I did some light load bench tests, before I fried my driver. The most I did was 8W for 15min in free air, with an XHP35.
No problem for the driver at all. The XHP35 was a few centimeters next to the driver, and was shining on it a bit (when I blocked the light shining on the driver with my hand, it got pretty warm). Even with that, the driver got just a bit warm. If I had to guess, it was around 40°C (with 24°C ambient).
The board that I used was the first revision and had bad thermal properties. The newer boards should do better.
There are multiple values given in the table with various resistor values, but when you scroll down to the graphs, there are two called:
OUTPUT-VOLTAGE SWING HIGH vs. TEMPERATURE
OUTPUT-VOLTAGE SWING LOW vs. TEMPERATURE
Now these values from the graphs are much lower than those from the table. For the “table values”, there is a resistor RL connected to the OpAmp, but I don’t know how it is connected.
Is RL connected between the output and VDD/2, to show how near the OpAmp can get to the rail with a load on it, and the graphs, when there’s no load?
I am far from an expert in data sheets but I have learned that the graphs generally do show better numbers then the official ratings. This is due to the built in error margin that the companies build in the way I understand it. Also most of the official ratings are max ratings, so what they guarantee the components will not go over.
Thus the real world specs can generally be a fair amount better in most cases and these are what you see in the graphs.
In “real” manufacturing they design off the max ratings and then just get a nice bonus if the parts preform better.
Also I noticed this at the top of the graphs page :
So looks like it has some resistance connected in the graphs as well.
Yeah, I am not sure this is the case, just my best guess. I have seen pretty big differences before and the ratings are listed as max ratings. But that is a bit extreme I admit.
The new 17mm high-power boost driver will open up some interesting possibilities for triple lights. It will be possible to use high-Vf LEDs like the Cree XP-G2 and Osram Black Flat in a triple configuration with a high current through each LED.
Yep, I plan to swap out the driver in my EDC with this one for just that reason. I still have a hard time finding a better tint then my mixed tint EDC. Although the 90CRI XHP50.2 I have been playing with looks promising, if I could find it in 4500k it would be perfect.