Fixture-ring lamp potting

There are two main ways in which lamps are held in place relative to reflectors. They are either held with a socket and pedestal assembly, such as with the MagLite's, or they are potted directly into the bore of the reflector, such as with StreamLight's 20X and 35X, and TigerLight's Gold System lights. SureFire P60, P61, P90 and P91 lamp assemblies are essentially the same as this second method, because the lamp and reflector form a single unit.

The socket and pedestal assembly allows for flexibility in focusing the tightness of the beam and does not necessitate throwing away the reflector along with a burnt out lamp. However, the filament is often not centered above the center of the socket to such a degree as to introduce serious beam artifacts. Bending the pins of a bi-pin lamp can sometimes fix this problem, but not always. In addition, the heat and IR radiation from the lamp can be a problem for many sockets, causing them to melt and deform.

The lamp/reflector assembly—where the lamp is fixed in place in the reflector via a potting compound or ceramic discs and springs—does not allow for any focusing of the beam and requires that the entire assembly be replaced anytime the lamp fails. However, one can almost always count on a decently centered hot spot and a beam with no serious aberrations or artifacts, as well as a high degree of heat resistance and shielding due to the ceramic material or potting compound.

From the point of view of the flashlight modder, potting lamps into reflectors is an unmitigated nightmare. There are three axis positions (x,y,z) and one angle (filament angle) to get right, and this simply cannot be done by eye. A jig is required to power up the light in the reflector, adjust the relative positions until a good beam is achieved, and then hold the lamp and reflector while the potting compound is injected and cured.

In the process of muddling my way through my TigerLight modding project, I was driven by several necessities to a sort of hybrid method which I believe has some small degree of interest for home flashlight-modders such as frequent CPF. I call it "fixture-ring lamp potting" or "ring-potting" for short. I highly doubt that this idea is original or remarkable from a professional's point of view, but I have not seen anything like it detailed on CPF and so I am detailing it here in case anyone would like to use it, or is interested in how I am putting together my lamps and reflectors for the TigerLight Upgrades.

The idea is simple: a Carley RF1940 2" reflector has a .377" bore with a set screw for securing a lamp in the reflector. ( Carley sells metal sleeves which somehow fit over the lamp, but the sleeves they have will not fit Welch Allyn lamps, and this idea has never appealed to me due to the danger of a short circuit across the lamp pins and the danger of cracking the glass capsule due to pressure from the set-screw. ) This allows the potting process to be broken down into two separate steps by first potting the lamp into a .375" cylindrical ring (3/8") and then sliding the ring up into the bore until the proper focal length is obtained. At this point the set screw is tightened to secure the ring-potted lamp in position. Thus a person can visually center the filament in the ring, pot it, and then move the ring in and out to get the desired focus and beam.

Thus, the immediate questions become "what do I use for fixture-rings?", "what compound do I use for potting?", and "how do I hold the potting compound from leaking out the bottom of the ring while potting the lamp into place?"

I use sections cut from a 3/8" o.d. stainless steel tube, with a wall thickness of .028. Unfortunately, many nominal 3/8" o.d. tubes are slightly larger than .375" and will NOT fit up through the bore of a 1940, but with some persistance a suitable length of tube can probably be found. The rings must be cut with an abrasive cut-off wheel in order to avoid deforming them and leaving overly large cusps on the edges.

For a potting compound, I am using Cotronics 940LE ceramic adhesive/potting compound. Anyone can buy from Cotronics, but the smallest size is a pint (for $65 or so) and it needs to be mixed by weight. A scale accurate to .1 gram is needed.

I cover the bottom of the hole with the blue, removeable masking tape, and then poke the lamp wires through the tape. I first make sure that the wires are straight and even, and then hold the ring up to the light so that I can see the positions of the wires when I hold them to the tape. That way I can ensure that they are centered before I poke them through the tape.

Below is a picture of two empty rings, and one ring covered with tape, with a WA lamp poking through. Notice that the rings have a hole drilled through one side and partially through the opposite wall. I had to start doing this in order to preclude the possibility of the potting and lamp coming lose from the ring with repeated thermal cycling and insertion and removal from reflectors.



