Custom made SSC P7 LED Bike lights

[EL34]

Is that the led current or the battery current?

Same thing, the 7135 current regulator boards are limiting the battery current and therefore the LED draw current to 2.8 amps.

When the battery packs are fully charged, test revealed that the LED would draw 3.2 amps with no current regulation, which is a good way to fry the LED.

The LED current is rock steady at 2.8 amps with the boards

 

[Tidra]

" The LED current is rock steady at 2.8 amps with the boards "

for how long, because if I know this driver then with the input battery voltage drop, the output current will drop too,...

 

[EL34]

Hi Tidra,

As soon as the battery voltage drops to a certain point, the current will of course go down.

Problem is, I am using 7500 +++ mah battery packs, and that can take quite a while and so it would be hard to measure.

In actual real world off road riding test, I have clocked almost 3 hours riding time with the P7's and a 10,000 mah battery pack with not a lot of noticable difference in light levels at the end of the ride.

Testing current level drops with large capacity battery packs would be more of a question of how many batteries in parallel, how many mah the total battery pack is and what is the actual capacity of a battery pack.

I have a couple really nice digital display chargers that can measure actual battery pack capacity. The nice thing about building your own battery packs is you can add as many 18650 Li-On's as you need for a your particular ride lengths.

 

[jbassett]

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.

I can't believe you really think it is unimportant to measure the temperature. I converted a cycle light to dual Cree XR-E R2 LEDs and the outer casing didn't get very warm. However, I measured the temperature of the aluminium LED mounting plate and it hit 110°C and still rising when I turned it off. After making and anodising a custom heatsink the temperature peaked at ~65°C, in static conditions at 20°C air temperature, which I can accept. If I hadn't measured it the LEDs would not have lasted too long.

 

[EL34]

I know what the temperature of my outer housing is.

It is just fine.

 

[jbassett]

Yet again, you completely miss the point. The temperature of the outer housing makes no difference at all. The only important temperature is that of of the LED. If you want to have your $20 LEDs fail every few hours go ahead; most people can't afford running costs that high. The easy way to ensure a good LED life is to check it is running at an acceptable temperature. The operating temperature range is -40°C to 85°C according to the SSC datasheet, and that's the LED temperature not the outside air temperature!

 

[EL34]

I don't agree and this is why.

My lights are designed to run at a certain temp using the complete housing/heat sinks.

Who in their right mind runs a P7 without heat sinks?

Run your car without any coolant in the radiator and report back to me what the engine block temp is.

Run a Dual core Intel processor without a heat sink/fan while playing a high end graphics game and report back to me what the core temp is.

My lights run perfectly fine the way they are designed to run, with the full metal jackets intact.
Stop asking me to run just a P7 without proper cooling.
Run one yourself and then report back to me, how's that.

Please get someone else to play that game and leave me out of it.
I am under no obligation to run silly meaningless test just for your amusement, I don't work for you or anyone else.
Thanks, I appreciate it.

 

[jbassett]

Run a Dual core Intel processor without a heat sink/fan while playing a high end graphics game and report back to me what the core temp is.
<snip>
Stop asking me to run just a P7 without proper cooling.

Run any CPU with an inadequate heatsink any it will overheat even though there is a heatsink fitted. Just having a heatsink doesn't mean it will stay cool.
NO ONE has asked you to run a P7 without proper cooling! We have been asking you measure the P7 temperature whilst fitted to your heatsink and case. The only other way to determine the effective temperature coefficient of your case and then calculate the running temperature, It sounds like you've not done this.

 

[greenLED]

My lights run perfectly fine the way they are designed to run, with the full metal jackets intact.
Stop asking me to run just a P7 without proper cooling.
Run one yourself and then report back to me, how's that.

Please get someone else to play that game and leave me out of it.
I am under no obligation to run silly meaningless test just for your amusement, I don't work for you or anyone else.
Thanks, I appreciate it.

EL, if I may, we're not exactly new to building custom lights in this forum. There are *years* of cumulative experience in building lights in this forum.

That experience has shown that if you don't dissipate heat from LEDs, they will suffer from thermal degradation. All we're trying to point out is that, as is, your setup:
A) does not dissipate heat as well as it could if you increased the outer surface area, and
B) your emitter will degrade faster than it would if cooling were to increase from your current design.

You'll find some of the most dedicated (read: obsessive with detail) people on this forum. They've done their homework, they've (we've) burned lots of emitters and circuits doing testing, they "may" know a thing or two about building lights. Nobody on CPF falls for the "trust me, I know what I'm doing" mantra. It's not personal, it's simply how we've advanced the knowledge around here.

The way I'm reading this thread, we're not doubting your experience or claims. We're helping you make a better product, one that will last longer and keep people happy. As with a lot of other projects showcased on the forum, it will get dissected, criticized, rebuilt, etc.

The easiest way for you to get people to back down and leave you alone will be to compare the temp between a finned and an unfinned setup. Simple test. You don't have to work for me or them to try this and see for yourself. You don't even have to report here. Just try it, for crap and giggles, if anything.

Worse thing that could happen is that we all learn from it (either way).

Good luck with your project.

 

[jbassett]

I agree with greenLED that comparing a finned and non-finned case would be good. However, I think it still important to measure the temperature as close to the LED as possible.
When you measure the temperature of heatsinks, they are much hotter at the source end than the free end when running at higher temperatures. This is due to 2 reasons.
1. All materials have a thermal resistance which impedes the heat transfer. In a built-up structure each joint between pieces of material adds an additional resistance. Where parts only touch with a limited surface area this resistance can be signficant.
2. The heat getting to the open end is less than the source end as the rest of the heatsink has already dropped the temperature.

 

[cmacclel]

Very simple way to tell if your heatsinking is working.

