I have one that is new, and I haven't used it much.  I had similar problems
a few times, but I think I was shooting hand-held with too low of a shutter
speed - too late in the day.  They tell you not to use the VR on a tripod,
unless you are using it for support, and don't have the head locked down.
(I often use the tripod, but leave my ball head unlocked so I can frame my
shot.)

Also, check the VR function off on switch (read the manual) - there are two
positions, two functions.

Also, don't waste your time with the LCD display on your camera.  It tells
you nothing, and is best used to view which images you want to delete.  I
never know how the shot is until I look at it on the computer.

I have gotten some good shots hand-held, but I wasn't zoomed all the way
out.  In spite all of the marketing hype, I think a good shutter speed for
full zoom would be at least 1/250 or higher (to start).  Practice your
shooting technique, and get used to the lens before you make any judgements.

The truth is that this lens will be a difficult lens to use on a D50 for
several reasons. Not impossible but difficult.
The maximum aperture of the lens, particularly at 5.6, may not let in enough
light for the autofocus to work properly, especially if you are shooting
under anything but bright sunlight and make sure the autofocus marker in the
viewfinder is over an area of high edge contrast. We are talking real world
capabilities here and not the optimistic specifications issued by the
manufacturer.
This lens will also be difficult, if not impossible, to focus manually with
a D50 in anything other than bright sunlight because the viewing system of
the D50/D70 is cheap and dim with an almost useless ground glass screen. The
D50/70 and the majority of dSLRs while posessing marvels of electronic
imaging capability are almost hopelessly compromised by their lousy viewing
systems for uses such as you propose. If you do not have experience with a
bright pentaprism viewfinder take a look through a medium to high end film
slr and you will be surprised at the difference. In fact, mount your 80-400
on one of these and I assure you the differences will be immediately
obvious.
You should sytematically test your lens on a firm tripod or other solid
support and compare autofocus and manual results aiming at the same target.
When using manual you may need to take test exposures that bracket the focus
in front/at/behind what you think is the proper focus. You also need to test
the diaphragm by bracketing the exposures through the range of f-stops
available on the lens. With a dSLR the good thing is that it only takes a
few minutes to do this and check the results on your computer (the LCD
screen on the camera is useless for this purpose).
You can use jpeg or raw: jpeg adds programmed sharpness to the image, raw
images will appear soft until software sharpening is applied.
The lens is probably fine, it just is not easy to use on a D50.

Your post is (ahem..) a bit unclear.  Do you mean to say that the pictures
are out of focus when you download them?
I have the same lens and the VR stuff is quite remarkable, I tried it on a
D-100 and the results were great.