I have also found it necessary to fashion some sort of jig to hold the lamp and ring during the potting process. I drilled holes in a piece of aluminum bar stock, and then drilled out shoulders in the holes, so that the rings with tape fit snugly in the bar and bottom out when the hole narrows, and so that the pins of the lamp poke down into the narrow section and are held above the table due to the thickness of the aluminum. My rings are .260 or so deep and I find that if they sink less than half way into the aluminum that this is ideal. I used 1/2" thick and 2" wide stock, holes 2" apart on center, with an inch clearance at the ends. A picture of one of these jigs appears below:



Notice that the lamps are pulled up high and are leaning out of the way. This allows for easier delivery of the potting compound. I apply the potting with a disposeable plasitc syringe. Note that you can carefully push the lamp down into the potting, and then pull it up again to add more compound if needed. The potting is quite paste like and will not leak past the lamp pins, but the pins must NOT leave the holes in the tape!

I have found that it is much easier to accurately center the filaments if you use a 1940 to help you. I carefully place the reflector down over the half of the ring that sticks up above the bar, and move the lamp around with a Q-tip until the filament is centered. DO NOT TOUCH THE LAMP GLASS during the potting process, as it could cause the lamp to explode or leak during running. Cleaning after the fact will not necessarily help as the portion of the lamp glass that is below the top of the ring but above the potting will be inaccessible. I clean the lamps thoroughly with isopropyl alcohol before I start--just to be safe--and then use a cloth or Q-tips from then on.

I go through a few round of placing the reflector over a lamp and adjusting it and moving on to the next one, until I am certain that everything is positioned correctly. Next I put the jig in the oven at 200F for 15 minutes, remove and let cool, then take off the tape and file off any bit of potting that got on the pins and outside of the ring. here are two finished ring-potted lamps:



Notice the red dot on the back of one of them. That indicates a WA 1185 lamp. I find this to be a good way of marking the ring-potted lamps, as they will all look very much alike otherwise.

Again, avoid touching the glass, but if you do it is not critical at this point. Simply clean with isopropyl alcohol before use.

And finally, slide the ring-potted lamp up into the bore of the reflector to a likely position and snug down the set screw. There is no need to tighten the set screw very much--ever. Now attach the lamp lead wires, loosen the set screw, turn on the light, and move the ring-potted lamp in and out to obtain the best beam and re-tighten the set screw. Here is what you end up with:



And that's all I can think of at the moment. I can sell a small amount of the Cotronics 940LE to interested CPFers. PM if you're interested.

Hi Jim,

Thanks for sharing all of this - it's good stuff. I have 1274's and 1160's that I'll someday attempt to pot in rings, so your info here will no doubt be of great help to me.

BTW, my 1940/1274 was waiting for me when I returned last night from my business trip. A few minutes on my lathe and it now resides in my Mag 4D/6x1/2D. I had previously made my own bi-pin PR adapter so the TL LA would plug right in. Now, swapping-in the 1940/1274 combo was child's play... /ubbthreads/images/graemlins/smile.gif

Thanks - John

After potting many dozen lamps, I wanted to add a few things to this thread:

First, the o.d. of the "3/8th" inch tubing should really be .374". Even .375 is too large for some of the 1940's, whoses bores are not precisely .377, but vary on the order of +/- .001 or so.

Second, it is best to test the rings in a 1940 whose bore is on the small side, to make sure they will fit in all 1940's. It is absolutely necessary to file down the entry drill-hole and also the outside surface directly opposite this because partially drilling into that side causes a small bump to protrude into the outside surface. Next, the whole outside of the ring must be sanded smooth. I use a combination of dremel tool accessories to do this: attaching the rubber peice from the midsized sanding drum attachment to the smallest sized shank creates a set-up which will hold a 3/8" o.d. ring when fully compressed by the tightening screw.

Third, clean the inside of all fixture-rings with a bit of pipe-cleaner and soap and water after cutting and drilling them, but before filing and sanding. Or it could be done after filing and sanding. Just as long as any dust, dirt, oil, and other contaminants are removed, and the rings are dry before the potting compound is applied.

