Titanium Sapphire "25"

[McGizmo]

Hi guys,
I figure it makes the most sense to start a new thread on the titanium Sapphire 25 instead of updating the original Sapphire Thread. I will borrow some images and probably some comments from there though. From the outside, the Sapphire 25 looks exactly the same as the Sapphire. The difference is in the converter and converter assembly. First some revisited pics:

The original Sapphire uses a converter I purchased from Arc Flashlights driving a Nichia 3 mm DS LED. The Sapphire 25 uses a constant current driver designed and manufactured for me in Japan by the same folks who have provided me with the 3S converters I am using in other lights currently.

The construction of the Sapphire Head is the same as the original as illustrated below:

The new converter is set at 25 mA constant current to the LED and the input voltage range for it is ~1V up to ~5V. This means that this converter will drive the LED at the target 25 mA whether regardless of the type of AAA battery used and it would even function properly on 2xAAA or a Li-Ion cell. The trace on both sides of the board is gold plated including any contact surfaces to enhance the ground path. I solder a .5 mm thick sterling silver ring on the PCB for the anode contact. The converter assembly (light engine) is retained in the head by virtue of an interference fit and seal around the 3 mm LED. It can easily be pulled out of the head if the need arise.

I decided to make my own PORON foam battery cushions and because they are hand punched, they each have their own unique eccentricities and form. My objection to the ones provided by ARC was in the large OD which put the foam into the thread area as well as beyond the PCB and onto the lip of the head. The cushions I have made and use here are smaller in diameter than the PCB and even those that have an obvious off axis cant (Leaning Tower of Pizza) do not interfere with the threads and remain attached to the backside of the PCB if removed from the head. A crude pic:

You will also notice that the LED does not sport a large cap as is the case with the ARC converter.

I should state that I have no issues with the ARC converter and for all I know it may be a more efficient converter. It certainly has a lot of history and a long track record! :thumbsup: My primary reasons for going with a new converter are the desire to have a source independent of another flashlight manufacturer, hopefully a converter less sensitive to contact resistance and a converter that is constant current across the voltage range of the various batteries, likely to be used.

I have been using a couple of these Sapphire's for a number of months now with pre-production converters and I am quite pleased with the performance of them.

For all intents and purposes, I consider this light to be a Sapphire but I feel I should make a distinction because of the new converter and since the new converter is set via a sense resistor to be at 25 mA I figure calling it a Sapphire 25 makes some sense. :shrug:

I also realize that since it is a new converter I should address the run time question. I did a runtime test in my integrating sphere with one of these and I used a Duracell alkaline AAA battery for the test:

I left the test run overnight having started it at about 10 AM and the next day, 22 hours later, the LED was still lit but at about .2 lumens or so. The important consideration was that the converter did not shut off but allowed the LED to carry on as best it could. You can see with the sample tested, the IS measured 5 lumens out the front. I don't know how accurate the IS is in testing the Sapphire but I am confident that the results are well with the realm of the population of these lights.

 

[Bullzeyebill]

Don, that is a nice mostly flat runtime there. Would you do a runtime with an Eneloop, or other LSD AAA? I can imagine that you might not want to monitor a NiMh AAA through its runtime to avoid over discharging.

Bill

 

[Haz]

That's a pretty nice runtime graph on alkaline batteries. I assume there will be longer and flatter runtime on lithium and rechargables.

 

[McGizmo]

Don, that is a nice mostly flat runtime there. Would you do a runtime with an Eneloop, or other LSD AAA? I can imagine that you might not want to monitor a NiMh AAA through its runtime to avoid over discharging.

Bill

Hi Bill,

My laptop was dedicated and away from the printers for almost 24 hours to create this one run time because I figured folks would want an idea of what to expect. I personally use the lithium AAA's in these lights. I don't have any rechargeable AAA batteries and candidly have no interest or desire in getting some or doing additional runtime tests. I would not be surprised if some of the variation seen in the first four hours of the graph isn't due to ambient temperature variations in the room where the test was conducted.

I leave run time tests to the experts.

 

[csshih]

If anyone every wants to get a runtime of this jewel when it's released, let me borrow it and I'd be happy to do so.

nice work! interesting to see the new converter

 

[fyrstormer]

What was the current setting on the original driver? Do you happen to know what the current rating is on the Arc Ti AAA as well?

 

[Henk_Lu]

That foam battery cushion looks very nice! In fact, the only thing I don't like on my Sapphire is teh battery rattle because of the missing foam...

I guess this cushion can serve on the "old" Sapphire as well? If so, I'll order one with my next light...

 

[McGizmo]

What was the current setting on the original driver? Do you happen to know what the current rating is on the Arc Ti AAA as well?

Others are better suited to answer this question. As I understand it, converter from ARC is not a constant current driver and the current to the LED is a function of V-in . I believe on a fresh battery the current could be as high as 60 mA but then drops as the voltage to converter drops.

To my knowledge, the present converter in the ARC AAA is the same converter used in the ARC Ti AAA and this is the converter I used in the Sapphires I have built, to this point. This converter has a long track record and many have experience with it in use.

 

[fyrstormer]

Oh, I see, you didn't just turn down current on the "stock" Sapphire driver, you actually replaced it with a different driver.

How high can the new driver you're using go? I assume 25mA isn't the highest setting, that's just what you specced...

 

[tygger]

Great work Don. Lost my original sapphire a few months ago. :sigh: Can't wait to pick one of these up.

