Of course the output may be acceptable.
The point is: The black ink is lay down in a particular way.
Not at all comparable to the gray square.
On my 9900 the back ink is lay down in such a particular way... the output is affected.
On both machines the issue is the same.
Mine is worse and effecting the output.
That's the only difference.
This is of some importance in troubleshooting this problem.
Maybe, maybe not. It's worth looking at, but if the output is acceptable and as expected on one machine but not another, and the "wave" pattern you're seeing from the multipass process is the same, it suggests to me some other issue.
I think you may be too focussed on what you think is wrong/unusual/different in that pattern and look at more "traditional" fault causes for your issue.
Personally, I would isolate the source data (ie try a different computer, change cables/connection type), change location (to ensure there's no external factors), change software and reinstall latest drivers, test strictly on Epson media with default driver settings for that media, change black ink cartridge, look for anything that might affect paper feed physically, test at all output resolutions with both bi-di and uni-di, print only completed prints (don't stop them part way through - I think it's distracting you from finding the real cause), ensure all calibrations (including colorimetric) are completed by an Epson technician, latest firmware.
I know you've done a lot of testing already, but again you seem focussed on an issue that isn't bringing you any results. So move past that, return to more traditional troubleshooting techniques and them investigate the results.
Whatever you might think, there's ample evidence that this isn't a widespread problem. After the initial tests that I did, I tried other media and still can not replicate the problem, for example.
Regarding replacing the printhead twice, it simply suggests to me either the first replacement head was faulty or the technician didn't perform the repair and calibration to spec as opposed to there being some actual benefit to replacing the head twice. If replacing a head becomes an option and doesn't work, I'd look at the mainboard and the power supply and ensure all the ribbon cable connections are individually checked as well as all the ink lines from cartridge to head (particularly seals).