Fourth, take care when centering the filament in the ring. As I mentioned before I use a fixture to hold the rings and a smooth Carley RF1940 reflector to help guide my eye. Use the same smaller bore reflector for potting that you used for testing the rings to make sure they were smooth and would fit nicely. Carefully place it over the ring and lamp without touching the lamp. Then rock it back and forth front to back and side to side in order to "map out" the play between the ring and the reflector bore. Position the reflector in the center of the region of play, then look squarely down into the reflector so that your eye is on-axis with the reflector. Now move the lamp around with a Q-tip until it is centered. Unfortunately, the position of the lighting will play tricks on you because it will light up one side of the filament more than the other. Thus, I have found that to get the best results it is necessary to rotate the jig 90 degrees during the centering process in order to truly ascertain the position of the filament. Be careful not to bump the lamp while removing the reflector. Also, make sure not to pull up the ring and lamp when you remove the reflector. If you do, re-install it into the jig without touching the bulb and start over.

I have many feet of some nice stainless steel .374 o.d. tubing and could sell small lengths of this as well as small amounts of the 940LE.

Fifthly and lastly, a scale accurate to .1 g is a bare minimum. A scale accurate to .01 g is much much better. I recommend the My Weigh 50g and 100g Durascales, and the My Weigh MX-50. These are relatively inexpensive ($75-$100) and highly accurate scales (+/- .01 g) which read to .01 grams. Be careful not to drop them. BTW, around 2 grams of the activator (liquid) with the appropriate amount of powder (liquid/.45) makes the right amount of potting compound for 6 lamps. The pot life is 20 minutes at which point the compound gets a bit too stiff to move the lamps around easily.

It's funny, really, but the most obvious of things often is the last thing we realize.

When I first started potting lamps into fixture-rings I was just eye-balling the placement of the lamp, by looking at the glass capsule. Then I would put it in a reflector and turn it on, and realize that it was off-center, and I soon had the revelation--duh!--that the filament was the important thing to center. But still, often I would secure the lamp in the reflector and look and it was obvious that the filament was not centered.

Then I got to thinking that if I was so obviously seeing whether or not a filament in a ring-potted lamp was centered when secured in a reflector that I should use a reflector to center it in the first place!

So I started doing that and there was a dramatic increase in my success rate. But not 100 percent, to be sure. Now I have had another "duh!" type realization. I am using a reflector to center the filament during potting, but I am not tightening down the setscrew on the ring. Instead I am moving the thing about to "map out the play" and mentally figure out the neutral position. But why not secure the ring in the reflector during potting in the first place, since that's what I do to check to see if the filament is indeed centered! I hope this is the last "duh!" but it probably isn't.

Anyway, I am now going to shorten the butt of the reflector I am using (or another one, maybe, just to be safe--in case this idea doesn't work out) so that I can tighten down the set-screw during the potting process to ensure that the axis of the ring is parallel to the axis of the reflector bore. I'll report back on how this works out. (Just on the off chance that anyone cares and to document my methods and procedures for future reference in case anyone ever wants to try his or her hand at this method. Personally, I really like it and think that it offers distinct advantages over potting into a PR base.)

Jim,

Another problem is that I have 6 or 7 Carleys, and every one of them slightly changes the centering of the bulb due to the very slight variation in bores.

I usually can get a good center, and I haven't (obviously) played with enough of your ring potted bulbs of other types than the 1160's that you did for me, but even though those all appear to be PERFECTLY centered in the ring, they obviously fit differently when centered in different reflectors.

I think at least some of the problem is the actual reflector manufacturing process. I have received brand new Carleys that are almost 1/16 of an inch different in bore. It only is important to me if I have to rebore because they are too small, but the bulbs do fit differently in them.

I have often wondered if Carley bores with different machines or if some are hot bored and others are bored at lower temps.

Bill

Bill,

Good observations. My experience mirrors yours. The bore in a 1940 varies in diameter and is sometimes off of perpendicular to the plane of the rim. I make sure that my rings are small enough to fit in even the smaller bored 1940's.

As for off-axis bores (I'm talking about the hole in the butt of the reflector and not any class of people, BTW /ubbthreads/images/graemlins/smile.gif ) there is nothing I can do about that. In my instructions, I suggest that people try rotating the ring to different positions relative to the set-screw to try to get a good centering. I myself am almost always able to get a decently centered filament this way--assuming that it turned out to be well centered in my test reflector, anyway.

And even when the filament is a bit off-centered, the beam is still pretty good, especially when compared to the general run of flashlight beams. It may not be SureFire quality, but usually it's pretty close.

