Audio Technica ART9 sounds awful

Just a last gasp:  The source of a phase issue (one channel out of phase with the other), if there is a phase issue, need not be limited to the connections of the color-coded leads at the cartridge; it could be downstream at any junction.  It could even be at the distal end of the very short leads between the cartridge and the headshell, if someone misconnected the leads to the pins on the headshell, for one channel.  (This also assumes you use a tonearm with a removable headshell.) I realize you are now satisfied, but just sayin'...

It's not even out of the question that the defect in your cartridge, if there is one, is simply that connections to the external pins for one channel are reversed internally, although that would be hard to imagine for a company like AT.

I offer this only because I find it unlikely that maladjustment of azimuth or of VTA could have the particular effect you describe, unless azimuth was WAYYYY off, maybe.

Safebelayer, Just to be sure you got the right idea from Almarg, me, and a few others, what we suggest is that one channel is 180 degrees out of phase with the other.  That condition can cause most of the symptoms you report. 

So, to check that problem, just reverse the speaker leads on ONE of the two channels, either one, leaving the other channel as is.  Perhaps you know this and have already tried it with no success, but it did occur to me that perhaps it wasn't clear.

For sure, no cartridge sounds its absolute best on day one, but not to the extent and with the peculiar symptoms that you report.

My point was that your symptoms are probably not due to tonearm/cartridge mismatch.

In my post above about the plasticity of the relationship between effective mass and compliance, I might better have written that one should make the calculation of Fr for a variety of values of M and C, in order to see that you can get away with some fairly unlikely combinations and still be within an "acceptable" range.  In part, this depends also upon what one considers acceptable.

Anyway, I still think Almarg nailed this particular problem. Which has nothing to do with M and C.

My money is on Almarg's hypothesis that the two channels are 180 degrees out of phase with each other.  Even if you cannot prove that by visual inspection, you might try switching leads on one channel just for the heck of it.

I am no authority on the ART9, but I found on the AT website the following: Compliance = 18 @ 100Hz; Static compliance = 35.  Wally Malewicz's web page says that the compliance at 10 Hz (which is the value to use in the resonant frequency equation) is related to the compliance at 100Hz by a factor of 1.5-2.  So one would predict a compliance of 27 to 36 at 10Hz, not out of the range of the above quoted number.  However, Wally also says that the static compliance is related to the compliance at 10Hz by a factor of 0.5, which would lead us to think that the compliance at 10Hz is about 17-18.
This topic is quite messy.  As a person who likes to boil things down, I note that the resonant frequency is inversely proportional to the square root of the product of M X C, which actually leads to the finding that there is quite a lot of tolerance in the relationships that allow one to end up with a tolerable resonant frequency. (Try plugging in a range of real world values for M and C and take the square root of the product, to see what I mean.)