Custom made SSC P7 LED Bike lights

[EL34]

I have been prototyping some new super light weight and small 900 lumen P7 lights. I have tested several of the P7 lights and they are blowing away my HID's and so the HID's are going on Ebay for sale.

I am thinking of having a bunch of the parts made and then offer the parts as a do it yourself type thing with web site instructions on how to do it.

I don't have time to assemble lights and I would like to keep the prices as low as possible on the parts so people can get into high performance lights for way less than the silly prices being asked by some of the commercial light companies.

Anywho, I haven't decided if I am going to have the parts made yet. I would appreciate feedback on the design and if there is enough interest, I will have a machine shop run off 100 or so and see how it goes.
I already have a web store for my main business, so the parts sales aspect is already in place.

This link takes you to my bike lights page, the new light link is at the top of this page.
http://www.el34world.com/Misc/bike/BikesLights1.htm

I am still working on a cheap Li-On charger design that will charge battery packs, not individual batteries.

Thanks for any feedback.

Here's a shot of a scratched up prototype, just to give you an image.

 

[LEDobsession]

Where's the cooling fins? You will definitely want some of those.

 

[EL34]

No need for fins so far like on thick bodied lights I own.

The super thin skin sheds heat way faster than my thicker skinned/finned light experiments.

Also, on bike lights, the air is always moving and removing heat. Not so with a flashlight type light.

Think of it this way regarding the skin of the light.

If you put a 1/2" thick piece of aluminum in an oven, heat it up, and then try to hold it in your hand. You will burn your hand and the thick slab will take a long time to cool down enough to hold it in your bare hand.

Now put a piece of aluminum foil in an oven, take it out. You can hold the aluminum foil in your hand within a second or two.

The foil will act as a great heat sink and will dissapate the heat way faster.

I have tested several types of lights and the flashlight P7 mods I did were the only lights that needed fins because the casing was way too thick.
You need a thick casing on a flashlight because it has to be strong enough to club someone over the head.

You do not need a thick casing on a bike light. In fact, the thinner the better, until you reach the point where the structure is in jeopary of being way too thin to support itself, or it dents too easily.

 

[Ryan7257]

I like it, nice weight too, and even less with an aluminum bolt.

 

[znomit]

If you put a 1/2" thick piece of aluminum in an oven, heat it up, and then try to hold it in your hand. You will burn your hand and the thick slab will take a long time to cool down enough to hold it in your bare hand.

You hand is measuring how much heat you have removed from the oven.
The foil has removed very little heat, cools immediately. The bar removes a lot of heat... ouch.

You want enough mass in the light so it doesn't overheat when you're changing a tyre for 10 minutes(if using your single mode driver). You want enough mass in the exterior that it heats up uniformly. You want enough surface area to allow transfer of the heat from the light to the air.

 

[EL34]

Ryan, yes, the bolt is a big steel bolt from the Marwi mount.
Almost 1/4" in diameter
It does not need that big steel,bolt for sure.

It is just a tiny little light, it just needs a small mounting bolt.
An Aluminum screw would be great.

You could even use a nylon screw.

 

[EL34]

I have been testing a newly completed light today.

Had a couple of design changes and things are looking good. No thanks to my Chinese supplier. You can never tell what you will receive, but the prices are cheap and so you have to put up with it.

I have black high temp plastic on the way to make black mounts. The white mount was just material I had on hand here and so I used that as a test.

The heat test shows that this light runs cooler than any of my other lights. I have spent a bunch of time getting all of the aluminum parts thin and in contact with each other so that the heat is sinked away and out to the outer skin. The reflector is even parts of the outer skin and is a heat sink. The entire light is a heat sink from tip to tail.

You can hold this light in your hand without any discomfort at all.
not so with some of my other LED lights that were converted from thick flashlight bodies.

I also have some black nylon 1/4 x 20 threaded rod and black nylon nuts on the way. The nylon is going to be replacing that big 1/4" steel bolt on the Marwi mounts.

 

[Scoot]

Hi all... new here... more to tell later later about my recent adventure here on this awesome forum:wave::candle:

I am convinced that there is strong merit to the heat management principles applied here. Do you know how much an anodize layer might interfere with the heat transfer? I love the seeming "floodiness" of the beam shot picture.:twothumbs See how my vocabulary has improved already?:sweat:

 

[EL34]

The main body could be anodized, but not the reflector.

