active cooling vs passive and Tj

pirg420

pirg420

Well-Known Member
So with the recent chatter about Tj affecting efficiency it would seem that an active cooled setup would be much better at keeping the Tj lower and thus allowing the chip to run more efficiently, and create more par etc
On my 1100 watt Bcblondes fixture the fans pull 16 watts at 110 volt. We are comfortable 'wasting' this power if our chips are running cooler and gaining efficiency etc.

Ive a few questions. How are you measuring Tj? When I use an laser thermometer on my heatsinks I see about 28 degrees C. I am unable to measure the temperature behind the chip on the heatsink surface(obviously), which is what i think Tj refers to, but I might be wrong. Is the 28C I'm measuring my actual Tj?

Ive seen estimates from Greengenes putting Tj at about 40C with active cooling and supra has mentioned about 50C with active cooling obviously this could fluctuate depending on any variables but they seem like decent estimates.

Would love to hear everyones thoughts on this.

thanks
 
CannaBruh

CannaBruh

Well-Known Member
Tj would have to account for each diode in the package. Cree makes note of this in the data sheet and suggests using forward current IsubF & case temperature TsubC

Determine maximum TsubC temperature before efficiencies dive, and then couple a sensor to that pad, monitor directly and see how different cooling approaches influence these temps.
 
pirg420

pirg420

Well-Known Member
Rahz said:
You can get a reasonably accurate reading using a thermocouple attached to the test point on a cob. It's close to the LES but hidden from view on your lamp by the reflectors.
Click to expand...
Cool thanks for the info, what do you think about using a laser thermometer pointed at that test point?

thanks
 
Rahz

Rahz

Well-Known Member
pirg420 said:
Cool thanks for the info, what do you think about using a laser thermometer pointed at that test point?

thanks
Click to expand...
It's something I haven't tried. To find out how accurate it is someone would need to use both a thermocouple and the alternate method and compare the results.
 
Abiqua

Abiqua

Well-Known Member
CannaBruh said:
echo in here...

as I said prior, RTFM, the datasheet lays out the method
Click to expand...
That is the proceducre, the method would involved calculating their version of Tj. Which isn't included in the procedure, right?

Thank you for sharing :joint:
 
CannaBruh

CannaBruh

Well-Known Member
Abiqua said:
That is the proceducre, the method would involved calculating their version of Tj. Which isn't included in the procedure, right?

Thank you for sharing :joint:
Click to expand...
How are you measuring each individual pn junction's temperature? What value does this have over the model the manufacturer has already defined?
 
CobKits

CobKits

Well-Known Member
Tj is usually calculated indirectly via Tc measurement and then using Rj-c to calculate it

Rj-c is relatively constant so the difference between Tj and Tc increases with current. at low currents, Tc can be 30C and Tj can be 32C, at high currents, Tc might be 100C and Tj 140C. you can play with citi or vero calculators to get a better understanding of the relationship. both calculate Tj for a given Tc
 
You must log in or register to reply here.
Top