Hi guys,
In addition to the Ti Sapphire that uses Arc's AAA driver hosting a Nichia 3 mm LED (310DS) I have also built a prototype using a Nichia 036 1/2 watt LED that is a High CRI version of the 036. This LED uses the same dual phosphor as the High CRI 083 LED's I have been using in the SunDrop.
It is a pretty cool little LED with some interesting potential. It is a single die LED so it can be collimated reasonably well. However in the Sapphire AAA Mule, my intent was to use the LED for full flood with no secondary optic.
Practically speaking, IMHO, the Sapphire with the 3 mm Nichia is about ideal for general purpose in a key chain light. I say this based on the design constraints of being a single level output and the pros and cons of intensity level, flood VS spot (throw) and what have you.
However, for my typical use of a key chain light, I have found the 3 mm LED a bit too bright often. Fortunately it doesn't have a hot spot but it is often more light than I need.
The thought of a High CRI AAA light with greater flood and less lux was appealing to me but how could I use the existing Sapphire host to accommodate a tiny surface mount LED that by rights should be mounted on a MCPCB? I also knew that I wanted the LED to be as close to the window as possible with the least amount of restriction of light getting out the front end. In terms of thermal management, one advantage the 036 LED would have is that it is rated to a max of 350 mA with target current in the spec sheets of 150 mA and behind the Arc AAA driver, it would see a max of around 60 mA or so. The LED would be well below its rating.
I remembered back a number of years I had experimented with mounting a Luxeon directly to a sapphire window and that the sapphire is a pretty decent heat sink in its own right. If I wanted the 036 as close to the window as posible, how about mounting it directly on the window?!?
I used some Norland 61 adhesive in a drop on the window and then seated the face of the LED down into the drop and curred the adhesive.
The purple masking tape was used because it has a known emissivity and allows for reasonable measurement of the temperature of the sapphire window. Driving the LED at 80 mA, I measured the temperature of the LED and window with a FLIR. Although in the image above, you can see a captured picture on the FLIR showing the window at 51.6 C, after leaving the LED powered up for a longer period until presumably steady state was reached, I then measured 54.5 C on the face of the window and 56.4 C on the back of the LED package. With the window removed, the temperature of the LED rose to 65 C. When installed in the head of the titanium AAA host, the head of the light does get warm to the touch indicating that further heat is spread to the light itself. Once the window and LED were installed in the head, I added further Norland 61 UV curing adhesive behind the LED essentially potting it in the head and adding thermal mass and transfer capability. I have been told that this LED is a pretty robust little fellow and the Tj max is 150C.
Beam comparison between the 3 mm and 036:
The lights are quite close to the wall and those of you who have used the Sapphire with 3 mm LED know that it is a flood beam itself. You can see that the 036 is a very wide flood with signficantly less intensity.
Since this light is about CRI, I took some comparitive shots of te Color Checker leaned against a "white" wall. The camera white balance was set on daylight sun.
The tint seen on the "white wall" behind the ColorChecker is in keeping with what the naked eye sees but a bit exaggerated.
I introduced both of the light samples to my integrating sphere and the comparative spectrum with some data overlay below is the result. I can not vouch for the accuracy but suspect the data is within reasonable bounds.
When I was told about the High CRI version of the 036 and asked if I wanted some samples, I requested the highest CCT bin they had. I suspect the higher CCT is a result of the higher spike in the blue and also results in the sample I measured being down near the lower limit of the CRI range for this LED. The minimum CRI for these is 85 and I measured 87. The 3 mm LED actually did pretty good in terms of CRI but at a CCT that is certainly well above the reality of incandescent sources with an obvious difference in tint resulting.
Those of you who have color checkers can compare the images here with your color checker. For those of you who don't, the 036 renders the color squares much more accurately than the 3 mm and especially in the cases of the "problem" squares.
The spectrum I recorded is subject to some uncertainty in its own right and I have no idea how much variation one might find from sample to sample. My experience with Nichia has been one of pretty good consistancy from part to part. Below is an excerpt from the Nichia Spec sheet and you can see what they claim is a typical spectrum for this part:
I would guess that the warmer bins of the 036 High CRI were used in deriving the spectrum above.
One thing I found exciting in this exercise was the consideration of using the ceramic package itself of the LED as a thermal bonding surface as opposed to relying on surface mount thermal bonding. This could be cause for revisiting the diamond bog again!
(I do have a desire to try a slury of a UV curing adhesive mixed with diamond powder. Unlike epoxies, you wouldn't necessarily be wasting expensive goo as you would apply just the amount needed) I have had dialog now twice with a company who makes all kinds of LED related UV curing adhesives and even lens materials but in both cases, I was told they would get back to me and they never did. Being a small fish does have some disadvantages...