Hot body = Good Thermal Path

Cool Body = Bad Thermal Path

Before even thinking about offering a run of these I would be sure you have covered all bases before you waste you money.

If I was making / designing a bike light one of the requirements would be a smart driver board with temp monitoring / throttling.


Mac

 

[EL34]

Mr Greenled, thank you, finally someone that seems to have a bit of logic and friendliness.

Here's an older image of one that I popped off the lathe. Is not really very polished looking yet.

This is the test I ran a couple weeks ago with the latest shell design with fins shown above:

A digital thermometer sensor taped down to the shell of the light right behind the fins. The fins are right over the top of the P7
.
The light was run on 3x18650 batteries in parallel at 3.6 volts at 2.8amps.

The light rose to 109 degrees in about 2 minutes and stayed there for about 40 minutes until the battery current started to drop a bit.
Then the temp gradually went down slowly until I stopped the test after an hour and a half or so.

I have several of these lights that have been running in the field without incident so far.

Sorry about the rant, but I really don't understand why some people think I am obligated to run test for them personally.

I am not selling this light to anyone here, so why would anyone care about my personal lights.

Here's the simple part:
If there is some sort of issue, I will change the design.
If not, I won't

That's pretty simple to understand if you ask me.

 

[jbassett]

The light rose to 109 degrees in about 2 minutes and stayed there for about 40 minutes until the battery current started to drop a bit.

Was that 109°C or 109°F?

 

[EL34]

Yes, it was 109 celsius of course.
I was running a P7 at 228 F for 1.5 hours. :whistle:

Or maybe it was Kelvin, Rankine, Delisle, Newton, Reaumur or Romer.
I forgot which thermometer I was using at the time.

Could have been the one my wife brought home from the hospital that was used where the sun don't shine. :party:

Someone give me a link to a site that describes in very prcise detail what the standard test setup is so that everyone is comparing apples to apples.
I am also gonna need a page of detailed results from all of the flashlight manufactuers and users here that have performed the exact test setup.

Then I can compare my results with a scientific standard set of known results.

 

[jbassett]

There is no need to be so maladroit, Fahrenheit and Celsius are used equally around the world.

I'm sure you're aware that the maximum LED junction temperature for LEDs is around 140°C and that is where the SSC datasheet stops; anything higher and breakdown is pretty quick. The thermal resistance of the LED is given as 3°C/W so, nominally, the LED junction will run about 30°C hotter than the solder connections and backplate at 2.8A. From your figures even a 100% efficient case design means the LED junction is at maximum temperature.

A point to note is that at 140°C junction temperature the light output has fallen to 75% of nominal. At this temperature the nominal 700 lumens output is down to only 525 lumens. That is a loss of nearly 200 lumens which can be drastically reduced by keeping the LED cooler.

 

[EL34]

You forgot to ask me where the test occured?
Maybe I cheated and did this in my freezer?
If not in a freezer, what was the temp of the room? C or F?
Was there a ceiling fan on?
If so, what speed was it on and how far away from the light was the fan.
Do I have 9 foot ceilings, 8 foot ceilings, a vaulted ceiling or maybe even cathedral. :eek:
What kind of tape did I use to tape the sensor to the body.
Did I walk by the light several times and create an artificial breeze which would throw off the average temperature results.
What kind, if any thermal paste did I use.
Was the digital thermometer of proper quality or was it from Wal-Mart.
Was the light sitting on any sort of surface that could act as a heat sink?

Hmm, so many possiblilties and possible ways to alter the test results.
I'll have to call this test void and null because the test setup was not standardized in any way meaningful and therefore, out it goes.

I'll set up another proper test when I receive those documents I requested.
I promise to abide by the standard test procedures and present them before the board in good faith.
lovecpf

 

[jbassett]

Really, are you always this facetious?

Whilst air temperature, atmospheric pressure and humidity do affect the result it is only by a relatively small amount.

The biggest influence would be air speed. However, most test in static air as that is worst case and a light may have to operate for prolonged periods with no air flow (yes, even a cycle lamp).

I tend to test at room temperature (~20°C) as that convenient and is higher than normally experienced at night.

I'm sure I could find you an International Specification if you really want one.

 

[EL34]

No that's ok thanks.
I only consider real world test results as valid.

I found out a long time ago that there is a big difference between someone calculating a (what if) result and an actual real world result.

One of them is what is actually really happening and one of them is a guy with a calculator in his hand.

Really appreciate it though. :poke:

 

[John_Galt]

You forgot to ask me where the test occured?
Maybe I cheated and did this in my freezer?
If not in a freezer, what was the temp of the room? C or F?
Was there a ceiling fan on?
If so, what speed was it on and how far away from the light was the fan.
Do I have 9 foot ceilings, 8 foot ceilings, a vaulted ceiling or maybe even cathedral. :eek:
What kind of tape did I use to tape the sensor to the body.
Did I walk by the light several times and create an artificial breeze which would throw off the average temperature results.
What kind, if any thermal paste did I use.
Was the digital thermometer of proper quality or was it from Wal-Mart.
Was the light sitting on any sort of surface that could act as a heat sink?

Hmm, so many possiblilties and possible ways to alter the test results.
I'll have to call this test void and null because the test setup was not standardized in any way meaningful and therefore, out it goes.

I'll set up another proper test when I receive those documents I requested.
I promise to abide by the standard test procedures and present them before the board in good faith.
lovecpf

There's absolutely no need to be so insulting. The people here answering your thread were just trying to clarify things for themselves, and help save you money. I first thought your thread was interesting, until you started becoming so discourteous, and omnipotent.

 

[cmacclel]

Not cool when people are trying to help you then you go insulting them :(

I wish you luck.


Mac