Build Log: Double Triple XP-G R5

[HarryN]

Interesting project - I like (and have built) some side x sides. They are much more challenging, aren;t they.

My suggestion is to skip O rings and seal things up with silicone. Its not like you are going to change LEDs every month, and when you do want an upgrade, it isn't that hard to dig out silicone based semi-permanent seals.

 

[sosemot728]

Interesting project - I like (and have built) some side x sides. They are much more challenging, aren;t they.

My suggestion is to skip O rings and seal things up with silicone. Its not like you are going to change LEDs every month, and when you do want an upgrade, it isn't that hard to dig out silicone based semi-permanent seals.

I wanted to try the o-ting route this time around. if it doesn't pan out, I will pursue something else.

 

[Linger]

O-rings - I like the image in post #14 better. You wouldn't even need the added metal lip containing the o-ring in a separate cavity, though you could include the slightest lip for peace of mind. If you can find ring of correct size a simple inner ledge will allow the ring to rest on it, downward pressure from the bezel will secure the o-ring between body and lens. Its nice a simple, two pieces squeezed together by water with o-ring in between.

Ideally for impact resistance your lens isn't compressed with anything metal at all, so double o-rings nice and tight will ensure nothing shatters if its dropped. Then you could ease off a bit on the lip on the bezel, no need to press it down with the bezel as the o-ring will hold it firm and water pressure make sure its sealed up tight.
That's really why I like the ring underneath - as pressure increases it closes more on the oring, instead of original proposal where pressure pushes lens away from o-ring

 

[paulr]

I mentioned on the other thread, it would be nice to be able to adjust the aim directions of the two clusters separately, so you could have a high and low beam like a car. But that only makes sense if there's a way to switch on just one cluster. At first I thought that would take two bflexes, but maybe it could be done with one bflex and some wiring to just switch one of the clusters out of the circuit.

 

[Geir68]

O-rings - I like the image in post #14 better.

I agree with Linger on the above. I would use O-rings between the body and lens like that first picture in post #14 illustrate.
In order to make the water pressure work against the O-rings, make a little gap between the bezel and body. The pressure at 30 meter depth will cause aprox. 40 kg additional force onto the O-rings. The deeper you go, the more force is applied to the sealing O-rings. I made a housing for a SLR camera once, and used this principle successfully.

Geir

 

[sosemot728]

I agree with Linger on the above. I would use O-rings between the body and lens like that first picture in post #14 illustrate.
In order to make the water pressure work against the O-rings, make a little gap between the bezel and body. The pressure at 30 meter depth will cause aprox. 40 kg additional force onto the O-rings. The deeper you go, the more force is applied to the sealing O-rings. I made a housing for a SLR camera once, and used this principle successfully.

Geir

This lights focus has changed from a dive light to a bike light, in order to keep it's small size. While I am waiting on this prototype run to be machined. I am working a a dive light version, that will be an all in one contained unit. It will only be a single triple tho. Pics soon.

 

[sosemot728]

Have updated some of the design for the production model.

Added double o-rings to front glass

Increased size of glass from 24.7mm to 26.8mm to suit.

Added rear wing reduction section.

Added smooth curve heat sinks

What does everyone thing of this new design changes? I think it is more 'organic' and less rough on the eyes.

 

[Geir68]

I really like it. Bike lights is what got me started in this hobby.

This is a bit massive for a bike light imo. I'm a weight weenie when it comes to things that bolt onto my bike. There is a lot of aluminum around the driver that could be removed. Can you calculate the weight(or volume) in Solidworks of your design?

Geir

 

[sosemot728]

The prototype is about 290g the new version above is about 20-30g less. There is not much room to move in terms of size. Because of the style of assembly... Screws

 

[sosemot728]

I have moved some stuff around. shortened it abit (at the expense of wiring room, it will be tight, but manageable)

Changed screw configuration of face

Changed size of rear heat sink wing (Increased... I was concerned about durability)

New version weighs in at 270g (which is now on the utmost limit) Including all electronics (except battery), Glass etc...

Pics below..... Comments??????

I want this light to be designed by this community, so any input would be much appreciated!

 

[John_Galt]

I really like it. Bike lights is what got me started in this hobby.

This is a bit massive for a bike light imo. I'm a weight weenie when it comes to things that bolt onto my bike. There is a lot of aluminum around the driver that could be removed. Can you calculate the weight(or volume) in Solidworks of your design?

Geir

Well, this has a lot of heat to disperse. Running 6 XP-G's at full tilt will produce some serious heat, so the more mass, the better, I say... Granted, he's not going to run it at full power continuously, but even still...

 

[sosemot728]

Decided on a switch type

http://au.farnell.com/jsp/level5/module.jsp?moduleId=en/508907.xml

All I need to decide is on what colour!? (I'm thinking red)

Here is the datasheet http://www.farnell.com/datasheets/35187.pdf

Any preferences?

