I wrote the post Mr. Yewlock quotes. There are only three electrical differences between the Mk IV and the MkV:
1. The Mk IV is a regular lead based chassis. The Mk V is lead free. This is nice for the environment (maybe) but there have been some fracture issues with lead-free solder joints.
2. The Mk V has modern five-way speaker cable binding posts, the Mk IV has the old barrier strips.
3. The Mk IV has input level attenuators (pots) -- just for the single ended inputs, that previous models had. The Mk V does away with them. McIntosh says they ran out of room on the Mk V after changing to the five-way binding posts -- but I just think they realized how ridiculous they were in a modern amplifier. If you have a Mk IV and you're running single ended, you should just turn the pots all the way up (clockwise) to get them out of the signal path.
If you look at a schematic for either the Mk IV or the Mk V, you will see that the V-1 and V-2 12AX7's share a 24V filament supply. So even when operating in the balanced mode, with XLR inputs, if you want your amp's left channel to work (the V-2 12AX7 is the left channel "gain-multiplier" tube) then you better have a 12AX7 with a working filament in the V-1 spot -- but it doesn't have to be a particularly good 12AX7, or even a new one ;--)!
The Gordon Gow Commemorative Edition MC275CE was the first 275 to offer balanced inputs, HOWEVER, the circuiting was such that the single ended input wiring was still connected to the amplifier sections when the mode switch was in the 'BAL' position, so you could not switch between two sources, one SE and one BAL. In the IV and V, the SE and BAL inputs are completely independent, so in theory, you could connect different sources (or two preamps) one to SE and one to BAL. Frankly, I don't recommend it, that's what preamps are for.
The amp WILL be quieter (better signal to noise ratio) using the BAL inputs; first because (as Hifiguy said) simply eliminating the V-1 tube from the signal path eliminates a source (the V-1 tube) of potential noise and distortion. There is a second reason for the sonic improvement however, and it's the fact that the BAL input signal is twice the strength of the SE signal, further improving the S/N ratio. And there is a THIRD reason as well! In BAL mode, the amplifier's input impedance is twice that as in SE mode. This guarantees flat frequency response, regardless if your preamp's output is a little on the high side, as it is in many tube preamps.)
Casouza mentions the 12AT7's as being possibly overworked ;--) Thats not really true, but it's especially not true in McIntosh amps which are extremely easy on tubes -- they run cool and last a very long time. The main reason for this is the McIntosh 'Unity Coupled Circuit' which takes current from both the plate AND the grid of the power tube. This requires TWO separate primary windings on the input side of the output transformer(s) AND it requires TWO 12AT7 driver tubes per pair of power tubes, instead of the usual one per pair, as in conventional push-pull power amps. This gets more power out of a pair of KT88's with less stress. It also eliminates the need to adjust/maintain precise bias; in fact McIntosh amps have their bias fixed at the factory, and that's it! This is great, especially if you have fluctuating wall voltage, however it makes it MORE important IMO, to use well-matched power tubes. McIntosh denies this, but that's because they don't even TRY to match the tubes they supply with their amps, just discarding the obviously bad ones ;--)
