I suspect this unit, as most modern speed lights don't actually dump capacitor power when they are turned down. They simply restrict the current using a IGBT semiconductor so as you turn them down, the "left over" capacitor charge is used for the next discharge.
FWIW, I strongly agree. Any speedlight does not "dump" like studio flash can do. They always fully charge the capacitor to 100% (which is why we wait for recycle to complete), and then flash power level is determined by duration to quench it off as necessary. Duration would normally have no fixed meaning if it does not always start at 100% voltage. So a too-fast repetition rate not allowing for recycle completion can be a problem, but doesn't sound like that this time.
Rather than a gross inconsistency, my guess is the situation is simpler. For example, if Auto ISO was on, a Manual flash cannot react to ISO changes. Cameras can probably communicate with their dedicated flashes to know they are present and Manual, to prevent Auto ISO from changing, but a dumber flash would be at risk of suffering from Auto ISO.
This one says "Compatible for Canon,Nikon,Pentax,Olympus and and Other Digital Cameras" and does NOT do TTL, which does NOT sound like a dedicated flash with camera communication. Therefore, one good guess would be Auto ISO issues. Turn Auto ISO off for use with a Manual flash. This is Not optional with a Manual flash (exceptions for dedicated flash).
Likewise, any camera automation mode that can change aperture will screw up a Manual flash, because it cannot react to changes. Manual must use the power level we set, regardless how we might make other changes. Also set camera Manual mode with a Manual flash. This is Not optional with a Manual flash (camera A mode with fixed aperture could be an exception).
Regardless, a Manual flash is NOT auto exposure, and we absolutely MUST set power level and aperture and ISO properly and manually for any picture. There was no discussion how that was done, but again, it is Not optional.
Much more detail (for example, reported ISO and aperture, etc) needs to be known about both situations to have much clue. The Exif data will show all this. If ISO or aperture are not the same in both, that is the issue. If both subject situations were not the same, then also detail about distance and power level, and bounce or direct, etc, etc.