I suppose I could look into other manufacturers of reflectors, but I think the Carley RF1940 is a great value for the money, and I like the people at Carley. Plus, Carley will sell small quantities to small fish like me. Actually, I'm probably more of a tadpole or a minnow. /ubbthreads/images/graemlins/crazy.gif

Jim, I'm glad it isn't just me, then.

The first Carleys I bought were from members here and I always wondered if they might have been "adjusted" by the members, versus straight from the factory, but some that I received in Group Buys were also the same.

I agree, I have always been able to eventually get a good pattern, but sometimes it is a little tougher!

As for Surefire quality, the problem is NOT your potting method, but (and I know YOU know this, but for those who may not) the size of the filament and the type of reflector texturing used.

I agree about the Carleys, they are GREAT for the money. I have actually looked at several other sources and not found any better for anywhere near the price. Of course, if you WANT to pay 100 bucks a piece, there are some awesome surfaces out there!

Just a kudos and congrats on all your work, including the M6 project. Great, innovative thinking from one of the Hotwire Emperors!

Bill

I'm so fond of this scheme that I am working on adapting it to the larger sized Osram lamps, such as the 62138, 64440, and 64447. I will be using 5/8" o.d. stainless steel fixture-rings for this, and I have already found some good "molex" type connectors (actually they are made by AMP/Tyco) which will grab onto the 1 mm dia. pins perfectly and which are rated to 9.5 amps.

More details and pictures and such-like to come.

I believe that this scheme has significant advantages over a pedestal:

1. Heat is not an issue for the ceramic potting compound or stainless steel rings and there is no need for a pedestal and no need to make a heat resistant pedestal.

2. The heat is actually somewhat shielded and contained in the head/reflector of the light, although it will still obviously conduct along metal surfaces--but at least the radiant heat is contained.

3. The filament is always more or less well centered. Because the relation of the filament to the glass and thus to the pins is not precise, a pedestal and socket arrangement must be able to be adjusted for a well centered filament. No need for that mechanical complication with the ring-potting scheme

4. Still offers focusing of the beam, even though this is not practical "in the field" and can not be done on the fly. Contrast this with SureFire's M3, M4, and M6 LA's, which although excellent, restrict a person to only one beam focus.

I have completed the preliminary testing of ring-potting the 100W Osram 62138 lamp into 5/8" o.d. fixture rings, using the Cotronics 940 LE potting compound. The 940 LE, being specifically engineered for this very application, has acquited itself with the highest marks: it has held up beautifully to the first half-dozen or so burns and looks just as it did when new. There is no indication of cracking or crumbling or burning or any other stress.

For the purposes of the initial testing, I enlarged the bore of a 1940 to fit the 5/8" o.d. rings, and I did so from front to back in a single diameter cut to make things simpler. However, on the next reflector I modifiy (after my back problems have been sorted out), I will enlarge the bore nearest the reflector surface to just under .500" to allow the bulb to pass through, with some alignment room to spare, and then I will enlarge the rest of the bore, out towards the rear of the relfector, to just over .625". Thus I will retain as much reflector surface area as possible, while still allowing use of the common 5/8" o.d. stainless steel tubing. (I did do an extensive search for 1/2" i.d. SS tubing, but I came up empty handed, and discontinued the search in favor of a more complicated reflector mod).

Here is a picture of a ring-potted 62138 next to the initial modded reflector. Note that I accidentally over-enlarged the bore, so that it is significantly larger than it needs to be.



And here's what it looks like from the top:



Note that the potting compound does NOT entirely fill up the ring. On the contrary, it is only just above the level of the locking holes in the side of the ring; thus most of the pinch is still open to the air.

For these intial tests I ran wires from my 12V 13A Cosel powersupply, with the AMP/Tyco connectors on the ends. At first, I covered these in general purpose shrink wrap. The heat generated by the lamp was such that it melted and burned this wrap in short order. So I cut off the connectors and attached new ones with kapton tape wrapped around them instead. The kapton tape has held up just fine.

The connectors, on the other hand, are unable to take the heat and lose their springiness and ability to grab the pins with enough force. I'll have to either locate heat tolerant connectors, or simply use a ceramic socket in place of the individual connectors. The only problem with this would be the extra expense and the loss of the ability to attach to ANY pin spacing. These aren't really significant issues, so either way will be fine.

