As requested, 3 XM-L Driver/Emitter module DIY

469 posts / 0 new
Last post
NightCrawl
Offline
Last seen: 5 years 10 months ago
Joined: 01/22/2012 - 08:20
Posts: 3071
Location: Karlsruhe, Germany

Still more efficient than any other lightsource iirc..

texaspyro
Offline
Last seen: 2 years 6 months ago
Joined: 04/29/2011 - 12:43
Posts: 4593
PilotPTK wrote:
[Royal blue, for instance, is quite a bit more efficient than white.

Underneath the skin of a white LED is a blue led. They use a phosphor to convert the blue light into white… and, like everything in the universe: a) You can’t get something for nothing. b) You can’t even break even. c) You will die trying. (The three laws of thermodynamics)

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA
texaspyro wrote:
PilotPTK wrote:
[Royal blue, for instance, is quite a bit more efficient than white.

Underneath the skin of a white LED is a blue led. They use a phosphor to convert the blue light into white… and, like everything in the universe: a) You can’t get something for nothing. b) You can’t even break even. c) You will die trying. (The three laws of thermodynamics)

Correct. Hence the reason I speciffically used Royal Blue as the example. The laws of physics demand that the blue must be more efficient than the white since a _conversion_(energy loss) has to take place in order to make white out of the blue.

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

scaru
scaru's picture
Offline
Last seen: 5 years 1 month ago
Joined: 03/22/2012 - 13:36
Posts: 6946
Location: Virginia

PilotPTK wrote:
texaspyro wrote:
PilotPTK wrote:
[Royal blue, for instance, is quite a bit more efficient than white.
Underneath the skin of a white LED is a blue led. They use a phosphor to convert the blue light into white... and, like everything in the universe: a) You can't get something for nothing. b) You can't even break even. c) You will die trying. (The three laws of thermodynamics)
Correct. Hence the reason I speciffically used Royal Blue as the example. The laws of physics demand that the blue *_must_* be more efficient than the white since a _conversion_(energy loss) has to take place in order to make white out of the blue.

So would a blue CFL be more efficient or not since they normally put out white light? 

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA
scaru wrote:

PilotPTK wrote:
texaspyro wrote:
PilotPTK wrote:
[Royal blue, for instance, is quite a bit more efficient than white.
Underneath the skin of a white LED is a blue led. They use a phosphor to convert the blue light into white… and, like everything in the universe: a) You can’t get something for nothing. b) You can’t even break even. c) You will die trying. (The three laws of thermodynamics)
Correct. Hence the reason I speciffically used Royal Blue as the example. The laws of physics demand that the blue must be more efficient than the white since a _conversion_(energy loss) has to take place in order to make white out of the blue.

So would a blue CFL be more efficient or not since they normally put out white light? 

Maybe – i’m not sure. But if it were more (or less) efficient, it wouldn’t be for the same reasons. CCFLs naturally create a white’ish light – there is no phosphor conversion taking place. I strongly suspect that if there is a gas that creates natural blue light it would be very similar in efficiency to a white ccfl – but that is no more than a guess. Either way, what I can say for sure is that no matter the answer, the reason for the difference (if any) would be a completely different one than in the white LED situation.

PPtk

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

texaspyro
Offline
Last seen: 2 years 6 months ago
Joined: 04/29/2011 - 12:43
Posts: 4593

Actually fluorescents do have a phosphor conversion. That tasty mercury filling makes ultraviolet light that excites the phosphor. Leakage of UV light from CFL tubes can be a problem. It can cause degradation/fading of pigments, paints, and dyes. CFL tubes that emit minimal UV light have extra filtering built in.

And there are ways of upconverting long wavelengths of light into shorter wavelengths. Most cheap green lasers work on this principle. The output of an 800 nm laser diode is down converted 2:1 to form 1600 nm light in a non-linear crystal. That 1600nm light is then upconverted 3:1 into 533 nm green light using a different non-linear crystal. The overall efficiency is very poor and the required alignment and temperature stability of the crystals is very critical.

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA

texaspyro wrote:
Actually fluorescents do have a phosphor conversion. That tasty mercury filling makes ultraviolet light that excites the phosphor. Leakage of UV light from CFL tubes can be a problem. It can cause degradation/fading of pigments, paints, and dyes. CFL tubes that emit minimal UV light have extra filtering built in.

