Fuel pump bewilderment and confusion

I've been chasing a dropping / inadequate fuel supply problem for a week or so and I'm completly out of ideas at this point.  As I gradually deconstructed the whole fuel system to try and isolate the issue, it seemed like I was getting inconsistent results.  To the point that, having not written down the test setup and results, I started questioning what I had done on previous tests.  

Fast forward to today, and I started the day with a new fuel pump.  But even this has devolved to where I am now.    My test setup is:  5gal bucket of diesel -> hose -> strainer/pre-filter -> fuel pump -> hose -> "calibrated" bucket (marked every 1/4 gallon).   In addition, I have a voltmeter on the terminals of the pump, and a clamp-style ammeter on the 12V wire to the pump. 

When I jump the fuel pump relay, it initially seems to be flowing great.  In the neighborhood of 1gpm, which is what I'm expecting.  Voltage is solid at 12V +/- 0.1V.  Ammeter shows about 4.5A.  But over the course of 1-2 minutes of continuous running, the flow gradually diminishes, while the amps gradually drops to more like 2A.  Still 12V though.  If I stop, wait like 5 minutes, and try again it does pretty much the same thing.  If I wait like 30 seconds it starts off stronger, but not what looks like "full flow".  In my 30 sec rest example, the current starts off on the 2nd round at about 3.5A instead of 4.5A and then gradually starts dropping.  

Under a normal setup, I could imagine heat causing this or the tank not venting adequately, creating a vacuum situation in the tank.   But neither of those make any sense in my dumbed-down test setup.  

So I'm completely flummoxed.  Any ideas, either for what's happening, or for another test?  I don't want to put the whole thing back together until I have some confidence it's working right.  Which it clearly isn't now!

Electric pumps don't suck.  They need to be in the fuel, or fed from a low pressure feed pump, or pulling fuel from a large diameter fitting in the bottom of the tank. If your inlet filter/screen is anything more than a strainer, it won't pull through that.  A foot or two of lift is often beyond their capability.

Your amperage is dropping to the electric motor portion because the pump portion isn't working as hard, because of the sucking situation.

This is a lift pump, meant for at least partially pulling.  The installation notes that came with it say "make sure pump inlet is no more than 24" above top of fuel tank".  As it is it's probably about 12". 

That also wouldn't explain why this setup was working for about 15 years. 

You are correct.  I shall change my advice to, "Well.  That's berkeleying odd."

Your comment did get me thinking harder about it, though, and I'm wondering now if my test setup is part of the problem - my pump is in its normal spot, but instead of pulling from the tank, it's pulling from a bucket, which is ... on the floor.  Which is definitely lower than where the tank usually sits.  It should still be within the lift height, but like you suggested, the drop in amps suggests it's pumping a lot of air in with the fuel.  

Modified testing setup in progress...

Use bench supply or car battery to rule out electrical system issues?