OK, long winded post which is not uncommon for me. I will be working on the next wave of Sapphires soon and my plan is to build at least a few of these "mule" versions. I have no idea what the assembly time will be like or how easy or hard it will be to buld some of these. Only one way to find out. I also suspect that the interest in these may be greater than the utility they actually provide. Without secondary optics, the Sapphire Mule doesn't reach out very far!! It is great in room bounce and up close illumination but as an outside light, the limitations are rather obvious! On that note, I don't think I have ever mentioned that in a pinch, you can hold a 1/2" diameter acrylic ball in front of any of the Sapphire versions and get some extended throw our of them! The ball seats cleanly down in the chamfer of the head and I have played with the idea of a molded ruber or silicone sleeve that would retain the ball and snap over the head of the light; a beam shaper if you will but not like the ones we are presently familiar with.
If you have a Sapphire and a 1/2" acrylic ball, check it out.
While throwing out ideas and thoughts, I'll add a few more.
I have also built a prototype Sapphire using the high CRI LED behind a reflector. I have some 8 mm reflectors from a few years back and I took one and chopped it down so that I could fit it into the Sapphire head without restricting the light exit.
It is in the middle below with the 3 mm "stock" sapphire on the left and mule on its right:
Beam shot exposed to show the mule's contribution (middle light and beam diameter covers both its neighbors:
Beam shot exposed to better represent comparison between "stock" sapphire and reflector version of 036:
I find the idea of using the 036 behind a reflector intriguing and this is the first time I have used a High CRI LED behind a reflector with results I find worthy of further investigation. The tint variation from spill to spot with the Nichia High CRI is much less obvious than what I found with the Seoul LED's I sampled. On the bench, I coupled the 036 with a full version of the 8 mm reflector and varied the drive current up towards the max rating with very cool results. I think the smaller and restricted reflector along with the lower drive level of the LED in the sapphire sample above does not do the LED justice in terms of a collimated beam application. Multi levels with higher drive seems an obvious direction for this. I also would like to come up with an appropriate driver to push a 7UP cluster of the 8 mm Reflectors with 036 High CRI behind them and a host of appropriate size and configuration. This is all speculation and just ideas here.
Many of you have plans to mod the sapphires so in that spirit I have included some of these ideas and thoughts. Like I said, I do plan to make some of the mule version I have prototyped but that is subject to a reasonable assembly and time considerations yet known...
In addition to the Ti Sapphire that uses Arc's AAA driver hosting a Nichia 3 mm LED (310DS) I have also built a prototype using a Nichia 036 1/2 watt LED that is a High CRI version of the 036. This LED uses the same dual phosphor as the High CRI 083 LED's I have been using in the SunDrop.
It is a pretty cool little LED with some interesting potential. It is a single die LED so it can be collimated reasonably well. However in the Sapphire AAA Mule, my intent was to use the LED for full flood with no secondary optic.
Practically speaking, IMHO, the Sapphire with the 3 mm Nichia is about ideal for general purpose in a key chain light. I say this based on the design constraints of being a single level output and the pros and cons of intensity level, flood VS spot (throw) and what have you.
However, for my typical use of a key chain light, I have found the 3 mm LED a bit too bright often. Fortunately it doesn't have a hot spot but it is often more light than I need.
The thought of a High CRI AAA light with greater flood and less lux was appealing to me but how could I use the existing Sapphire host to accommodate a tiny surface mount LED that by rights should be mounted on a MCPCB? I also knew that I wanted the LED to be as close to the window as possible with the least amount of restriction of light getting out the front end. In terms of thermal management, one advantage the 036 LED would have is that it is rated to a max of 350 mA with target current in the spec sheets of 150 mA and behind the Arc AAA driver, it would see a max of around 60 mA or so. The LED would be well below its rating.
I remembered back a number of years I had experimented with mounting a Luxeon directly to a sapphire window and that the sapphire is a pretty decent heat sink in its own right. If I wanted the 036 as close to the window as posible, how about mounting it directly on the window?!?
I used some Norland 61 adhesive in a drop on the window and then seated the face of the LED down into the drop and curred the adhesive.
The purple masking tape was used because it has a known emissivity and allows for reasonable measurement of the temperature of the sapphire window. Driving the LED at 80 mA, I measured the temperature of the LED and window with a FLIR. Although in the image above, you can see a captured picture on the FLIR showing the window at 51.6 C, after leaving the LED powered up for a longer period until presumably steady state was reached, I then measured 54.5 C on the face of the window and 56.4 C on the back of the LED package. With the window removed, the temperature of the LED rose to 65 C. When installed in the head of the titanium AAA host, the head of the light does get warm to the touch indicating that further heat is spread to the light itself. Once the window and LED were installed in the head, I added further Norland 61 UV curing adhesive behind the LED essentially potting it in the head and adding thermal mass and transfer capability. I have been told that this LED is a pretty robust little fellow and the Tj max is 150C.