 

[souptree]

So how does the 25 compare to the original Sapphire in terms of real world performance (brightness, beam shape, runtime, etc.) ? Any noteworthy distinctions?

 

[SFfanman]

So how does the 25 compare to the original Sapphire in terms of real world performance (brightness, beam shape, runtime, etc.) ? Any noteworthy distinctions?

-great question. I might add tint and dare I say, price?

 

[McGizmo]

There is an offering for these, HERE. The real world difference is only in output brightness. I am using the same LED's in these as the original Sapphire. I gave serious thought to getting some warmer tinted Nichias as well but for this type of light and based on some opinions I was given off the forum, I decided not to add a second flavor and complicate things at the inventory and assembly level. If it seems that there is enough demand or concern for a warm tint version, I might do a small run of them at some future time.

The converter actually has a tiny flea poop size sense resistor that sets the current level. I understand it has a max drive level of 40 mA.

 

[tygger]

Thanks Don. My keychain will soon be whole again. :twothumbs

 

[Oddjob]

Hey Don, any chance of converting a Sapphire to a Sapphire 25? I still get the occasional flickering on my Sapphire.

 

[McGizmo]

Oddjob,

If you are having issues with your Sapphire, please contact me and we can see about getting it sorted out. I have found that with some of them I needed to enhance the ground path contacts via the silver solder blob crimps. It is my hope that this new converter and its gold trace along with it's different circuit design will be less sensitive to contact resistance and such but the proof will be in the performance of these once they are out in the field.

 

[fyrstormer]

There is an offering for these, HERE. The real world difference is only in output brightness. I am using the same LED's in these as the original Sapphire. I gave serious thought to getting some warmer tinted Nichias as well but for this type of light and based on some opinions I was given off the forum, I decided not to add a second flavor and complicate things at the inventory and assembly level. If it seems that there is enough demand or concern for a warm tint version, I might do a small run of them at some future time.

The converter actually has a tiny flea poop size sense resistor that sets the current level. I understand it has a max drive level of 40 mA.

Hrm. Well, I see that most of the 5mm LEDs on LEDSupply are rated @ 20mA, so I guess that's actually a useful amount of current. I thought 5mm LEDs actually ran closer to 100mA; shows what I know.

Is your source for these capable of making small batches? I could probably do some interesting things with some of these drivers set a bit higher.

 

[paulr]

Wow, this is interesting, a buck-boost converter in a 1aaa light? I do believe from that graph that there may be some efficiency loss compared with the Arc converter, but the graph is certainly acceptable. One thing I wonder is whether the oscillator can start when the battery is very low, i.e. whether you can run the light til it's quite dim, then still be able to turn it on and off; also, whether it has the "recovery" characteristics of the Arc (run light til battery is dead, turn off, wait 5 minutes, and the light is good for another 1/2 hour of operation or so--repeat until bored).

I know Fred (Photon Fanatic) used some MJ converters in his TI Killer series, which he has unfortunately discontinued due to not liking the Nichia GS led.

How feasible is it to disassemble the Sapphire 25, if there got to be a reason for replacing the converter and/or led?

 

[fyrstormer]

Based on his own description, disassembly would consist of melting the solder in one of the epoxy fill holes, blowing it away with compressed air, and then using tiny pliers to hold the LE by the fill holes and then wiggling the LE out of the head.

 

[McGizmo]

Wow, this is interesting, a buck-boost converter in a 1aaa light? I do believe from that graph that there may be some efficiency loss compared with the Arc converter, but the graph is certainly acceptable. One thing I wonder is whether the oscillator can start when the battery is very low, i.e. whether you can run the light til it's quite dim, then still be able to turn it on and off; also, whether it has the "recovery" characteristics of the Arc (run light til battery is dead, turn off, wait 5 minutes, and the light is good for another 1/2 hour of operation or so--repeat until bored).

I know Fred (Photon Fanatic) used some MJ converters in his TI Killer series, which he has unfortunately discontinued due to not liking the Nichia GS led.

How feasible is it to disassemble the Sapphire 25, if there got to be a reason for replacing the converter and/or led?

I have no idea how this converter compares in terms of efficiency with the ARC converter. I know Peter told me that the ARC converter is quite efficient and more efficient than a constant current converter. I have a zip lock bag where I put my dead batteries and I just grabbed two Duracell AAA's out of it and I believe one was the one that I used in the run test above. Both fired up the LED in my Sapphire 25. No idea how long they would stay lit but they did fire up and not trivial in brightness.

The converter does behave as a buck/boost converter and on the bench, I found I could take it up to 5 V-in which got me thinking the converter could be used with a USB power source and possibly a keyboard light driver on a LapTop. I was told through translation by the designer that a 5V source could have spikes that would damage the converter and that the USB idea was not a good one. 3 AAA batteries in series though should be fine as well as a Li-Ion. With a max output current of 40 mA though, this particular driver is limited in scope when considering any scaling up and possibly using it with 1/2 watt or more powerful LED's. It was designed specifically for the Sapphire AAA format and single lamp style LED and it seems to be quite effective in its intended application.

The only thing holding these LE's in the head is the interference seal of the LED in the O-ring. You can carefully grab the foam cushion and pull the LE out of the head. If the foam cushion were to pull off the PCB then there is a small ground path via at the perimeter that you can put a needle in and use as a lifting pin by cocking it off angle and pulling it up. You could also remove the window and just push the LED out. These LE's are not potted nor are they staked with a solder blob as I did with the original Sapphires. LED replacement would be rather straight forward with the necessary skill and soldering equipment.