[PlantedRich]

The new impeller should also come with rubber bushings if I remember correctly. Make sure there is one on both ends of the shaft as it is easy to lose the one out of the bottom of the well and not see it missing. The Eheim shaft is ceramic so it will not be bent. Breaks if dropped but not bent. Another place to clean is inside the impeller itself. this is a spot where algae/grunge can collect and then a bit of sand can stick in that to wear and score the shaft.

 

[PlantedRich]

Okay if the bushings are on each end and the cover fits, you are likely to have it placed right as the cover won't fit otherwise. The bit of "pop" as it fits the bottom bushing sounds good as it needs to be tight. The blades should be loose on the other parts. The mag drive type motor has so little torque to get moving, this is a way to let the magnets get just a tiny bit of start before the pressure from the blades is added. A good look at the blades might help spot a nick or chunk missing? Not unbalanced?
But then it is possible for the shaft to wear quickly if a bit of sand gets stuck between it and the impeller. I check for wear by looking but then I place the shaft on a flat surface with a bright light like an LED behind it. Slowly rolling the shaft and looking for spots where the light shows under can help spot wear that you might miss just looking.
If you spot nothing yet, it is possible to take the magnet off the plastic part to look at it. Squeezing somewhat carefully on the knobs at the end will let them slip through and come out. No tiny little fine parts but just don't break the plastic?
Since the impeller is the only moving part other than water, you about have to be looking in the right place.

 

[PlantedRich]

If the impeller gets pulled to the side of the well when you hold the shaft centered, there is a problem with the bottom. Holding the center with a finger so that the sides are not touching, you should be able to spin the impeller.
At this point a good bright light like a little LED flashlight to shine down in the well to look might be my next thing. The problem with the warranty idea is that you kind of need to know what it does so that 4-5 years from now you will know if it does it again. If the bottom rubber bushing is lost in some way, knowing that will be very helpful. And I have to admit that is one of the things that bedevil me at times.
Most of the time the bushing stays in the well but when it comes out on the shaft and I don't know it, I have wiped it off in the rag while cleaning. Then when I put it back together, I get noise until I look in the rag and find the bushing!
The "feel" may feel right but, who knows, you could be sticking it down in a snail shell instead! I try to be smart but at times, I do really dumb things like that.
The blades should be loose to spin part way. No real drag but a solid stop when it reaches a point on the rest.

 

[PlantedRich]

At this point it might be worthwhile to move on to some testing to be certain where the noise is being generated?
I had not suggest it earlier as it seems to be a lot of trouble so your choice on whether to go with it.
Can you see a way to take the top of the canister, just the lid part with the impeller and all installed and hold it in a bucket, pan or tub so that the impeller can pump water as normal but not shoot water all over the room. I've never done this so give it some thought as to practical or not. What I'm thinking is that we may need to be certain that the noise is really from the impeller rather than noise from air somewhere. I trust you to hear the difference but then I also know I can fool myself at times.
Somebody once said to "trust but verify"?
Maybe the head in a bath tub with enough water to fully cover the lower head to the normal level and then weight the tubing down under the water so it draws normally and the output fixed to catch the outflow water to leave it in the tub and not all over?
Once you get the head set up, filling the inflow tube at the faucet until it runs out the other side might act the same as priming? Let it run this way for a bit to see what it does?
See why it was not the first thought!! Desperate times require desperate means.

 

[PlantedRich]

As a second step of the testing I proposed, I would want to start looking at ways air might be getting into the impeller. But then the way I search for trouble is to first be certain that I'm looking in the right spot. When you describe it running quiet and then after some time becoming noisy, it doesn't really seem like an impeller/shaft problem as much as an air problem. Some other less likely ideas for noise is that the media is packed too tight (or dirty) so that it creates problems. Something weird like a piece of loose media down in the outlet elbow where you don't see it but it rattles around and moves at times?
But you seem to have ruled out some of those. At times, I have to step back and redo the search.

 

[PlantedRich]

Reason for some of my thinking is that I did have a "barrel" get into the elbow and give me some fits. It would block flow and then when I cleaned it would move and flow would seem better until it moved again. And I have run into putting too much media in the canister on a 2075. That was a problem getting the thing to prime and run quiet.
So past experience with lots of equipment has taught me one thing.
When gizmos are running right, we can say what they are doing. When they are not, they can do anything they want and let us try to figure it out!