Umm, the voltage does not matter, as long as you give the driver what it needs. The supply voltage to the driver under load has to be more than the driver's minimum input voltage.
If you do give the driver the voltage it needs, it will regulate the current by giving the load just enough voltage for the desired current to flow, and adjust this voltage, in order to keep the current constant. You can change the input voltage, and as long as it is within the driver's input range, the output voltage will not depend on the input voltage in any way - it will always be exactly what the load needs for the current you set.
If you give the driver too little voltage, things become dangerous! You may think the diode current is 200mA, but in reality, your dummy may have a lower Vf than the actual diode. Because the dummy has a lower Vf, the driver was able to regulate the current through it. But as you then switch to the actual load with a higher Vf, the driver will not be able to regulate the current through it! This is where things become interesting. As the diode works (at a lower current, than you think it works at), it slowly warms up. It's Vf drops, and as it drops, the driver can suddenly pass more current through it! After this it warms up faster from more current, and passes even more current, and theoretically, it could drop in Vf enough for the driver to start regulating the current. Or it could just climb towards the current that is actually set.
The really dangerous part is this: If you "set" (or rather think you set) the current through a dummy when the driver does not get enough voltage even to regulate the current even through the dummy, the current you see on the meter will be lower than what you actually set the driver to. As you then connect the diode to the driver, thinking that the current is where you want it, the driver will try to do it's best, but the voltage will just not be enough. If you then suddenly do give it enough voltage, the current will shoot up to where it is actually set to!
A driver out of regulation with the Vf dropping with heat, allowing more current to flow, could explain the power climbing. As could a combination of this and the wrong current setting made with a driver out of regulation.
This is why i asked you how sure you were of your current. Do you really know if it was 200mA? Was it 200mA from the start, or did it climb there as the diode heated up? Could it have been set higher (set outside of regulation)?
This is why it is VERY important to verify that the voltage is in the right range when dummy loading. If you set a constant current driver under the right conditions, the current will always be the same. If you set the driver under the wrong conditions, the current could actually be set much higher.
But you can test this, if you haven't changed the driver setting yet. Put a dummy load on the driver, and measure the current. Then make sure the driver is getting what it needs (whatever that is, different for every driver, with linear drivers it depends on the load).
To simplify: When you gave the driver 5V, the current was not 200mA. When you gave the driver the right voltage, the current was 200mA (if it was set correctly, under the right circumstances).
This could be the reason for your diode being dead. So it is quite important to set everything up correctly, or at least to figure out what happened, or the data is useless. If your power changes with the drier input voltage, then you're doing something wrong - the current is not what you think it is.
If you haven't turned the pot yet, test the driver on a dummy with a high enough input voltage. It probably is 200mA, and the climb was just from a funky diode, but it would be good to know for sure.
I practiced on the IR diode and I think I found a really good way to remove the diode with any visable damage. I still have not touched the red or BR diode though.
