11 hours. I also purchased one of the burn bags (I mean, charging bags ).

 

[ToddK]

I began with a 35mm reflector for the LED, and turned off the lip of the reflector, and also removed all of the threads on the bottom of the reflector. I then bored the hole out to 3/8”, to allow for mounting a SSC P7 LED in the reflector. The LED is epoxied to an aluminum star heat sink using Arctic Alumina thermal adhesive. I originally pre-tinned the LED leads, then placed the epoxy on the LED, centered it, and then proceeded to solder down the leads to the heat shrink star. I think if I did this again, I’d let the epoxy fully dry before doing the soldering.

Leads attached to the LED star:

Next, I attached two leads to the LED driver, which will eventually tie into the Carrol cord at the back of the housing. The attached shot shows the driver, and the heat conducting pad that comes with it:

I then turned down a 6061 aluminum rod to a diameter of 1.300”, and then boring a 0.86” diameter hole to a depth of about 1.00”. The outside of the reflector is a cone with a 23 degree (?) taper. I put the base of the reflector into the hole that I had bored, and then aligned one of the sides of my quick change tool post to the angle of the reflector. Once the right angle had been set, I then continued to bore the taper, such that the star heat sink, LED where sandwiched between the bottom of the hole in the aluminum, and the reflector. When everything went together with minimal slop, I knew that I was done boring the taper. Note that the 1.30” diameter is not wide enough to fit a full cone that fits the reflector, so the reflector sticks out of the aluminum rod by about 0.20”.

Now that the LED & reflector are taken care of, I changed the lathe speed to 125rpm, and cut a 0.25” wide groove to a depth of 0.35”, starting 1.1” back from the LED side of the bar. This groove allows for running wires between the LED and the driver.

The LED side of the heat sink:

After the grooving, the heat sink was parted off at about 2.05”, and on the side opposite the LED, I bored a 1.1” diameter hole to a depth of about 0.57” to accommodate the HipCC driver.

Just after finishing the face for the LED driver compartment, prior to drilling and boring:

I then carefully drilled and tapped two holes for 3-48 screws to hold the driver in place. Two more holes on the driver side, and two on the LED side allow wires to pass between the driver and LED.

Running the wires on the LED side:

LED full installed:

Installing the driver:

 

[ToddK]

I began the light head with turning down a Delrin rod to 1.90” for the first 2.50”, and then open up to a 2.10” diameter for the next 1.25”. I then bored a 0.55” hole into the first three inches of the center of the rod, and tapped it with the M16x1.5 to accept another Agro gland.

At this point, I parted off the piece at a little above 3.75”, rechucked the piece on the smaller diameter, again using a single piece of electrical tape on each jaw, and finished the other surface so the total length was 3.75”.

I started boring out the interior of the light head, beginning with boring out a 1.75” diameter hole to a depth of ¾” of an inch. The groove for the spiral retaining ring is the next thing to be cut, and I reduced the lathe speed to 125 rpm. I use a special interior grooving tool that cuts a 0.062” groove width, and I place the groove such that the glass disk is held in place with no slop. After the retaining ring groove is finished, it is time to cut the groove for the o-ring face seal. I ground my own trepanning tool for this task from a square tool bit by grinding down to the correct width, and then grinding the trailing outside corner of the bit to a circular shape, so that the bit would fit in the groove without the outside trailing corner hitting.

Next, I bored a 1.300” diameter hole to a depth of 3.00” as measured from the end of the piece. After this is accomplished, I bored down another 0.50”, with a diameter of 1.1”, to allow room for the wires to fit inside the light head.

The inside of the lighthead, fully machined:

Next, the holes for mounting the goodman handle were drilled and tapped, again marking the drill bit at the maximum extent:

With the heatsink, LED & reflector installed:

After the o-ring, heatsink, reflector and glass were installed, the spiral retaining ring can be fed into place. In general they go in fairly easily, and are more challenging to pull out without slipping back into place with a snap. Smalley’s has some nice images showing how they go in and out. http://www.smalley.com/pdfs/spiral_retaining_ring_install_and_removal.pdf

The finished lighthead:

 

[ToddK]

After getting everything together, I took the full light for a ride in my new Cornelius keg, to a simulated depth of 140’ for one hour. I did a 20 minute burn test in a bucket of water, and everything was fine, the beam stayed nice and white, and didn’t start to turn blue. Beam shots show a spread of 21” at 10’, with lots of spill outside the hotspot, giving an angle of 10 degrees.

Beam shots in the dark:

Beam shots with lighting in back:

I have a dive trip planned for right after Christmas, but my buddy doesn’t know the real reason yet.

 

[ToddK]

Canister drawing:

Canister Lid drawing:

Aluminum Heat Sink:

Lighthead:

Components & costs:

And now that this is done, I think I'm going to have a beer!

 

[trappist]

Wow... top notch work here ! I'm impressed, though it's a question of having access to equipment. Still the finished light is more than professional !

I got my main light head dead in the past few weeks, I have replaced the HID 10w inside and my bro's one being also dead, I will try to mod it into a LED one. I chose to use an SST-90 (2300lm) LED (9A !!), we'll see how that goes. Total cost of the mod since I have only to change the content of the head is ~100$. Ill try to take a few pics if I don't forget ! Usually I want to get it done so I forget about building the full case to post :anonymous2:

 

[ToddK]

Wow... top notch work here ! I'm impressed, though it's a question of having access to equipment. Still the finished light is more than professional !

I got my main light head dead in the past few weeks, I have replaced the HID 10w inside and my bro's one being also dead, I will try to mod it into a LED one. I chose to use an SST-90 (2300lm) LED (9A !!), we'll see how that goes. Total cost of the mod since I have only to change the content of the head is ~100$. Ill try to take a few pics if I don't forget ! Usually I want to get it done so I forget about building the full case to post :anonymous2:

Definitely post some pics when you are finished! The P7 LED has been out for a while, and I keep wondering about the efficiencies of a newer LED. It would be nice to not generate as much heat.

What are you using for the driver?

 

[stano]

Nice work! Keep us posted.

 

[=C=]

really impressive