We have a number of Mantis Compact at work, I think it was the first version with LEDs as light source. After a while the LEDs started to die and the light became dim and blueish, the light assembly was replaced on most of them but recently I discovered that one of the least used microscopes had failing LEDs. I found out that a new light source costs 280$ so I took the failing one apart and found it was a horse shoe shaped circuit board with 20 LEDs on it and four resistor networks and one power connector, that was all. I measured the current trough the LEDs and they crammed 40mA through each LED, that seemed a bit to much for a 5mm LED and that is probably why they were failing.
Feeling a little grumpy about the price of the replacement light and the fact they were failing because of poor engineering decisions I decided to replace the LEDs on the old board and I managed to bodge in a 1N4004 diode in series with the power connector to drop the voltage a little. The hole operation was relatively easy and the current dropped to 26mA which is still a bit high but under the absolute maximum of 30mA in the data sheet. There are a bunch of plastic parts aligning the LEDs so focusing their beams is not a problem, after screwing everything back together again it worked perfectly.
I used these LEDs as replacements:
https://www.elfa.se/elfa3~se_sv/elfa/init.do?item=75-100-21&toc=20292When I compared the old and new LEDs (under the microscope of course) I found they looked
exactly the same, the internal construction of the metal parts, the bond wires, the colour of the phosphorus layer, could not find anything different on the Sloan LEDs I used. Do the manufacturers all use the same tooling?
Other that that there have been no problems with the microscopes we have and they are really excellent to work with.