Great new thermal switches

100 degrees is pretty low......... too bad they don't make those that small adjustable.... I've seen those used as rollout switches on newer furnaces, but they have a manual reset button when used for that application.
 
Well, the data sheet for the XR-E shows max led junction temp at 150C. How hot your heatsink should be depends on the quality of the mechanical connection between it and your emitter, the design and efficiency of your sink, how well it is connected to the light body.... In otherwords, there sure is a lot of factors to consider.

Theoriticaly, an absolute perfect sink would be the same temp as your led. I doubt if that is ever the case, but going out on a limb, a total guess I would think at least 100C.

Rather than a thermal switch per say, if I was doing this I would look into the use of a PTC thermistor. Depending on how much room you have in your light, perhaps one with an external probe that you could butt up against the emitter.

How to wire it in would depend on the driver, if using a CC driver with a sense resistor than you could parallel it to that sense resistor. Or, worst case it could be between the led and driver, it would still limit current to the emittor by increasing resistance, but it wouldn't do anything to protect the driver.

Others here have show that the use of NTC's for soft start in incan lights has the drawback of varying greatly according to ambient temps. I don't think that would be a problem with the use of PTC's though, as ambient temps that effect the PTC would correlate directly to led temps as well?
 
If I'm reading the specs right those temps are in C. So they should be perfect for this application. I'm not quite certain, but I think the proper range for safty is around 60-70C for the heatsink. I can run my M60 for around 10 minutes in a surefire g2 later to see how hot it gets, that's a good safe upper range for longetivity IMO.
 
Well i think im going to go for an 80 degree one and silver epoxy it onto the back of my sink.

The sink is a 10mm thick, 63mm across disc of solid aluminium that is press fitted into a 3mm wall thickness tube of 9 inches long and surround by the whole British ocean.....

Im just hoping this is going to be enough for 7 XRE at 900ma. but one of these switches would be a great saver if it turns out it's not. To much time and money has gone in now to just left it burn up....
 
100 degrees is pretty low......... too bad they don't make those that small adjustable.... I've seen those used as rollout switches on newer furnaces, but they have a manual reset button when used for that application.


It's "C" not "F"

Mac
 
Americans!! LoL...

Just to let you know, i just checked and the they are definitely in degrees centigrade.

So what do you think 80 or risk it and go with 100.

After all i don't want this thing going **** up when im down at 40 meters in a dark wreck...
 
Well it actually seems as though you have a heck of a heatsink already, and being surrounded by water to boot... but, better safe than sorry, huh? Are you using bare emitters or stars? Is the output adjustable?

I am thinking you are going to need to do some testing here. Do you have a meter with a thermocouple? Perhaps start with the 80C switch, tie it into the heatsink back and insert your thermocouple as close as possible to at least one emitter. I am just going to guess pick one at the center, seems those would be most likely to have issues. I would think do this on dry land just for test purposes.

There is such a thing as thermal modeling software but it is SO far over my head, I don't begin to know how it might help you.
 
Yeah thought about trying to test it. But the heatsink is a one time only compression fit, perfect contact with the tube, metal to metal no glue or other gunk required...

As for the spec it's shaping up like this so far:

6 Cell 5000mah Lithium Polymer battery pack max voltage 25.38 thertical.
Maxpro voltage monitor / fuel gauge
Blue Shark set on max output
63 by 10mm heat sink in 3mm thick tube
7 Cree XRE q5 on 14mm round PCB's ~( 25.2fv in worst case)
1 Ultra fast silicone diode ~(To raise fv by around 1v).
80 deg Thermal saftey switch
7 13 deg Carclo Optics
Silicone covered silver wire throuout
Customed designed housing tested to 150M depth.....
2 Mode On / Off.....hopefully not permanently!!

Now if it wasnt for chrismas and everywhere closing i might of had a chance of getting it finished, but my Led's arrived from DX last friday and one had lens missing, so there sending me 2 replacments! Got to love that for service.

Would a PTC increase in resistance at it's specified trigger voltage which would then trick the driver circuit into pushing more and more volts through to try and maintain the rapidly decreasing amps to a point where ethier the PTC gets to hot (adding to the problem) and fails or the driver board blows up beacuse it max;s out and then also over heats?
 