You have to dip the parts in several nasty solutions to do anodizing.

The shiny reflector surface would be harmed and there is no way that I know of to anodize the outside, but not the inside.

 

[georges80]

I'd highly recommend you put a thermocouple on the LED and measure how hot it is getting versus measuring the body of the light with your hand.

You have a reasonable amount of surface area, but you still have about 10W of heat from the LED plus more from the shunt regulators (how much depends on your input voltage/battery versus LED Vf). Given this is a P7 light and you are using shunt regulators I'm presuming your likely battery source is a single li-ion cell or a stack of parallel li-ion cells.

Surface area is what counts when it comes to dissipating heat to the outside air - not how thin you have made the housing. Surface area is achieved either by a larger housing or my putting fins/grooves in a smaller housing (to create more surface area). Anodizing the outside of the housing black would also radiate more heat - given the light would be used in moderate/full darkness.

cheers,
george.

 

[EL34]

The LED is screwed down to a capsule that is in direct contact with the outer skin and so, what ever the temp is on the outer skin nearest the LED capsule is accurate enough indication of how well the heat is being pulled away from the LED and being distributed to the outer skin. There is no air gap at all between the LED capsule and the outer skin. I am also using thermal grease between the LED and the capsule.

The inner area of the main body is machined with two diameters that match the LED capsule and the rear tail cap exactly. Surface 1 and surface 2 in the photo below.

The heat is being sinked to the outer skin very efficiently. The LED capsule is even in contact with the rear body switch cap piece.

I designed the machined parts to all be touching each other so that they all sink the heat and share in the dissapation.

Unlinke the P7 flashlight I got from DX, which has a small pill with two threads that barely makes contact with an inner cylinder. Very poor design and it still works, but it get extremely hot.

It would not make sense to measure the LED capsule heat by itself. That would be like taking the fan blade off a car to see how hot it gets. All three alumnum pieces in the photo below are engineered to work together as one. The reflector is also part of the whole outer skin and also adds to the heat dissipation since it is also aluminum.

 

[znomit]

It would not make sense to measure the LED capsule heat by itself. That would be like taking the fan blade off a car to see how hot it gets. All three alumnum pieces in the photo below are engineered to work together as one. The reflector is also part of the whole outer skin and also adds to the heat dissipation since it is also aluminum.

A good test is to run the light for a good while and then pull it apart and see how hot the MCPCB or even surface 1 is. If the light isn't hot after 10 minutes sitting on the bench your driver might not be giving the full 2.8A, some of the DX, Kai ones aren't well regulated. Have you checked the current output?

 

[EL34]

I understand that heat can be a problem.

A little background info:

I have been building lights and battery packs for many years going back to Halogen systems.
I own a company that deals in electronic items and have engineered many a product since about 1993.

I have already run through many heat test with several of these lights.

I own many bike lights of different types and I can tell you this, they operate way cooler than any of my other LED conversions or HID's.

I have let them run for hours just sitting with no cooling fan just to get a firm grasp of how well they shed heat. Heat has been an issue

I am not sure how to get this point across without some sort of industry standard measurement procedure that can be taken and then compared with other P7 lights running at 2.8 amps.

The operating current is 2.8amps, see this page for more info
http://www.el34world.com/Misc/bike/BikesLights20.htm
and this page
http://www.el34world.com/Misc/bike/BikesLights21.htm

 

[georges80]

Um.... I'm an engineer too and have been designing driver boards for LEDs for a bunch of years now. I can safely say I started designing driver boards just as the 1W Luxeons first came to market.

You DO need to measure the LED temperature. It really doesn't matter how good you 'think' your thermal path is, it is the temperature at the LED that counts, especially in a light that isn't just turned on for a few seconds/minutes like a torch/flashlight.

Your analogy of removing the fan blade makes no sense, unless you don't understand my point. I'm not saying to take the 'pill' assembly out to measure the temperature, I'm saying you need to get a thermocouple INTO the housing in contact with the LED and run the entire unit as an assembly while monitoring the temperature. Until you do that you are just assuming the temperature is 'cooler' because of touching the outer casing. You need to run the test for an extended period of time to determine the steady state temperature of the LED. Run it with no airflow and then with airflow.