And there are ways of upconverting long wavelengths of light into shorter wavelengths. Most cheap green lasers work on this principle. The output of an 800 nm laser diode is down converted 2:1 to form 1600 nm light in a non-linear crystal. That 1600nm light is then upconverted 3:1 into 533 nm green light using a different non-linear crystal. The overall efficiency is very poor and the required alignment and temperature stability of the crystals is very critical.

You learn something new every day. I’m quite familiar with laser conversion – I work with lasers a lot.. CCFL, however, I’m no expert on at all. I didn’t realize that UV was the native output. Cool.

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

Piesoup
Offline
Last seen: 11 months 3 weeks ago
Joined: 05/05/2012 - 12:16
Posts: 49
Location: UK

Here is a housing I have drawn up in Solidworks. It weighs 90g and can easily dissipate the 30W of heat when riding at 12mph.

Rendering takes ages so there is actually one more shallow fin which you can see in the following pics. Simple little adjustments meant the stabilised temperature fell from a hot 114C to 65C. A chamfer here, extra depth groove there makes all the difference. I found the 'firewall' as I call it, the depth of solid metal behind the LEDs, to be at the mose efficient at no more that 4.5mm. Any more and the weight went up with no noticeable improvement in cooling.

The test was done with an initial solid temp of 60C to speed up the calculations. Air speed at 12mph, my normal riding speed when needing 3000 lumens! Air temp at 20C, a little hopeful here in England. Three LEDs on a copper disc simulating roughly Pilots module, each at 10W.

More pics to follow...

Piesoup
Offline
Last seen: 11 months 3 weeks ago
Joined: 05/05/2012 - 12:16
Posts: 49
Location: UK

Piesoup
Offline
Last seen: 11 months 3 weeks ago
Joined: 05/05/2012 - 12:16
Posts: 49
Location: UK

Here is the side view. Note the extra shallow groove a third of the way in from the left. It's not shown on my first render. The hole on the the thicker portion is for mounting. There is one the other side too as it will be mounted with a cradle held on by the bolts on my stem/face plate. 

The scalloped portions seem to help with the turbulence, breaking up the boundary layer to draw the heat away. Without them, the air seemed to flow over the fins. 

I know it would be much more efficient to have the fins running into the air flow but there is no way I'm going to be able to easily do that on my manual lathe! The model said it would stabilise at 34C though. 

There is a hole at the back, either for a plug of a cable gland. I'll have the switch on a fly lead. 

Piesoup
Offline
Last seen: 11 months 3 weeks ago
Joined: 05/05/2012 - 12:16
Posts: 49
Location: UK

Here is the graph. I didn't run a transient (timed) test at my pc would have thrown its toys out of the cot, but you can still see the rise of the LEDs in red against the max temp of the housing in green. Once the templevelled out, the test stopped. Like I said earlier, I started at 60C to speed up the calculations. With a fine mesh setting and narrow channel refinement it could take upto 2 days to complete! I placed a local mesh at the front as I noticed that portion played the biggest part in getting rid of the heat. The rear section had a rough mesh to speed up things. 

NightCrawl
Offline
Last seen: 5 years 10 months ago
Joined: 01/22/2012 - 08:20
Posts: 3071
Location: Karlsruhe, Germany

Its just plain awesome what today's software can do in gifted hands..

wquiles
Offline
Last seen: 4 years 11 months ago
Joined: 04/07/2012 - 09:09
Posts: 79
Location: United States

PilotPTK – I know I am late to this thread, but that is a very nice design. Great job with the thermal vias and the large copper areas for thermal transfer.

I really liked the small oval metal “posts” that solder to the board, yet by being hollow make it easy to attach a test probe – where can I find those? Who makes them?

texaspyro wrote:
If I need a piece of equipment,  I can almost always buy it off of Ebay for less than I can rent it for a month.

Same here. I try/look for it first on Ebay :bigsmile:

Will

Custom machining and electronics - www.atdms.com  Email: wquiles (at) gmail [dot] com

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA

wquiles wrote:
PilotPTK – I know I am late to this thread, but that is a very nice design. Great job with the thermal vias and the large copper areas for thermal transfer.

I really liked the small oval metal “posts” that solder to the board, yet by being hollow make it easy to attach a test probe – where can I find those? Who makes them?

Thanks Will – A lot of time and thought went into the design, and I’m pretty happy with it. The software/firmware is just as neat as the hardware though.

The test points are awesome. I use them for practically everything. Small enough to not be a real-estate burden, big enough to probe and super easy to attach test-leads to. They’re keystone part number 5015
http://www.keyelco.com/products/specs/spec130.asp

Mouser usually has them in stock, and I’m sure other places as well.
http://www.mouser.com/ProductDetail/Keystone-Electronics/5015/?qs=qOqV1E...

