The spec sheet had no thermal specs.
I found another diode with similar package (SMA).
The calculations will be based on it's specs below.
Specs for SMA Package
Here are some precalculated values
This seems really confusing ..
If we use,
we get $100^{\circ} C$ per watt
If worst case is 100% duty cycle we now get, $200^{\circ}$ per watt
Thus at 0.5 watts we get a final temp of $125^{\circ}C$ assuming $25^{\circ}C$ ambient
Let's do the calculations manually,
For the 'leads'
:opt no-lint
import Text.Printf
degCperWatt = 645*125/11.8
putStrLn $ printf "DegC per Watt for Leads Min Pad Size %8.2f deg C per watt" degCperWatt
DegC per Watt for Leads Min Pad Size 6832.63 deg C per watt
Well, the leads are no help .. !
Now let's try junction to ambient ..
Spec says $270^\circ C$ per watt,
So at 0.5 watts we get a final temp of $160^\circ C$.
Not sure why it's different .. The SMB package is a bit lower -> final temp of $140^\circ C$. Maybe that's why.
I tried an online calculator (not completely sure I'm using it properly) and got,
At least we're in the ballpark ..
There are two questions remaining ..
Three choices from Jay for 'analysis' but I found an interesting axial diode that's cheap.
Uses DO201AD Axial Package .. And it's cheap!
If we mount the diode with one centimeter leads on each side and about 3mm in the air we can achive about $80^{\circ}$ C per watt. We could probably live with 5mm lead length on both sides.
At 0.5 Watts,
$40^{\circ} (diode) + 25^{\circ} C (ambient) = 65^{\circ} C$
If we assume absolute worst case is $100^{\circ}$ C then the
reverse leakage power is about 50mw .. No problem
I consider this a success.