Canister light build. Need input.

[kallehornstrom]

I might place an order for the 7-array and 4-5 loose xp-g's and see what works best. These are the lenses you (Packhorse) used right: http://www.dealextreme.com/details.dx/sku.13903
As for driver I'm thinking the RCD-24 could be a good choice. I really like the idea of dimming via a potentiometer. However it might not be mechanically feasible within the small confines of the housing. http://www.dealextreme.com/details.dx/sku.13903

BTW. Here's the light head. (Unfortunately it is in German)
http://shop.strato.de/epages/61162903.sf/de_DE/?ObjectPath=/Shops/61162903/Products/%22G%20LK%2050N%20AL%20(Set)%22&ViewAction=ViewProductDetailImage

 

[Packhorse]

Yeah those are the lenses.
Looking at the housing I see its 60mm ID but has a throat of 50.5mm that is going to limit the size of your heatsink lens assembly. Thats a pity cause 60mm would be great!

 

[kallehornstrom]

Yeah those are the lenses.
Looking at the housing I see its 60mm ID but has a throat of 50.5mm that is going to limit the size of your heatsink lens assembly.

Yea, this also makes the connection between heatsink and canister wall a somewhat tricky business.

Thats a pity cause 60mm would be great!

Did you have anything particular in mind?

 

[Packhorse]

60mm ID would easily allow the use of 7 aspherics. Or perhaps the use of 3 of Ahortons aspherics, maybe 4

 

[kallehornstrom]

I guess I'm gonna have to experiment a little to make this project come together. One thing that I haven't decided on yet is what battery pack to use. The canister I have in mind gives me roughly 55(dia)x280mm to work with: http://shop.strato.de/epages/611629... 60 250 AL"&ViewAction=ViewProductDetailImage
Using standard c-cells is not going to make for very efficient use of the space. I might actually go for AA's and end up with something like 28V.
Using LiIon's should give me more power but I'm put of by the relative complexity when it comes to dealing with these (Especially since I plan to solder the pack myself). My expereince with batterys (and LED's) are after all rather limited. If it wasn't for CPF I would be left completely clueless. lovecpf

 

[Fluor]

Its not just the ostar but all multi die emitters
I guess this is why I am such a fan of aspherics. But they too have their problems. Efficiency being a big one.
I built a IST aspheric for a customer 200-250 lumen (R2) he had access to the ability to measure the torch lumen and claimed it was only 90. He still claimed it was one of the best throwing lights he ever owned though.

Packhorse,

Do you know where the light is lost on the way? Is it when passing through the lens or not collecting all the light on the other side?

 

[jspeybro]

it's probably both.

Transmission of light through uncoated BK7 (optical quality glass) is around 93%. In case of acrylic lenses, this is around 90% so this introduces some losses. since the lens is placed behind a glass window, you will loose about 15-20% of the light that is incident on the lens.

also, since the emitter is not sending all its light towards the lens, not all emmitted light is incident on the lens so this is a second origin of losses but these are more difficult to estimate. You would need to know the light distribution (I think I've seen this info in Cree datasheets).

A sollution could be to use a reflector but that will most likely convert the light that was first lost into light that is spilled around the central spot, since the light will not come from the focal point of the lens but from somewhere else via reflections. Under water, this spilled light is most likely not noticable so this makes you wonder why you would even bother combining a lens with a reflector.

The better sollution would be to use lenses that you place over the LED to capture as much light as possible. These lenses are usually more complex in design (and more difficult to design such a lens that does the job well) and are usually made for specific LEDs and output angles.

Another option could be to use a larger lens, but that is usually not an option when modding lights.

Johan

 

[kallehornstrom]

I've had a look at the data sheet of the optics for the 7 xp-g board at Cutter. The beam is definitely not going to be narrow enough therefore the 40mm board is not an option for me any more.
I'm going to order 7 xp-g on 10mm board and 21mm aspherics from DX - cut them down to 18 mm (and they should fit). I also found another version of the housing that has no throut behind the glass which is going to make the heatsinking considerably easier.

 

[Packhorse]

it's probably both.....

Johan

Yeah, what he said.


Thats why I think the Ahorton lenses are so great. They are glass and (about) 30mm with a smaller focal length than the 21mm Dx one.

I would much prefer to use 4 of the Ahorton lenses than 7 DX's ones.

 

[wquiles]

I wonder when the XP-G's will ship?

Packhorse,

Somebody here just got their XP-G's, so it looks like they started shipping :thumbsup:

http://www.candlepowerforums.com/vb/showthread.php?t=239719


Will

 

[Packhorse]

I just got mine today!

These XP-G's are tiny as!

Anyway it didnt take me long to hook one up twith one of Ahortons aspherics.
I would like to say I was really really impressed! But that would be a lie. To be honest the XR-E's perform better. My guess is that the XP-Gs have a wider dispersion angle and less light is gathered by the lens.

 

[jspeybro]

using a bigger die but keeping the primary optic the same size... not the best idea I would say, but that depends on what you expect the led to do.
We want to focus as much light as we can so a bigger primary optic would probably be a better solution.

When you say 'perform better', do you mean that the XR-E's can be better focussed or do you mean the light output, hotspot vs spil, ...?

Johan

 

[Packhorse]

They both focused to the same size beam which confused me some what. I was expecting the XP-G to be larger. While the XP-G R4 was a slightly warmer colour than the XR-E R2 it wasnt quite as bright. I then tried the R4 in a reflector and it seemed to work really well although I didnt have a R2 in a reflector to comare it against.

