The "DC cut" option is implemented differently in the 880ex and the 1680/1880.  


880ex:
The original 880ex did not have an option for "DC cut". It was added later in an operating system update (possibly 2.012) .  
	In the 880ex, the  "DC cut" only seems to affect the signals going out of the digital outputs.  The analog ins & outs, digital inputs, mix bus, and record buses are unchanged by "DC cut".  This means "DC cut" will not be recorded, but is only affecting playback at the digital outs.  
	"DC cut" acts like a high pass filter. In the 880ex only, it begins to noticeably attenuate frequencies starting around 200 Hz.  It is -3dB around 50 Hz.  It is -9dB around 20 Hz. It appears to resemble a normal, first order, high pass (low cut) filter.  In fact, the typical 90 degree phase shift associated with first order filters was clearly visible.


1880/1680:
In the 1880, the "DC cut" seems to be applied to the mix bus. (This result agrees with the 1880's manual as I discovered later).  Only signals routed through the mix bus are affected. With normal track recording, this means "DC cut" won't be recorded, but will only affect playback. (Presumably, if the mix bus is routed to a track, "DC cut" will affect the recording of that track.)  The analog and digital outputs are both affected by the "DC cut" option (this is the main difference from the 880ex where only the digital outs are affected).
	In the 1880, the "DC cut" affects the signal differently than in the 880ex.  In the 1880 it is also a high pass filter, but with a much lower corner frequency.  In the 1880, "DC cut" starts to noticeably attenuate the signal at around 70 Hz.  It is -3dB at around 20 Hz.  It is -7 dB around 10 Hz. This also appears to resemble a normal, first order, high pass (low cut) filter.  Once again, the typical 90 degree phase shift associated with first order filters was clearly visible.  According to others on VSplanet, the 1680 behaves like the 1880.


The problem with first order filters is the big phase shift. This phase shift starts to "muddy" the bass frequencies at 10 times the corner frequency.  Theoretically,  this means the bass begins to "muddy" at 500 Hz on the 880ex and 200 Hz on the 1880.  Obviously, it is best to turn off the "DC cut" unless it is known to be absolutely necessary. 


In order to determine the effects of the "DC cut" option, I performed several experiments using a VS880ex, VS1880, sinewave generator, frequency counter, and an oscilloscope.  

While performing the experiments, I also verified the frequency response specs of the VS880ex and the VS1880.

VS880ex:
Running the test signal into the analog ins and measuring the analog out on the oscilloscope, I found the VS880ex seems to live up to its advertised frequency response specs.  However, I did note a few observations that may cause a little trouble on the low end. 
	First of all, I noticed the overall time delay (or latency) between analog in and analog out (with no recording) to be a pretty steady 1.15 milliseconds through most of the audio spectrum.    
	Next, I measured the high frequency response (44.1 Khz sampling rate) at -.2dB at 20 KHz.  This agrees with Roland.
	On the low end, it went so low (well below 10 Hz) that I couldn't actually measure a value for attenuation from input to output. However, a few things were noticed. First of all, at low frequencies, the time delay began to reduce.  This indicates a probable phase shift.  I measured 40 Hz at 1.07ms (instead of 1.15ms).  20 Hz was at .88ms This is only 2 degrees at 20 Hz and is pretty insignificant. 
	Also, I noticed the analog input started to load down the sinewave generator at low frequencies.  This significantly reduced the output of the generator's low frequencies even though the input to output of the VS showed no attenuation. Because of this, there may be a percieved loss, in bass response, in the VS, when it is actually being lost in a mismatch in impedance (from source output to VS input) or a weakness of the external sound source. This is one reason why an external preamp is probably a good idea.  

VS1880:
The results were basically the same on the 1880 except for the latency value.  It was actually slightly higher at 1.63ms.  Also the phase shift on the low frequencies was slightly higher. At 20 Hz the delay was 1.24ms, which translates to about 3 degrees (also basically insignificant).  



I guess the main lessons I learned from this are:
1. Avoid using "DC cut".
2. Use a decent preamp to be on the safe side for low frequencies.