Measuring Temps

Would anyone who takes temp readings care to detail the process?

I'm wanting to start taking measurements but not sure how to go about it. I have a temp probe but can't see tightening the screws down with the wire between the COB and heat sink. Is the measurement taken off the side of the contact?

Does it look like this?

Bumping this thread hoping someone with experience will see it.

It's something I'm really interested in so wouldn't mind spending some money on proper equipment.

I don't have the cash for PAR measuring equipment, but being able to measure temps at Tj would provide a pretty accurate estimate of output, and would show variances between various cooling methods.

Kinda busy right now but I can say I like the cheap handheld dedicated heat meter thing I got off of Amazon.

It has two slots for inserting the included thermocouples. I prefer reading it off that to a heat gun. Normally just kapton tape it to the test spot on the cob or nearby on the heatsink. Think it's more accurate that way..

Doh, the test spot is new to me, thanks. I've been over datasheets dozens of times. I guess my eyes ignore what they're not looking for.

As far as getting a probe between the emitter and heat sink, from electronics-cooling.com:

"All three types (J, K, and T) are available as insulated duplexed pairs from0.001-inch diameter on up. For accuracy, and minimum system disturbance, the smaller the wire the better, but wire smaller than 0.003-inch diameter is very fragile."

So perhaps it's possible to get a wire between the emitter and heat sink without harming the emitter. It might be worthwhile to compare such a reading to the test spot to see what kind of variance pops up.

From my little time testing, the test spot on the cob and the heatsink in the immediate area, whether right next to the cob on the face, or right behind the cob read the same. The heat pretty much will equalize.

If there is a bad interface between cob and heatsink then I could see a difference happening. Definitely smart to check that in particular.

Putting a wire under the cob sounds like a bad idea I think. Shouldn't be necessary and would probably only serve to tell you what the temp would be with a wire under the cob..

Get a thermometer, and monitor the readings. Sometimes you can find a digital unit that measures temps, and humidity.....

For about 3 bucks

Here's the majic spot (on a cxa?) :

I just used a cheap handheld meter.. but set it in place/ aimed on the spot (before turning the lights back on each time).
Might not be too accurate temp wize but shows up accurate variation/relative temps and doesn't seem to need to be precise ie. somewhere around the spot for the same temps (as Pos mentions).

Here is Bridgelux's demonstration on pages 26,27 and 28....
http://www.bridgelux.com/wp-content/uploads/2013/10/AN30-Thermal-Management-of-Vero-LED-Modules.pdf

Also:
http://www.omega.com/prodinfo/thermocouples.html

and these:
www.electronics-cooling.com
www.coolingzone.com
www.thermalnews.com/
www.qats.com/Qpedia-Thermal-eMagazine/Current-Issue/18.aspx
www.engineeringtoolbox.com
hyperphysics.phy-astr.gsu.edu/hbase/thermo/heatra.html www.electronics-cooling.com/1997/01/notes-on-using-thermocouples/


Couple of things...
1)
The closest thermocouple I could find @ omega.com was the
40 awg...$89 for the 5 pack...I am 99% sure they would work in the place of the 36-36 that is recommended as per Bridgelux specs above ^ ...[Cree may be different, but a datasheet hunt is on....]
recommended.......
5SC-TT-(*)-36-(**)360.13 (0.005)PFA


[available]
5SC-TT-(*)-40-(**)400.076 (0.003)PFA

http://www.omega.com/pptst/5LSC_5SRTC.html


Also from the Bridgelux Thermal pdf......

Use a 36 gauge or higher (smaller wire diameter) thermocouple for measuring Tc. Thermocouples of different size and type can be found at www.omega.com. A 36 gauge K type thermocouple with connector is PN: 5SRTC-TT-K-36-36.

Click to expand...

There should be no air gaps between the thermocouple tip and the surface of the Vero LED Modules. A visual inspection using a microscope or stereo scope is beneficial when attaching the thermocouple to inspect the bead contact with the surface

Click to expand...

Apply enough adhesive around the thermocouple bead to secure the thermocouple in place and allow the adhesive to cure. The adhesive should not be excessive such that it spreads to the top surface of the plastic housing or onto the LES area.
o Place a small strip of Kapton tape over the thermocouple junction to insulate the thermocouple from the surrounding air temperature and tape the wire down to the top side of the housing.

Click to expand...

#2....