 

[kallehornstrom]

I haven't gotten around to order the XPG's yet but it's nice to see that they are shipping. Has anyone gotten the R5?
Anyhow, I have the canister/light head now. They seem extremely well built with a rotary magnet on the head with lathes every quarter turn. I'm going to try to use four (mini) magnetic reed switches, one NC and the rest NO, wired to resistors to provide appropriate control voltage to the RCD24. Thus giving me the option of three current settings.

The next step will be to determine the total length of the lenses-led's-heatsink and pcb assembly. What is the focal lenght of the 21mm aspherics from DX?
Then I need to find someone/somewhere to machine the heatsink which needs to be hollowed out to fit the pcb.

 

[wquiles]

Its not just the ostar but all multi die emitters that have this problem.
And as Will points out to get a decent tight focus you need a large reflector.
Thats not to say you cant get a decent result with a multi die and a large reflector or a singe die and a small reflector, its just not optimal.

I guess this is why I am such a fan of aspherics. But they too have their problems. Efficiency being a big one.
I built a IST aspheric for a customer 200-250 lumen (R2) he had access to the ability to measure the torch lumen and claimed it was only 90. He still claimed it was one of the best throwing lights he ever owned though.

There are now a few new LED's on the market since I started building my aspheric based can lights. XP-G and SST-50 come to mind.
Problem with these is the die of the LED is larger so the beam will be wider. With careful selection of apherics it may not be a problem though.

Im not saying you should build an aspheric based light but you should consider it.

Packhorse,

I started playing a little bit my SST-50 Emiter - the die is HUGE!:
link to LED sub-forum's post ...

This 2" dia glass Aspheric lens is likely not a perfect Aspheric for the SST-50, but there is no doubt that the larger die will make it harder to get a super narrow beam. This particular sample is an unkown bin, but it is still very bright at 5Amps, so if we assume 800-900 lumens, the question is, would something this "narrow" but very bright still be useful for a diving light?

Will

 

[Packhorse]

With a aspheric I think a 10 degree beam is pretty good. That is all all the light with in 10 degrees.
I have though about using a SST-50 and I think it could be good when used with an aspheric. Probably the best candidate for a adjustable focus LED based light.
At 800-900 lumen at 5 amp its probably not the most efficient LED out there but I think this could be over looked if it can be used with a 30mm aspheric ( like Ahortons) and get a 10 degree beam. Which sounds like its possible. Would make for a very small light head although at 5 amp a separate battery cannister would be a must.


If anyone is keen to send me a SST-50 for a test I am more than happy to post the results!

 

[kallehornstrom]

Packhorse, could you please tell me the focal lenght of the 21mm lenses from DX? ( http://www.dealextreme.com/details.dx/sku.13903 )

Thank you for the valuable advises I've gotten so far.

 

[wquiles]

With a aspheric I think a 10 degree beam is pretty good. That is all all the light with in 10 degrees.
I have though about using a SST-50 and I think it could be good when used with an aspheric. Probably the best candidate for a adjustable focus LED based light.
At 800-900 lumen at 5 amp its probably not the most efficient LED out there but I think this could be over looked if it can be used with a 30mm aspheric ( like Ahortons) and get a 10 degree beam. Which sounds like its possible. Would make for a very small light head although at 5 amp a separate battery cannister would be a must.


If anyone is keen to send me a SST-50 for a test I am more than happy to post the results!

Cool - I really like the idea of an adjustable beam/aspheric combo :devil:

This particular unknown bin SST-50 sample has a vf of 3.39V at 5 Amps, which is actually excellent since the typical value on the data sheet was 3.6V at 5 Amps. The 800-900 is just a conservative number I came up with - depending on the actual bin, this LED could be producing between 800 and 1035 lumens. The output is significantly brighter than the P7 lights that I have, so it might actually be pushing closer to 1000 lumens. I am tied up the next 4-5 days, but next week I hope to do some outdoor beamshots to start looking at the SST-50 and the relative output power in more detail.

You are also right on the power side -> 4-5 Amps is tough on a hand-held light. A canister might be the ticket here. We would of course have to come up with a driver(s) that can safely do 4-5 amps for extended continuous runs without burning up.

I only have this one SST-50 sample, and it is already thermally glued to that heatsink, otherwise I would have send you this sample to play with. He is currently sold out, but I got mine from PhotonFanatic (Fred) here:
link to sale thread ...

Will

 

[Packhorse]

Packhorse, could you please tell me the focal lenght of the 21mm lenses from DX? ( http://www.dealextreme.com/details.dx/sku.13903 )

Thank you for the valuable advises I've gotten so far.

I dont know what the focal length is.

What I can tell you is that I use the 16mm star Q5's from DX which I guess is 1 mm thick. The spacer I use is 11mm tall. So if you dont use a star then the spacer will be 10 mm.

This does not give a perfect focused beam. The spacer will need to be slightly taller. Or by making it shorter you can get a wider beam. 1mm difference will change the beam alot. You will need to experiment a little to find the beam that best suits you.

Remember that the fatter side of the lens points away from the LED.

 

[Packhorse]

Will, when you say the SST is brighter than the P7 is this at the same current input? If so then an W200 aspheric SST50 at 2.8 amp could be a nice light.