Custom made SSC P7 LED Bike lights

[StevelKnievel]

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.
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

Jeez. These guys know what they're talking about EL and they really are trying to help you out! CPF has been far and away the most helpful technically oriented online community that I've experienced and some of the guys giving you advice have been playing the high power LED game for years.

Your test may not have been standardized but it does give a good bit of insight into the thermal performance of your design. If you really want to develop a reliable product it will need to perform under the worst possible conditions, not just the average bike ride.

Even if your test was in completely still air at 30C ambient... and there is perfect heat transfer from LED to the outside surface of the housing... and you had perfect heat transfer to your thermometer... and there was no additional heatsinking help from atmosphere or test surface... you're light is still running that P7 at the ragged edge of its thermal envelope. If any of the hypothetical assistance you mentioned was in effect during your test then your design is pushing the P7 OVER the edge.

FWIW, I think you are have designed a nice assembly for bike use. It's light weight and should give a nice beam. Given the questionable thermal performance though you might consider driving the P7 below 2.8A. At 2A you'd still get lots of light and because the thermal situation will be much improved you will have fewer heat related losses. Just a thought.

Good luck!

 

[EL34]

Just having a friendly go at the pocket saver/slide rule guys, sorry.

It's all done with tongue in cheek on my side here.
Don't take my comments too seriously, it's just an exchange of words and different people do it in different ways.

Seems like the post count is what triggers some of these guys into the what If design mode. They just assume they are dealing with someone who does not have a clue and procede to quote from their text books. This is just as offensive to me as my comments back to them may or may not be.

Not really a big deal, just silly internet stuff which seems to happen on large forums only where you get an overwhelming number of know it alls.
Does not seem to happen much on smaller forums.

If they are going to throw silly comments out, they can expect some back, that's only fair.
They are allowed to speak and then then it is my turn.
I do get a turn don't I ??? :huh:

Anywho, the guy at the top seemed like he wanted me to take the light apart and run it without the casing which is the actual heatsink and would not take no for an answer.
The guy that thought I would actually run a casing as thin as tin foil was another fun comment.

There were several more silly ones and there were some great comments also.

If you start at the top and put yourself in my shoes, you may or may not get the feeling that some of the replies were just spewed out by people who want to make sure everyone knows they have their slide rule out and it is working. :wave:

Basically, the light runs at 108 F (better throw that F in for the pocket saver guys)
+ or - any number of degrees depending many other factors.

It's not a big deal, just let it all slide and it will be over in a very short time.

 

[jbassett]

EL34, you are giving the impression you don't have the first clue and are coming across as a real idiot.

Why not just forget trying to be clever (and failing) and address the comments made; YOU DID ASK FOR THEM AFTER ALL

 

[cmacclel]

Seems like the post count is what triggers some of these guys into the what If design mode. They just assume they are dealing with someone who does not have a clue and procede to quote from their text books.

Actually the post count had nothing to do with it. It was when you stated that a thinner body dissipates heat better than a thicker one which is completely the opposite to what everyone else has found.

Anywho, the guy at the top seemed like he wanted me to take the light apart and run it without the casing which is the actual heatsink and would not take no for an answer.

I assume your referring to George who is the owner of TaskLed.com and who has been designing high quality boards for years. He also creates boards SPECIFICALLY designed for Bike Lights. He was trying to help you like the rest of us, but you made short work of that. This a the most friendly forum out there. Maybe you should try another forum if you cannot deal with people trying to help you.

Mac

 

[EL34]

No, it was an example of two extremes which is always a good way to demonstrate to people what happens at both ends of the spectrum.

Logic: If tin foil did not shed the heat, it would be hot to the touch after hetaing it up. Where did the heat go?
That does not mean that tin foil makes a great heat sink. I am suprised that another person thinks that is what I am saying and does not understand. Look it does not matter to me.

Somewhere between those two extremes of tin foil and a very thick piece of metal, there is a proper balance that will work.

Hey basset, I get the same impression of you and so that seems to work out just right. Love the picture, fits perfectly.

 

[jbassett]

OK, EL34. I quite happy to take comments; was anything I said wrong?