Last edited:
I'm seeing max 50C on my surefire C2 body with malkoff M60 in 85f/29c ambient temps in still air. So I think 70 would be about right. 80 should be okay but any higher than that and I think you should just design a better heatsink :) Actually get an IR thermometer and see what real world temps you're getting, then get a proper thermocouple to match.
 
Wow, we are dealing with a lot more power here than I originally realized. I imagine there would be a ptc availible in that range, but I was thinking of a single cree light when I suggested it.

With a cheaper build and a driver that wasn't adjustable like the shark, the PTC between the led and the driver, the ptc would steal some of the power away from the led and the driver should keep steady output, still putting out say ~1amp but a portion of that would be re-directed into the thermistor, dimming the light (depending on the amount of resistance chosen).

With the shark it is even more ideal because it would act as a variable sense resistor to actually control output from the driver. Since the sense resistor is only seeing a small portion of the output from the driver, it would ideal.

I don't know it for a fact, but I am guessing that is how the magled module works, or probably most any drop in that "throttles" back the output when hot.

Here is one article that discusses the use of them for led drivers:
http://www2.electronicproducts.com/...nt_control-article-farrepcs-jun2007-html.aspx

So, in theory it would work very well. But, in practice, it may be a bit harder. Just finding the right value and rating for one thing, and proper placement of the PTC in your design. Obviously the design is dependant on accurate temp measurements, and I would think as close to the leds as possible.

Now, this isn't to say your thermal switch wouldn't work just as well, there is more than one way to skin a cat after all, and this is your design. Just attempting to offer you some options.

One other slight weakness I see in your design is the MCPCB. It is just one more thermal junction to deal with. A bare emitter with the corners trimmed off should give you a better mechanical connection.

But, I have to say, you are on your way to one heck of a dive light! I can't wait to see it :twothumbs
 
I was just thinking what would be really neat would be an onboard display. Push a button and get different read outs. Batt voltage, temp reading, led voltage, and current reading! A bit much I know, but just fantasizing.
 
What a cool idea, and i know of a way it would be possible!! http://www.swanghobby.com/product.php?productid=128&cat=15&page=1

But that's maybe for a future development.

I have end up going for an 80 deg. Switch from rapid electronics, no idea when it will get here though....Also waiting for a replacement LED from DX.

Ideally i would like 6 digress optics but couldn't find any at a cheap price and 19mm or less in diameter.

I will keep you updated as to how it goes.

Thanks so far for your ideas and help.
 
80* for a sealed unit, 60* for an in-air unit sounds about good, anything is better than nothing, at least with these there is less resistance then from a curcuit breaker (fuse) thing, because with those the circuit breaker itself heats up to get the bi-metal to bend, so there is much more resistance to CAUSE the heating , so it breaks on a certian amperage.

these are speced at <30 miliOhms, still adding in resistance that would need to be accounted for, but much less than most of the curcuit breakers.

but beware of Riveted spade connectors, they can become loose over time, if possible solder the rivet to the spade, without destroying it, for longevity.

myself i might even go Lower (they dont cost that much) then force an overheat (on purpose) and see if it clicks off for you, then up the temp just a bit from the one you KNOW worked in practice.
 
Gav, I reckon your heatsinking will be fine. I've had no problems with my dive light which is also 7 x XR-E at 1A. My sink is 20mm thick ally, 50.8mm diameter, held in place with cheap thermal glue from DX. The housing ID is 52mm. It's done 2 dives now without issue. My buddy reckons it's too bright !
:shrug:

Someone on here has used a 10mm thick ally sink for a dive light which uses 4 x Osram Ostar at 700ma. That would produce way more heat than yours. The gap between his sink and the housing is around 1mm and he uses thermal tape for heat transfer.

http://www.tauchfunzel.de/
 
Last edited:
Hey steve Worcester isnt all that far away from me in Gloucestershire?

I already have a mate "Tony" from your way who i dive with about 4 times a year.

Do you fancy meeting up for a dive sometime next year? We could take a look at each others diving torches.

PM me.

Cheers.
 
Top