As I mentioned before, the advantage to this method is the inherent heat resistance of the potting compound and stainless steel fixture-ring, as well as the space saved from having no pedestal, and the advantage of being able to pot the filament so that it is well centered in the ring, regardless of how it is positioned in relation to the pins and the glass bulb.

Also, the Carley RF1940 set-screw reflector holds the ring very securely and firmly in place throughout the entire run, from being cold at turn-on to being hot enough to de-temper metal.

A WORD OF WARNING:

These high wattage lamps are not toys. They produce an incredible amount of heat and IR radiation. If anyone is using these for a homemade or modified light, be very, very careful. NEVER leave such a light unattended, or shining on anything flamable, and for God's sake, NEVER ever ever--even for an instant--set the light bezel down while it is on. Never let a child play with or handle one of these lights. 100 Watts is a serious amount of power, especially in a small host, such as a Mag Lite.

Be careful, people.

Thanks Jim!!

Awesome report and caution about these lamps!!

I send out a caution with the WA lamps I sell, and the Osrams are MUCH, MUCH more potent!!

Looks like a great new option!

Once again, a MASTER finds a new plan!

Bill

Hi All,

McMaster sells cat#7551A23 4oz tube of ceramic adhesive putty, good for 2300degF. They say set time is one hour and cure time is 24 hours or 2-4 hours @ 250F. Abt $13 They also have higher temp-rated/higher$ stuff.

Larry

Larry,

The other consideration when choosing a potting compound is the coefficient of thermal expansion. Quartz glass has an extremely low rate of expansion, and the potting compound really should have a similarly low coefficient. This is exactly what the 940 LE is designed for: the "LE" stands for "Low Expansion" and is the same as for quartz glass.

So, what is the coefficient of thermal expansion of this McMaster stuff? Because if it is low enough, this is a much more convenient option for most people than ordering a quart of the 940 LE from cotronics corp. to the tune of $64.95. For myself, however, I really appreciate the 5-15 min. @ 200F drying time of the 940 LE: pop the lamps in the oven for 15 minutes and TA-DA! They're done.

Hi Jim,

Great writeup and excellent update on one of the more significant stealth modding innovations of the year. I think that if you go with the sockets I'm using, you won't be giving up much if any compactness. They are only 9.09mm in total thickness and 19.00mm in diameter. I use these in the Aurora, my 12V lantern mods and my GY6.35/5.3 test rig and they have performed well. Drop me a line if you need details.

This is a distinct arrangement that is uniquely useful and even if pin receivers cannot be employed. Nicely done.

Wilkey

Jim,

The catalog states it is aluminum oxide-based, so the cte should be pretty low. The also list mica-based adhesives supposedly suitable for bonding glass-to-metal but minimum is a 1# can for ~$33. (Of course $83 gets you 7#/1/2 gal.) cat# 7482A31

The king seems to be the Zirconia stuff: "One-part paste adhesive dries to form a hard surface and easily handles temperatures to 4000degF. It resists thermal shock, molten metals, and most chemicals. Use to bond ceramic, glass, metals, and graphite..." ~$57/pint can. cat# 7564A11

I posted for others who might want to try your techniques and need a smaller qty, lower cost option.

Larry

Thanks for the info Larry. Can anyone order from McMaster-Carr? I thought you had to have an idustrial account.

Wilkey

Wilkey,

I *think* anyone can. They have a warehouse with walk-in counter in Santa Fe Springs. It's too convenient for me at work (on line)- I just have to be careful which credit card I click. (The company card or my own. /ubbthreads/images/graemlins/grin.gif ) In the greater LA area you have the option of courier delivery for the same or less than UPS, and it's same or next day delivery. They even have a bulleten board!

Everything they sell can usually be purchased from McM's sources for less, but the convenience is priceless!
(Yes, I am a McMaster-Carr junky!)

Larry

[ QUOTE ]
js said:
I'm so fond of this scheme that I am working on adapting it to the larger sized Osram lamps, such as the 62138, 64440, and 64447. I will be using 5/8" o.d. stainless steel fixture-rings for this, and I have already found some good "molex" type connectors (actually they are made by AMP/Tyco) which will grab onto the 1 mm dia. pins perfectly and which are rated to 9.5 amps.

More details and pictures and such-like to come.

