Don't think my preamp likes feeding signal to two seperate places.

Erik, note in the initial post that he is connecting the LS17 to the DS450 power amp directly, via XLR.

Treble gets a little less extended.

If the cable length from the LS17 to the crossover is particularly long, and/or if the cable has high capacitance per unit length, that could be a contributing factor in addition to the impedance issue that has been cited. The capacitance of the cable to the sub crossover will affect the signal received by the DS450 essentially the same as if the only cable present on each channel was the one connecting the LS17 to the DS450, and if the capacitance of that cable equaled the total of the capacitances of the two cables you are using. In either case, given that all of the preamp’s output connectors are driven by a single output stage, a low pass filter would be formed by the interaction of the preamp’s output impedance with the total of the cable capacitances that are present. If either cable is particularly long and/or has high capacitance per unit length the high frequency rolloff and/or phase shifts introduced by that filter could occur at frequencies that are low enough to be audible .

Would the preamps up the range not have this phenomenon?

While I don’t have specific knowledge of all of the recent high end models, just about every one of the many spec sheets I have seen for ARC line stages and preamps at www.arcdb.ws indicates a recommended minimum load of 20K. And just about every one of the many schematics I have seen shows the output connectors that are provided for each channel all being driven from a single output stage, with the RCA connector connected directly to XLR pin 2, and with the two XLR connectors (if provided) connected directly together.

Regards,
-- Al

An additional problem is that the RCA is in parallel with the balanced output on the ARC preamp. This means that as soon as the RCA connection is used, the output of the balanced connection is no longer really balanced. There's a lot here that is pointing to the preamp as the bottleneck.

OTOH, I'll take that LS17 off your hands for no charge...

Almarg and atmosphere, you have it correct.  I believe the paradigm crossover unit is 20k, SE, as per my referenced thread. (That user had the exact same crossover, but with a ref 3. It seems I would need to have a balanced crossover, as the ARC doesn't like to have both the SE and the balanced outputs running together. Imho, equipment costing this much should be built with more robust or better implemented output stages. Many of us will utilize two seperate amplifiers or in my case an amplifier and a sub crossover. We shouldn't have to worry about the output stage "choking". The difference in sound quality once the crossover unit is disconnected is amazing.  I had no idea how much I had been missing before.  I will not connect the sub until I can figure out what I will do to remedy this shortcoming, because the bloom and air is too good to give up just to extend the bass. From Preamp to crossover, I'm using a set of 1 meter silver interconnects that were made by a gentleman for me that I no longer have contact with. I do not know it's capacitance. The signal from the crossover is fed mono to the 555se with a  no frills radio shack gold single cable which is two meters.  I would assume the capacitance of the radio shack cable, and the he input impedance of the 555se doesn't matter because the paradigm crossover is acting as a buffer itself, and that the issue is from the paradigm to ARC relationship.  

Many subs can operate off of the output of the amplifier terminals. You might contact the manufacturer and see if this is the case for your model.