Was using mine in manual as well, when i had the issue, barely last week. My trigger takes two CR2450 batteries, and has two LEDs, so somehow assume one was getting the right charge, while the other was just emitting a weak light. Think Inon adjusts the light output based on a signal somehow. You can ask Pavel Kolpakov, he is a guru if it comes to strobe triggers.
Initially, one strobe was sometimes flashing sometimes not. After noticing red light on the trigger, I have then changed one (out of 2) batteries (in flash trigger) for a new one (only had one with me...), and it was working fine. After a few dives the older battery (2nd one) got a bit exhausted as well, with trigger flashing red again. Strobe, while it was still working, was emitting visibly weaker light, despite being on manual and the same settings as the other one. After i've exchanged the older battery as well, for a one salvaged from a spare Suunto, issue went away immediately.
Interesting. I have a Nauticam D-850 housing with the built-in optical trigger that uses two CR2450s. I will have to look around for some extra CR2450 batteries. In the meantime I just did a the following test with my strobes mounted to the housing and facing a white wall (diffusers removed).
1) Fully charged all my AA batteries and installed them in the strobes.
2) Verified that all the settings on the strobes match and set them both to the same power (Manual -6EV)
3) Turned off the right strobe and took an image of the wall with only the left strobe (suspect strobe) firing.
4) Turned off the left strobe, turned on the right strobe and took another image with only the right strobe firing.
5) Turned the left strobe back on and took an image with both strobes firing.
6) Swapped the fibers going into the strobes to test each strobe with the opposite trigger/fiber. Repeated steps 3-5.
7) Adjusted the power on the left strobe up 2 clicks (now on Manual -5EV setting) to visually match the right strobe. Repeated steps 3-5.
The images taken during these tests can be viewed here: https://www.dropbox....AdhjrdSUKa?dl=0
From these tests it seems the suspect strobe is firing approximately at approximately 1EV lower power (2 clicks) than the other one. After swapping triggers/fibers without any noticeable difference I still believe the issue is with the strobe itself rather than the triggers or fibers. I will try to reach out to Pavel Kolpakov to see if he has any theories.