Also here is the cable that will be included with the light.
http://au.farnell.com/jsp/search/productdetail.jsp?sku=1204333

Or this in electric blue....

http://www.newark.com/lapp-kabel/0012422/cable-ctrl-in-sf-5core-0-75mm-50m/dp/15P6036

and the cable gland
http://au.farnell.com/jsp/search/productdetail.jsp?sku=1204192

 

[Geir68]

Running 6 XP-G's at full tilt will produce some serious heat, so the more mass, the better, I say...

You need mass to transport heat from LEDs to the housing surface, but you need surface area to get rid of the heat.

matthewm!
I wouldn't change the astetics of the light. Its very clean looking.
You need to add a tapped hole for the mount.

I'm confused by your choice of switch. You are linking to a switch for mounting onto a printboard and its only dustproof. You need a watertight (IP67) momentary switch to control the bflex. Switches from APEM and ITW are often used by bike light builders. Example link
You probably need to move the beflex closer to the LEDs to make room for the switch and cabeling -> less mass ;-)

Geir

 

[sosemot728]

I wouldn't change the astetics of the light. Its very clean looking.
You need to add a tapped hole for the mount.

I'm confused by your choice of switch. You are linking to a switch for mounting onto a printboard and its only dustproof. You need a watertight (IP67) momentary switch to control the bflex.

Yes I specifically am looking for a PCB mount, it is more compact than having a separate LED indicator with tabs and just as nice to solder, the one you linked to is too large.

You need to add a tapped hole for the mount.

This is specifically not included in the machining to allow changes

 

[HarryN]

Hi, very approximately, it takes 1 sq in of surface area per watt to dissipate the heat if the air movement is static (no wind). 6 x LEDs = something close to 20 watts, so in theory, it will need 20 sq inches of surface when you are at a stop.

If the light is bike mounted and you are riding, then the optimal finning is front to back, not side to side. Obviously what you are have there is more aesthetic and is probably ok in a breeze in any event.

 

[sosemot728]

Hi, very approximately, it takes 1 sq in of surface area per watt to dissipate the heat if the air movement is static (no wind). 6 x LEDs = something close to 20 watts, so in theory, it will need 20 sq inches of surface when you are at a stop.

If the light is bike mounted and you are riding, then the optimal finning is front to back, not side to side. Obviously what you are have there is more aesthetic and is probably ok in a breeze in any event.

That is a VERY approximate guess... it also depends on many other factors. At the moment, excluding front and rear faces there is 13,500mm^2 of convective surface area (~20 inch^2)

for this case, the direction of the fins has a very small effect on the heat sinking capacity. You are right, I am relying on forced convection to remove heat from the device, but in terms of fin direction. The orientation used now would be slightly better (Relying on the structure orientation to create turbulence within the flow field as opposed to laminar plate flow over "aerodynamic" fins (front to back finning). This is what some of my original CFD analysis has pointed too. But as mentioned previously, the difference is minor. Greater benefit is achieved through a small surface area increase.

The sinking surface area on this design is already quite large... I do not doubt the adequacy of heat sinking.

 

[HarryN]

LOL - I guess it is not so terribly far off if you came to the same number using a computer model as my back of the envelope paper analysis. Yes, there are a lot of assumptions in the model, but as you note, it is a good starting point approximation.

I have done a fair amount of thermal analysis, but that was in the day that we did it by hand and a calculator, not these slick modern software simulations.

You might be interested that the 1 watt / in2 number matches a quite detailed thermal model that Lumileds has in one of their older data sheets.

If you don't mind my asking:
- What reynolds numbers are you coming up with for the two situations for the fin orientation?
- Was there much difference in temperature profiles for the two situations?

Interesting project - similar to my project "Tornado" concept in some ways.

Good luck with it.

 

[sosemot728]

If you don't mind my asking:
- What reynolds numbers are you coming up with for the two situations for the fin orientation?
- Was there much difference in temperature profiles for the two situations?

I did both some quick hand calc's for convective heat conduction (ignoring radiation to external surfaces and conduction to the front and rear faces) and an ANSYS CFX analysis. The hand calcs were done using an annular fin approximation assuming turbulent flow (Re~10^5). The CFX analysis confirmed heat transfer coefficients in a similar range. I did not find much difference in the temperature profiles (except the obvious change in streamline orientation/direction) and average temperature difference between the two was only 2~3 Degrees C

 

[sosemot728]

I have had a few concerns in my head over the current design that I am not sure about. Specifically how the glass lens is held in place. I am concerned that it will be cracked by the inside o-ring holders, if too much screw force is applied. I could reduce their height, but this would change how it is supported.

Here are some photo's







Also does anyone have some cad models of the Bflex and Maxflex? I am trying to position a header section mount for the Maxflex, I do not have the latest version here yet.

 

[sosemot728]

Update: Added mounting and cooling plate for Taskled's Maxflex driver. Now light has the provision to be run on either driver, Thru the use of double sided thermal tape

Update: Increased side thickness for mounting of mount hole. Changed rear plate to suit. (No net weight Loss/Gain)