Performance Viewpoint/Marwi halogen-to-LED conversion (UPDATED)

[cityevader]

Re: Viewpoint LED conversion

Oh yeah, I checked some temperatures on low and high once temps leveled off after about 20 minutes on each level, with zero airflow and ambient temp of about 56°.

Anyody know what temperature (F, I've no idea C) to not go over...not terminal/death temps, but running temp for long term lifespan prolonging?

 

[BrianMc]

Re: Viewpoint LED conversion

The max junction T for the SSC P7 is 125 C or 257 F.

Four XP-G's are 13.2 compared to your 10.4 watt and this guy has used a hobby motor heatsink to help out. If the fins were intergral with the body they would have excellent heat path. Hopefully, it clamps on tight and has enough metal-metal contact to help. The ano won't help as I understand it. You could try one in the hobby store with your light and thermometer to see, if you liked the idea. Then buy one for each light if it worked well enough that it was worth it.

Please tell us if you try one out with the thermometer as it is an interesting solution for these lights.

 

[cityevader]

Well, my temp (with zero cooling airflow) maxed out lower than that, (although emitter temp is certainly higher than housing temp) but maybe on a warm night ride in summer it may be a problem.

One thing about copper is it conducts heat super fast. When I was grinding down the cup at the very edge, it seemed like the moment the grinder hit it the entire thing was too hot to hold. There will of course be loss where temp flow is restricted from cup to housing.

But to be honest, there is so little difference is apparent light output between 1.4 and 2.8 A that I will be running it on low probably the entire time. Funny how things are perceived. Kinda like sound, where twice as loud (db) is actually barely perceptible. Therefore further cooling requirements aren't likely necessary?

 

[pe2er]

Your light turned out very nice I Like the floody P7 light - I use my magicshine P7 for commuting on High constantly (must be my age, that I need that much light )

Would not worry about the housing temp too much (168°F = 75°C). A housing heating up fast means that the transfer from LED to housing is good, so that is a good sign. In my experience the lights heat up far less when some air is flowing over them, compared to 'on the bench'. I suggest you take some temp readings after a half hour ride and see for yourself.

PS: The 140°C limit for the P7 is the temperature of the die itself. This temperature is higher than that of the housing. Energy (heat) flows from the emitter to the surrounding air. Each material or junction between materials has its own thermal resistance, resulting in a temperature drop over this resistance. Just like an electrical current causes a voltage drop over an electric resistance.

So the maximum allowable temperature of the housing is less than that of the die (140°C), but depends on the amount of thermal resistance between the die and the outer surface of the housing and the thermal resistance between the housing and the surrounding air.
The better the heat conductive path LED to housing, the higher the temperature of the housing will be.
The better the heat conductive path housing to air, the lower the temperature of the housing will be.

Thermal resistance for the P7 is 3°C/W. About 20% of that is emitted as light. So at 9 Watt drive, the die will be 80% x 9W x 3°C/W = 21.6°C higher than the heatsink or PCB that the LED is mounted to. So absolute maximum temperature of the heatsink directly under the LED is 140°C - 21.6°C = 118.4°C. Of course, keeping the LED at lower temperatures increases LED life and efficiency.

For more information: See the Thermal management Guide and P7 datasheet on this page: http://www.seoulsemicon.com/en/product/prd/zpowerLEDp7.asp and I am sure that there are more than one thread in this forum on this subject. The picture above was taken from that guide.

To further complicate things for you: The constant current driver also has a finite electrical efficiency, and also creates heat that increases the housing temperature.

 

[BrianMc]

The thermal process is dynamic. As pe2er points out, the speed of warming of the body of the light is a better indicator of an effective heat path from the junction to the air, than is the final temperature. The heat storage capacity of the light, its surface area and the ambient temp make comparisons esssentially meaningless between lights, but for the same light design at close to the same ambient temps, you can judge whether things are better or worse.

It is like moving water in a siphon. You can siphon from a higher potential energy raised barrel of water to a lower tub. You can fill that tub (same temp) with a smaller or a larger hose, but the larger one will do it faster. Any crimps in the hose (resistance at interfaces) slow the flow additively. Shorter hoses are faster (less resistance from the walls of the hose). Small lights have short distances so faster heat transfer to the body but less area to transfer from the body to the air (smaller tub, so they fill more for the same flow), so they often are finned to add area.

With air being stripped away so that cooler air is always present, the thermal potential to transfer heat is higher and speeds the process to a new equilibrium with a lower junction temperature. It also removes the insulating aspect of a boundary air layer. So air flow over the lamp has a big effect on the temp at the LED junction. So you can have a cooler light because the heat isn't getting to the body (essentially insulated by resistances along the way which makes for a hot junction temperature) or because it is moving heat very well and it is getting dumped into the air efficiently (a low junction temperature).

Copper is a very effective spreader of heat so the primary resistances beyond those in the star itself (and out of your control), will be from the star to the copper and from the copper to the aluminum body. Having polished the inside of the newer housing for better heat sink contact and heat transfer, I can tell you they have high and low spots. I see no reason why the older style would be any different in that regard. The low spots fill with thermal grease which is way better than air, but worse than metal to metal contacts, and the high spots will make metal-metal contact. Anodized aluminum is supposedly a little less efficient an interface to metal than polished aluminum and closer tolerances mean thinner thermal compound layers and more heat transfer. Like lapping the star and pipe cap, this is likely a small 5-15% range improvement, and very time consuming.

So if your light got hot fast, (and this isn't very hot), then considering air flow and lower power when in use, yeah you're good to go.

 

[cityevader]

Well, I finally got around to going on a ride with these last week, and of course, I didn't have my thermometer since it's back in my toolbox at work.

Now I wasn't counting during the static tests, but I'd guess that on low I could hold the lamp for 8 to 10 seconds before letting go in discomfort. On high it was more light 2 seconds before feeling too hot to hold.

The night ride was at the same ambient temp as the static tests right at 51°F , and I was riding fairly slowly at around 10mph with the nervous wife. I had mine on high for about 45 minutes, and kept feeling the lamps....I couldn't feel them.

My fingers very fairly cold, certainly below typical body temp of 98°, and the lamps felt absolutely neutral, neither warm nor cold, so I'd estimate the lamps with a mild breeze at only about 80°?

Not too shabby!

Now if I can only find a spot reflector to squeeze in!