I order them by the thousand, because I really do put them on everything. One board I designed that’s full of analog and needs to be calibrated has over 800 of them on it. One Board. Ugly. Takes a good tech about 2 days to calibrate it. Couldn’t imagine doing it without good test-points.

PPtk

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

wquiles
Offline
Last seen: 4 years 11 months ago
Joined: 04/07/2012 - 09:09
Posts: 79
Location: United States

PilotPTK wrote:
wquiles wrote:
PilotPTK – I know I am late to this thread, but that is a very nice design. Great job with the thermal vias and the large copper areas for thermal transfer.

I really liked the small oval metal “posts” that solder to the board, yet by being hollow make it easy to attach a test probe – where can I find those? Who makes them?

Thanks Will – A lot of time and thought went into the design, and I’m pretty happy with it. The software/firmware is just as neat as the hardware though.

The test points are awesome. I use them for practically everything. Small enough to not be a real-estate burden, big enough to probe and super easy to attach test-leads to. They’re keystone part number 5015
http://www.keyelco.com/products/specs/spec130.asp

Mouser usually has them in stock, and I’m sure other places as well.
http://www.mouser.com/ProductDetail/Keystone-Electronics/5015/?qs=qOqV1E...

I order them by the thousand, because I really do put them on everything. One board I designed that’s full of analog and needs to be calibrated has over 800 of them on it. One Board. Ugly. Takes a good tech about 2 days to calibrate it. Couldn’t imagine doing it without good test-points.

PPtk

Awesome – thank you very much. The links you provided also have the SMT pad, so I will try to incorporate the pads in future boards :bigsmile:

Will

Custom machining and electronics - www.atdms.com  Email: wquiles (at) gmail [dot] com

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA

wquiles wrote:

Awesome – thank you very much. The links you provided also have the SMT pad, so I will try to incorporate the pads in future boards :bigsmile:

The keystone suggested pad is .135 × .070. I use a smaller pad, however, as I had trouble with the test points floating around too much.

My pad is:
Copper: .120 × .055
Mask: .125 × .060
Paste: .105 × .040

PPtk

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

wquiles
Offline
Last seen: 4 years 11 months ago
Joined: 04/07/2012 - 09:09
Posts: 79
Location: United States

PilotPTK wrote:
wquiles wrote:

Awesome – thank you very much. The links you provided also have the SMT pad, so I will try to incorporate the pads in future boards :bigsmile:

The keystone suggested pad is .135 × .070. I use a smaller pad, however, as I had trouble with the test points floating around too much.

My pad is:
Copper: .120 × .055
Mask: .125 × .060
Paste: .105 × .040

PPtk

Wow, thanks again man – much appretiated sharing the smaller footprint data. I use spring-loaded pogo pins for programming my Atmel Tiny processors, and for testing some small boards, but these SMT test points will come super handy as they free your hands and allow for a good/solid connection while testing/trying things out Big Smile

Will

Custom machining and electronics - www.atdms.com  Email: wquiles (at) gmail [dot] com

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA

For sure, the solid connections and free hands are awesome. Once you start using these things, it gets addicting – I put them on every signal line I think I might ever need/want to probe. They’re small enough that it doesn’t usually have any real effect on board size, and they are so nice to have when you need them.

The pogo pins are what I always use for programming headers

Unless I’m really tight for board space, then I use a tag-connect layout

PPtk

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

wquiles
Offline
Last seen: 4 years 11 months ago
Joined: 04/07/2012 - 09:09
Posts: 79
Location: United States

Very cool. My programming setup is not fancy, but works well for the relatively low volume of my projects. Here I am programming a Tiny85:

And here is the programmed board on its battery carrier (cover off in picture, but bolted in place for end customer use). On this photo you can see the 7 test/program test points (some with silkscreen labels):

Will

Custom machining and electronics - www.atdms.com  Email: wquiles (at) gmail [dot] com

zeremefico
zeremefico's picture
Offline
Last seen: 53 min 16 sec ago
Joined: 03/27/2012 - 02:44
Posts: 1391
Location: Greece

Hi, does anybody know a free or shareware program for 3d design?

