Hi Folks. Below is a part translation of original thread:
https://www.swiatelka.pl/viewtopic.php?t=22176
I want to present the latest UV lamp by myself. Some time ago it turned out empirically that UV beam formation, even 400 nm, due to the refraction in the TIR optics is highly risky. Because if we hit a lot, where the main emission peak of the diode oscillates within 390nm or even lower, it can quickly end as below:
I had to withdraw the product and began to combine how to bite anew … It was only to put on the reflector and look for a real high power diode UV 400nm diode, with a low Vf, which can be placed centrally on MCPCB. Unfortunately, I did not find such diodes and according to my idea what was created below, that is 4 pcs of LED UV 3W, connected in parallel, with dimensions 1.6x1.6mm. Centrally placed, they have the size of an XP-L diode and even fit into the centering ring for this diode. Emitters have a very low thermal resistance (3.5C / W) and according to the DS, a maximum current of 700mA (with a voltage of 3.4V). Currently, I’m testing a prototype built on the basis of Convoy S6 shell an extended 8x7135 3.04A driver (with a temperature monitoring), so 760mA goes out to the emitter 
Full stabilization of 3.04A current is up to about 3.55V on Li-Ion cell. In 100% mode the lamp works with the power of ~ 11W and the radiation in the UV 400nm, according to the sheet is slightly over 4000mW (applying the hand to the output you can make yourself express tan 
Below are some photos. The last one is an attempt to show the spot on the ceiling (there is a small donut with Convoy S6 reflector but it’s not bad).
LEDs dedomed with no centering ring:
