How hot is too hot

Wondering how best to check temperature of cobs? Mine are running on passive sinks. I have an infrared thermometer, but heard you can't get good readings from bare aluminum. How about anodized pin sinks... where do I check and what is the danger zone? Running citis if it matters. Just finished the build and would like to avoid melting cobs if I made any mistakes along the way.

I'm sure it's not recommended but I scan the COB from a couple inches away and grab the max reading. I installed a thermocouple on one luminaire and results were similar.

I want things below 85C; I'm running CBX3590's.

measure voltage at startup and steady state

less than 3% voltage drop = no sweat
3-5% voltage drop = check your datasheet or manufacturer-provided product simulator

in general if you can put your hand on the heatsink no worries

CobKits said:

measure voltage at startup and steady state

less than 3% voltage drop = no sweat
3-5% voltage drop = check your datasheet or manufacturer-provided product simulator

in general if you can put your hand on the heatsink no worries

Click to expand...

Awesome. I was worried about messing up thermal paste and having poor transfer. Ill checl em out tonight. Thanks!

upsetter. said:

I was worried about messing up thermal paste and having poor transfer.

Click to expand...

voltage drop will tell you a lot about that

stable, slowly dropping voltage as heatsink heats up = woohoo!
unstable voltage that drops quickly = ruhroh

althiugh i seem to get blow back for teaching people how to monitor actual temps instead of guessing.......

this is still damn good
http://www.mechatronix-asia.com/LED_heat_sink_calculation_simulation_thermal_design.html

A little context helps too...for instance, i personally would never run a 80C Tc but that is perfectly acceptable from the manufacturer.

imho 50c is too high as well, but more an enough compromise off the manufactures max Tc rating.....

I use the pain test. I grab the heatsink with my uncalloused fingers. If it hurts before a second has elapsed, things are too hot. (Of course, if the COB is incorrectly seated the heat will not be transferred to the sink.)

If you can make s'mores( thats melted chocolate y'all) under them...they are too damn hott

Abiqua said:

althiugh i seem to get blow back for teaching people how to monitor actual temps instead of guessing.......

this is still damn good
http://www.mechatronix-asia.com/LED_heat_sink_calculation_simulation_thermal_design.html

Click to expand...

I don't know what's wrong with this approach. No blow back here. I used similar information/calc from a different web page. Nothing wrong with doing the math up front. Take the guesswork out of it.

sanjuan said:

I use the pain test. I grab the heatsink with my uncalloused fingers. If it hurts before a second has elapsed, things are too hot. (Of course, if the COB is incorrectly seated the heat will not be transferred to the sink.)

Click to expand...

for efficiency snobs the "nuts test" is also pretty accurate

CobKits said:

voltage drop will tell you a lot about that

stable, slowly dropping voltage as heatsink heats up = woohoo!
unstable voltage that drops quickly = ruhroh

Click to expand...

woohoo!