I believe that this scheme has significant advantages over a pedestal:

1. Heat is not an issue for the ceramic potting compound or stainless steel rings and there is no need for a pedestal and no need to make a heat resistant pedestal.

2. The heat is actually somewhat shielded and contained in the head/reflector of the light, although it will still obviously conduct along metal surfaces--but at least the radiant heat is contained.

3. The filament is always more or less well centered. Because the relation of the filament to the glass and thus to the pins is not precise, a pedestal and socket arrangement must be able to be adjusted for a well centered filament. No need for that mechanical complication with the ring-potting scheme

4. Still offers focusing of the beam, even though this is not practical "in the field" and can not be done on the fly. Contrast this with SureFire's M3, M4, and M6 LA's, which although excellent, restrict a person to only one beam focus.

[/ QUOTE ]

Of course I don't agree with points 1 or 3, as far as the pedestal. Slow focusing or fixed focus pretend M*glites are not my ideal of 'stealthiness', just the way I and some others view it. We have to agree, to disagree on that. Neither pedestal & socket OR permanent fixed socket arrangements are ideal, not advantage is not significantly greater in either scheme. Heat is always an issue, as is resistance, as is the ability to deliver ample/maximum current to the bulb.

Umm, is there an verified need for 2000 degree compounds, not saying that overkill is a bad thing? Just curious if there are other lower high-temp compounds that could be used?

I guess you guys are more anal than I am about the necessity of perfect centering. While I certainly would like to devise a scheme for even more precise, and simple non-permanent adjustment mechanisms, I do so for ultimate luxury/preference rather than absolute need. Axial filaments are even in less need of perfect centering, IMHO. SF LA are not always as perfectly centered as the posts above would have one believe. They would cost more if they were to be able to ensure such 'perfect' centering.

I wonder what the loss of 'springiness' in the connectors means? Just trying to determine if there are alternatives that would be possible...or not.

'Larger size Osram bulbs'..umm, how about one that is 2mm larger in diameter than the 64447IRC (not to be confused with the non-IRC style Osram bulbs of same numbering system, which is of a different form factor).

Can we get Carley to do a special run on deeper parabola than the 1940, maybe Ginseng has investigated this possibility already?

"Emperors"? Uggh, what's that saying about clothes? /ubbthreads/images/graemlins/wink.gif I am but a mere peasant, Robin Hood /ubbthreads/images/graemlins/grin.gif

Larry,

Thanks very much for the info! As you say, these are good options to investigate for those who will only need a small amount. According to what you've said, it sounds as if these would work fine.

Wilkey,

Thanks for the kind words. I appreciate them. I've been meaning to call you and will do so sometime early next week.

Well, I've had some more experience with ring-potted Osram GY6.35 lamps, and I have to say that I'm less enthusiastic about it than I was at the start.

The ring-potting method has worked very well, even with the extreme power and heat levels involved here (100W) yet even so, I now think that this method is really only worth doing for the Welch Allyn lamps. Why?

Well, for starters the Osram lamps have a tip-off which prevents a person from looking down from the top to judge the position of the filament. One must look from the sides in order to ascertain the filament position and this is inherently less accurate.

Next, the heat involved at the pins is such that only a ceramic socket specifically designed to handle the high temperatures will do over the long run. So as long as you have to use a socket anyway, why not mount it and avoid the whole ring-potting and reflector modding hastle?

Which brings me to the reflector modding issue: the necessity to accurately bore two different diameter coaxial holes in the reflector means that it is quite a bit more involved than the mods for a mag or the Tigerlight. And not much less involved than machining a fixture to hold a ceramic socket.

This whole ring-potting method was developed as a way to make lamp assemblies for the TigerLight, where there is no room for a pedestal or socket (due to the rubber shock isolation gasket). It works great in this application with the 1940 and WA lamps. (Although, even here, a major manufacturer would be better advised to simply choose to have the lamps potted directly into a spun reflector, which in point of fact, is what TL and SL do.)

However, in the case of something like a mag mod or any mod where the head can be turned in and out on the body, this ring-potting method yeilds no advantages over using a socket(or only one or two slight advantages), and the loss of the ability to focus the light on the fly, outdoors, with everything buttoned-up is significant.

Thus it is my considered opinion that the ring-potting method is not suitable in this class of applications.

Thanks Jim.

Wilkey