Should I eliminate my preamp?

The OP was asking about getting rid of the preamp in the case when the DAC has a volume control. In that case, there is no downstream volume control, and no need to provide a low impedance buffer for the interconnect, as the amplifier will take care of that.

This is correct, and the source having a digital domain VC it will have a low impedance output, probably lower than many preamps especially tube ones, and it will drive any interconnect just as well if not better.
Besides, driving interconnects with a 10kohm passive is not an issue either, unless the interconnects are"low quality high capacitance" and are 3mt or more long. I can give you the maths if you want. 

Cheers George

"I can give you the maths if you want."

An offer that is hard to refuse, if not too complicated for my pay grade... 

I like 10kohm pots as passives, as the can be driven by just about any solid state source, just some high impedance tube sources have trouble, and being 10kohm as a voltage divider passive they have the lowest output impedance at worst (mid possition) around 2.5kohm.

With 2mt of poor quality interconnect with high capacitance at (100pf picofarad per ft) that 2mt is approx 650pf total

6.28 (2 pie) x 2500ohms (2.5k) x 650 (650pF) ee to the -12 1/x = gives a very respectable -3db point at 98khz! and this is the worst interconnect at 2mt at the worst mid volume setting.

So as you can see driving the capacitance of poor quality 2mt interconnects with a 10kohm passive at worst volume mid position of 2.5kohm Z is not an issue at all as we are at only -3db at a still very high 98khz!

With one meter of that same interconnect, it becomes -3db at 196khz!!

With one meter of better quality interconnect which has only 50pf per foot then the -3db is at 392khz!!!!   

Cheers George

A good preamp as DSP? A joke.

"I can give you the maths if you want."

An offer that is hard to refuse, if not too complicated for my pay grade...

Here is some additional math:

With a 10K pot in series with the output coupling cap, a timing constant is introduced.

The formula for a timing constant is this:
1,000,000/Resistance x capacitance x 6.28 = F

Resistance is in ohms, capacitance in microfarads, F is frequency in Hertz and 6.28 is Pi x 2.

If you have a solid state source like a CD player, quite often it has a coupling cap at its output (probably an electrolytic capacitor). Depending on the designer's expectations, the cap is probably between 1uf and 10uf. But without knowing that, if we do know the amplifier's input impedance which has to be included, we have enough information to draw some conclusions.

If the amplifier input impedance is 10K (worst case scenario) we can see that the timing constant is raised an octave if the volume is all the way up. Depending on the design, this can introduce phase shift in the lower frequencies, which is perceived as reduced impact.

At volume control settings at less than full volume, a series resistance is introduced by the control. This effectively raises the output impedance of the source as far as the amplifier is concerned. To reduce the volume by 6 db, the control is at mid position, with 5K driving the output and 5K to ground. The series resistance thus introduced with such a low impedance load of 10K at the input of the amplifier can be audible as a reduction of dynamic punch at all frequencies.

Now the reason I mentioned the formula is that with different controls, the timing constants are different. Sometimes it works out well and sometimes it does not. But in all cases the interconnect cable is uncontrolled by such a high output impedance so its artifacts will be much more audible than if driven by a lower source impedance.

So audition is important- and its important to remember that while it might sound alright with a passive, it could well sound better (less colored for the reasons stated above) with an active. If you spent some money on the amps and speakers, it seems a shame to me to not get the most out of them.

This is an example of you get what you pay for.