If a device is produced as standard, I expect its standard software to show the same result on every device, unless there is a hardware problem.
It should be noted that your "problems" or rather, in this context, phenomena could not be reproduced by any other SDS3000X HD owner, so the standard is given.
I think that such problems should not be solved by things like changing the calibration procedure, that is why I am surprised.
Shortening the calibration time is also not the solution to your phenomenon, but simply a shortening of the calibration time.
Anyway, as a result, my problem was solved, that's what matters.
Updating the firmware is like reforming the memory addresses.
A possible abnormality is virtually overwritten, which is rare but not unusual.
We have also experienced this phenomenon at work.
We programmed a circuit with a controller on it, then tested it and one thing didn't work as expected.
We did some troubleshooting, found nothing, then programmed the controller again with the same file - and it worked perfectly afterwards.
That's the way it is sometimes, wondering for a moment, then accepting it in the end because there's nothing else to do.