 

[jbassett]

I do take exception to someone making snide personal comments on my picture, at least I could be bothered to put one up

 

[EL34]

No, you said everything correctly.

 

[John_Galt]

Hey basset, I get the same impression of you and so that seems to work out just right. Love the picture, fits perfectly.

Please stop with the snide comments. If you don't want to, try joining the Flashlight Reviews forum. From what I've seen, they take great pride in their immature and idiotic (not to mention insulting) comments.

 

[znomit]

At a guess I'd say metal bit one is making loose contact with the outer housing, the small region that contacts metal bits one and two is a bottleneck for the heat flow.

I would think that after running for 10min, if you pulled it apart piece one would be too hot to hold.

There are some nice mag head thermal modeling posts here that show how the heat is greater with some mass removed.

 

[EL34]

Thanks guys, it has been entertaining but almost completely pointless and a useless drain of perfectly good energy.

Got lots of stuff going on today and so I'll be signing off.

Where is that Un-subscribe button

Oh yes, there it is - click :thumbsup:

 

[brad72]

Hi El34. The solid works thermal analysis is interesting. I'm very surprised that the fins caused the main housing to get hotter. I wish autodesk inventor had the thermal modelling feature.

I am just in the process of designing my first led lamp, so these will all be joys that I am yet to have. (i have made plenty of halogen lamps)

Just shows that unless we do thermal modelling on all the designs of light out in the market place, no one would really know if heat will kill the led and how many hours the led has to burn to produce a failure. The thermal modelling I think also needs to cater for the heat produced by the driver which is why I suppose the most efficient driver one can find would be the most suitable.

At the end of the day as you said this light is a protoptype. If thermal management or heatsinking is not correct then you will soon know.

Good luck with the project

 

[sampie]

George, if you looked at the links I provided it

Where can one get the mount for the lume

 

[chris_m]

No, it was an example of two extremes which is always a good way to demonstrate to people what happens at both ends of the spectrum.

Actually it was a totally pointless example. Whether you get burnt by the slab of metal but not by the tin foil has absolutely nothing to do with the thickness of metal which is useful for the outer body of a lamp.

Looks nice - shame you don't seem to either have a clue about thermal paths, or to be prepared to listen to those who do. Therefore it's likely to be performing nowhere near as well as it could if it was properly designed. Personally I'm always dubious about "snug fit" for a thermal path - you really need either a compression fit or a glued fit.

 

[Frobe22]

LED temperature is the ONLY important temperature here, case temperature only matters if you are going to get burned by touching it. As a bike light the case will be cooled when it's moving. A small thermostat in series with one of the 1400mA drivers would be a simple way to protect it from damage if used with no airflow.

In order to measure the LED temperature you may have to drill a hole in the body for the temperature probe. A small termocouple probe will require a 1-2 mm hole or notch. You can either drill it directly in the side of the case, or if no glass fitted yet you can make a notch in the reflector for the sensor to reach the LED body. (If you have large enough holes between the driver and the LED you may even be able to sneak the sensor in together with the wires). The sensor should touch the base of the LED body or the star it is mounted on. A small dab of thermal paste will ensure the sensor thermal contact.

I miss Newbie. He even drilled a small hole in a 5mm LED to measure the chip temperature. https://www.candlepowerforums.com/threads/128801
Lambda proved on his homepage that Luxeon LED with barely enough cooling even dropped in brightness when increasing driving current. Cooling first, current later if you want bright high power LED light.

 

[jbassett]

In order to measure the LED temperature you may have to drill a hole in the body for the temperature probe.

Argh, the horror of it. I wouldn't entertain drilling a hole in a nicely machined case to measure the LED temperature as it's not actually required if you plan ahead. It usually possible to feed the probe through the mounting hole for the switch or bracket.

 

[NoShoulder]

If you've ever read ANY of my posts you'll know that I make no claim to engineering knowledge. However, I do know a lot about communication and EL 34, I can't tell if you're kidding or really not understanding the basic question being asked.

While I'm sure everyone is interested and impressed by your experience and opinion (no sarcasm - I'm serious) we ALSO would like actual measurement that only you can provide, that is, temperature of the LED in use in the light.

If that is not clear, please explain and I or someone will clarify.

NoShoulder