If you want to be scientific rather than 'touchy/feely' you need real numbers from real measurements. Most of the folk in the subforum have custom built or modified many high powered bike lights, not just a single P7 and have a quite firm grasp and experience of thermal issues of these kinds of designs.

Given that you are using a shunt regulator your battery choice is very specific to reduce heat losses in the driver. The driver has no thermal management so you are relying purely on the heatsink provided by the housing. You DO need to cater for the light being on for a period of time with no airflow - either because you are stopped or a 'friend' clicked on your light to check it out and 'forgot' to turn it off. BOTH scenarios are possible and happen.

From your pictures, your shunt regulator boards are sandwiched between insulators (both electrical and heat wise), they will get extremely hot if your battery choice is not good. Say you have a Vf of 3.3V for the P7 (not unheard of) and you have a fresh li-ion at 4.2V. Now you have 4.2-3.3 = 0.9V x 2.8A of heat loss in the regulator boards. That's 2.5W that will cook those boards until the li-ion battery voltage drops down to a lower level.

Anyhow, I'll REPEAT again, the ability of a heatsink to shed heat is proportional to surface area, NOT thickness. Thickness provides thermal mass which affects how long it takes to heat up or cool down.

I'm not trying to shoot you down here and it's neat that you are looking to put together a kit of bits for folk to build their own lights. What some of us in this thread (with a LOT of bike light design experience) are trying to do is to help point out some issues that if addressed would improve your design.

cheers,
george.

 

[EL34]

I have read all the data sheets on the SSC P7 on the Seoul Semi web site.
I have run my lights out of the case and tested them.

If you would like to see some sort of camparitive results, I will need some sort of test results and standards to compare my results with.

What will we be comparing this against, other lights?

Exactly what results are you looking for?
Give me exact figures I can compare my measurements against please.

Not trying to be a pain, but in order to give you any sort of meaningfull bench test results, you will have to tell me what the standard is I am testing against.

 

[georges80]

Very simple - MEASURE the temperature of the LED (i.e. where the LED attaches to the MCPCB). That is all we need to know if the LED reaches and maintains steady state temperature.

Again, have you considered the heat losses in the shunt regulators and how they will cook - what battery pack/configuration are you suggesting should be used?

It helps in a thread if you address all questions (if possible), rather than a selective few...

cheers,
george.

 

[EL34]

Yes, out of the case, it runs and stays at a constant temp once it warms up.

My web site has info on the 1400ma boards I am using, betteries, chargers, etc.

You can use any 3.6 v configuration you like.that's a personal choice based on how mucj run time you need.

Personally, I like 3 to 4 x 2400 mah batteries in my LED battery packs for a 2-3 hour burn time.

I run the lights on a low setting while climbing up the mountains and use a 2.8 amps setting for the ripping downhills.

See the web site for more info on that.
http://www.el34world.com/Misc/bike/BikesLights1.htm

 

[georges80]

Ok - I guess we aren't going to get numbers.

And if you are saying the pill runs stable temperature ("out of the case") at 2.8A output I'd be VERY surprised (my nice way of saying skeptical).

Ok, parallel li-ion cells. Fresh off the charger you are going to have some serious high temperatures occurring in those shunt regulators if you have low Vf P7's.

Yes, I'm aware of the boards you are using, they are shunt regulators, any extra 'voltage' is converted to heat.

Anyhow, good luck with your project.

cheers,
george.

 

[znomit]

The operating current is 2.8amps, see this page for more info
http://www.el34world.com/Misc/bike/BikesLights20.htm
and this page
http://www.el34world.com/Misc/bike/BikesLights21.htm

Is that the led current or the battery current?

 

[EL34]

George, if you looked at the links I provided it clearly shows I am using two of the DX 1400ma boards in parallel with 4 x 7135 current regulators on each board . I am not interested in running around and wasting time giving some sort of heat measurement, it does not matter. Please just forget about this post and move on, thanks.

For those that are interested, I added some pics to the web site that shows size and weight comparisons between several of my other lights, batteries and this new P7 light

Here's the page with the size and weight comparisons.
http://www.el34world.com/Misc/bike/BikesLights6.htm

For example here's a Lume Strada HID next to the new P7 design.