₪₪₪₪ ΟΥΔΕΝ ΚΡΥΠΤΟΝ ΥΠΟ ΤΟΝ ΗΛΙΟ ₪₪₪₪

My YouTube channel

Flashlights & edc gear

K40M F16

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA

zeremefico,

Google Sketchup is actually not horrible for the cost of nothing..
http://sketchup.google.com/download/

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

zeremefico
zeremefico's picture
Offline
Last seen: 53 min 16 sec ago
Joined: 03/27/2012 - 02:44
Posts: 1391
Location: Greece

PilotPTK wrote:
zeremefico,

Google Sketchup is actually not horrible for the cost of nothing..
http://sketchup.google.com/download/


If you design your module with that, it is great for me!
Nice project PilotPTK Wink

₪₪₪₪ ΟΥΔΕΝ ΚΡΥΠΤΟΝ ΥΠΟ ΤΟΝ ΗΛΙΟ ₪₪₪₪

My YouTube channel

Flashlights & edc gear

K40M F16

JonnyC
JonnyC's picture
Offline
Last seen: 9 months 1 day ago
Joined: 01/14/2011 - 19:12
Posts: 1148
Location: Green Bay, WI - USA

Sorry I haven’t quite been keeping up with this thread, but where are things at? Pilot, you have some already produced (fully assembled), are we just waiting to see what hosts will work well for it?

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA
JonnyC wrote:
Sorry I haven’t quite been keeping up with this thread, but where are things at? Pilot, you have some already produced (fully assembled), are we just waiting to see what hosts will work well for it?

In terms of torches, yes – I still haven’t found a great host for it yet. A lot of people are inquiring about using these for non-torch applications though (bike-lights, off-road lights, you name it lights). I have some modules available right now for anyone who wants one – I haven’t announced major/general availability just yet because I’m still waiting for my shipment of U2 LEDs, and I don’t plan on running the big production run until I have everything (it costs big bucks to run on a U.S. production line for half a day).

Cool White T6 and Neutral T5 modules are available for purchase right now – I don’t have a lot of them though.

PPtk

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

JonnyC
JonnyC's picture
Offline
Last seen: 9 months 1 day ago
Joined: 01/14/2011 - 19:12
Posts: 1148
Location: Green Bay, WI - USA

There was discussion in this thread about using this module with a D Maglite. After seeing this thread (http://budgetlightforum.com/node/10341), I think it’s very feasible to fit this module and a heatsink into the top end of the Maglite with a bit of boring (just above the switch) The stock head could then be milled down, or a head like the one cut on the lathe in that thread could be used to hold the module and optics in place. If the momentary configuration is desired (I would want that), the module could be hard-wired and the stock switch could be converted to momentary action for mode changing (or the stock switch could just be replaced entirely). Along with cutting down the battery tube and either using 3 × 14500’s or have it bored for 3 × 18650’s, this could be a pretty compact powerhouse.

PPtk: Quick question – can the momentary switch be wired to pull up (to positive) instead of pull down (to negative)? That way the stock switch converted to momentary could work as-is.

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA
JonnyC wrote:
PPtk: Quick question – can the momentary switch be wired to pull up (to positive) instead of pull down (to negative)? That way the stock switch converted to momentary could work as-is.

Unfortunately, no. The switch must be wired between the Momentary Contact Input and Ground. It is pulled up on the circuit board, and the switch pulls it low to toggle an input. Pulling the Momentary Contact Input to Battery+ would be very bad, in fact.

PPtk

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

simplec6
Offline
Last seen: 6 years 5 months ago
Joined: 07/15/2011 - 19:52
Posts: 78
Location: US

I want three with U2 binned XMLs. What is the price of one of these?

tomloew
Offline
Last seen: 9 years 7 months ago
Joined: 06/23/2012 - 01:48
Posts: 1
Location: Chicago,IL

How much time is consumed in completing this extraordinary project?..I would be definitely learning something from you for sure !!! What are the prices?

My LED lights

tom S loew

raccoon city
raccoon city's picture
Online
Last seen: 2 min 13 sec ago
Joined: 10/06/2010 - 02:35
Posts: 17608
Location: रॅकून सिटी Palm Desert CA USA

Welcome to the site, tomloew!

Best0270

PilotPTK
Offline
Last seen: 7 years 11 months ago
Joined: 09/04/2011 - 21:53
Posts: 1736
Location: Michigan, USA
simplec6 wrote:
I want three with U2 binned XMLs. What is the price of one of these?

Hi simplec6,
Cool white U2 Flux 1C Tint: $93.00 plus shipping

Since you’re in the U.S. – Priority Mail would be about 6 bucks for all three.

I am currently extremely busy with work. Please do not expect a response from me quickly. I will be dropping in as time permits, but the amount of time I can dedicate to responding to topics and PMs is very limited.

Pages