Thanks to @salmonetin

He sent me the info about this Luxeon 7 diode star base called SinkPad-II
http://www.luxeonstar.com/any-7-rebel-leds-mounted-on-a-7-led-round-sinkpad

and has this NTC Thermistor / Surface Mount Chip buried into its "True T-case"
http://www.luxeonstar.com/assets/downloads/nths.pdf

^^definitely worth a look for some reverse engineering...

Tazbud said:

Here's the majic spot (on a cxa?) : View attachment 3374696

I just used a cheap handheld meter.. but set it in place/ aimed on the spot (before turning the lights back on each time).
Might not be too accurate temp wize but shows up accurate variation/relative temps and doesn't seem to need to be precise ie. somewhere around the spot for the same temps (as Pos mentions).

Click to expand...

This paper agrees with you.
www.lrc.rpi.edu/programs/solidstate/pdf/Perera-SPIE-2013a.pdf

Trove of info, thanks. Especially interesting in the pdf the info on variation between test spot and actual Tj. Those sink pads look nice! Seem expensive by the watt but having a few of those with 550nm mixed in with the COBs would be great.

Abiqua said:

Here is Bridgelux's demonstration on pages 26,27 and 28....
http://www.bridgelux.com/wp-content/uploads/2013/10/AN30-Thermal-Management-of-Vero-LED-Modules.pdf

Also:
http://www.omega.com/prodinfo/thermocouples.html

and these:
www.electronics-cooling.com
www.coolingzone.com
www.thermalnews.com/
www.qats.com/Qpedia-Thermal-eMagazine/Current-Issue/18.aspx
www.engineeringtoolbox.com
hyperphysics.phy-astr.gsu.edu/hbase/thermo/heatra.html www.electronics-cooling.com/1997/01/notes-on-using-thermocouples/

Couple of things...
1)
The closest thermocouple I could find @ omega.com was the
40 awg...$89 for the 5 pack...I am 99% sure they would work in the place of the 36-36 that is recommended as per Bridgelux specs above ^ ...[Cree may be different, but a datasheet hunt is on....]
recommended.......
5SC-TT-(*)-36-(**)360.13 (0.005)PFA


[available]
5SC-TT-(*)-40-(**)400.076 (0.003)PFA

http://www.omega.com/pptst/5LSC_5SRTC.html


Also from the Bridgelux Thermal pdf......







#2....

Thanks to @salmonetin

He sent me the info about this Luxeon 7 diode star base called SinkPad-II
http://www.luxeonstar.com/any-7-rebel-leds-mounted-on-a-7-led-round-sinkpad

and has this NTC Thermistor / Surface Mount Chip buried into its "True T-case"
http://www.luxeonstar.com/assets/downloads/nths.pdf

^^definitely worth a look for some reverse engineering...

Click to expand...

heckler73 said:

This paper agrees with you.
www.lrc.rpi.edu/programs/solidstate/pdf/Perera-SPIE-2013a.pdf

Click to expand...

Good'O. I didn't consider light source radiation or those needing to apply the accuracy of a shielded thermocouple but

Thinking out loud....I don't see why you couldn't drill a hole in a heatsink and place the thermocouple thru a pinhole in the heatsink and flush it with the mounting surface of the cob.....and have it directly read the the "true" Tc....

I did brush over the Tc - True Case temp conversion....so I really need to go back and calculate all the way up to 80C to get an idea of the range, but I don't see why a thermocouple couldn't read up to 250c or so...maybe Im wrong though...

Abiqua said:

Thinking out loud....I don't see why you couldn't drill a hole in a heatsink and place the thermocouple thru a pinhole in the heatsink and flush it with the mounting surface of the cob.....and have it directly read the the "true" Tc....

I did brush over the Tc - True Case temp conversion....so I really need to go back and calculate all the way up to 80C to get an idea of the range, but I don't see why a thermocouple couldn't read up to 250c or so...maybe Im wrong though...

Click to expand...

Actually, that would probably work, but I think it is unnecessary in this instance. I've used Type T thermocouples in the lab for measuring temps from liquid Nitrogen (77K) to steam (~400K). Thermocouples should have no problem so long as the temps are below their melting or oxidizing points.

http://www.asrichards.com/products/temperature/thermocouple_types.html

EEVBlog made a good vid on this subject (although he doesn't go into the physics). He even discusses the Type T TC.


NIST data tables for designing your own TCs:
http://srdata.nist.gov/its90/main/