Beam comparison between the 3 mm and 036:
The lights are quite close to the wall and those of you who have used the Sapphire with 3 mm LED know that it is a flood beam itself. You can see that the 036 is a very wide flood with signficantly less intensity.
Since this light is about CRI, I took some comparitive shots of te Color Checker leaned against a "white" wall. The camera white balance was set on daylight sun.
The tint seen on the "white wall" behind the ColorChecker is in keeping with what the naked eye sees but a bit exaggerated.
I introduced both of the light samples to my integrating sphere and the comparative spectrum with some data overlay below is the result. I can not vouch for the accuracy but suspect the data is within reasonable bounds.
When I was told about the High CRI version of the 036 and asked if I wanted some samples, I requested the highest CCT bin they had. I suspect the higher CCT is a result of the higher spike in the blue and also results in the sample I measured being down near the lower limit of the CRI range for this LED. The minimum CRI for these is 85 and I measured 87. The 3 mm LED actually did pretty good in terms of CRI but at a CCT that is certainly well above the reality of incandescent sources with an obvious difference in tint resulting.
Those of you who have color checkers can compare the images here with your color checker. For those of you who don't, the 036 renders the color squares much more accurately than the 3 mm and especially in the cases of the "problem" squares.
The spectrum I recorded is subject to some uncertainty in its own right and I have no idea how much variation one might find from sample to sample. My experience with Nichia has been one of pretty good consistancy from part to part. Below is an excerpt from the Nichia Spec sheet and you can see what they claim is a typical spectrum for this part:
I would guess that the warmer bins of the 036 High CRI were used in deriving the spectrum above.
One thing I found exciting in this exercise was the consideration of using the ceramic package itself of the LED as a thermal bonding surface as opposed to relying on surface mount thermal bonding. This could be cause for revisiting the diamond bog again!
(I do have a desire to try a slury of a UV curing adhesive mixed with diamond powder. Unlike epoxies, you wouldn't necessarily be wasting expensive goo as you would apply just the amount needed) I have had dialog now twice with a company who makes all kinds of LED related UV curing adhesives and even lens materials but in both cases, I was told they would get back to me and they never did. Being a small fish does have some disadvantages...OK, long winded post which is not uncommon for me. I will be working on the next wave of Sapphires soon and my plan is to build at least a few of these "mule" versions. I have no idea what the assembly time will be like or how easy or hard it will be to buld some of these. Only one way to find out. I also suspect that the interest in these may be greater than the utility they actually provide. Without secondary optics, the Sapphire Mule doesn't reach out very far!! It is great in room bounce and up close illumination but as an outside light, the limitations are rather obvious! On that note, I don't think I have ever mentioned that in a pinch, you can hold a 1/2" diameter acrylic ball in front of any of the Sapphire versions and get some extended throw our of them! The ball seats cleanly down in the chamfer of the head and I have played with the idea of a molded ruber or silicone sleeve that would retain the ball and snap over the head of the light; a beam shaper if you will but not like the ones we are presently familiar with.
If you have a Sapphire and a 1/2" acrylic ball, check it out.
While throwing out ideas and thoughts, I'll add a few more.
I have also built a prototype Sapphire using the high CRI LED behind a reflector. I have some 8 mm reflectors from a few years back and I took one and chopped it down so that I could fit it into the Sapphire head without restricting the light exit.
It is in the middle below with the 3 mm "stock" sapphire on the left and mule on its right:
Beam shot exposed to show the mule's contribution (middle light and beam diameter covers both its neighbors:
Beam shot exposed to better represent comparison between "stock" sapphire and reflector version of 036:
I find the idea of using the 036 behind a reflector intriguing and this is the first time I have used a High CRI LED behind a reflector with results I find worthy of further investigation. The tint variation from spill to spot with the Nichia High CRI is much less obvious than what I found with the Seoul LED's I sampled. On the bench, I coupled the 036 with a full version of the 8 mm reflector and varied the drive current up towards the max rating with very cool results. I think the smaller and restricted reflector along with the lower drive level of the LED in the sapphire sample above does not do the LED justice in terms of a collimated beam application. Multi levels with higher drive seems an obvious direction for this. I also would like to come up with an appropriate driver to push a 7UP cluster of the 8 mm Reflectors with 036 High CRI behind them and a host of appropriate size and configuration. This is all speculation and just ideas here.
Many of you have plans to mod the sapphires so in that spirit I have included some of these ideas and thoughts. Like I said, I do plan to make some of the mule version I have prototyped but that is subject to a reasonable assembly and time considerations yet known...
I may end up getting a sapphire after all 
