Not to nitpick but an FP24 is the same as a Mixpre, not an MP-2. The FP24/Mixpre gain can be turned all the way off; not so with the MP2.
My apologies, I seemed to remember the gain structure being the same when I owned the MP-2. Regardless, I think the 680 can take a pretty hot input & attenuate at the deck. When I first started using my FP24 I was running 18db attenuators at the 680 with the pads engaged at my mics, something didn't sound quite right on playback & I found the gain was very touchy on the FP24, one little adjustment & the preamp would overload. Recently I started to instead use the attenuators at the input from the mics into the FP & found its allowed me to use more of the gain on the preamp without fear of overload & the sound improved to my ears. Interestingly while I was searching the term attenuators on here a few months ago I found a post from resident wiz DSatz where he mentions the use of the pad on microphones does reduce input db but also affects the noise floor which may explain what I heard in my previous recordings, not sure but I definitely felt something was "off" when I used the mic pads, once I switched the location of my attenuators & stopped using the pads on my mics I noticed an improvement in my recordings, YMMV of course. What I also found interesting is that I didn't have to attenuate the the difference on the 680 of what I used to use with the pad on my microphone vs not using the pad.
Edit, DSatz informative post below.....
Roving Sign, the "pad" switch on a condenser microphone can be designed in various ways, so you have to find out on a case-by-case basis what the switch for a given type of microphone is actually doing. The great majority of such pad switches reduce the voltage at the input of the microphone's own electronics, i.e. the signal delivered by the capsule itself. This is done either by reducing the capsule's polarization voltage, or more commonly by placing an additional, constant capacitance across that of the capsule.
Either way, such pads reduce the microphone's overall output level while allowing the microphone to handle higher sound pressure levels. That in turn helps to prevent overload at the input of the mixing board or recorder, so why not leave them on all the time? The answer is that the noise level of the microphone's own amplifier circuitry stays roughly constant regardless of the pad switch setting. If you reduce the output signal levels by 10 - 12 - 14 dB while the noise floor stays constant, you're basically adding 10 - 12 - 14 dB of noise to that microphone's signal by the time you bring that channel's gain back up the level you wish to record at.
Thus any built-in pads of that type (which, as I said, are the most common type) should never be used unless they are needed to prevent the microphone itself from overloading. Unfortunately a lot of engineers don't seem to get that. There are studios I've been in where the KM 84s and U 87s all had their pads switched on all the time--because years ago, some input on some board may have been driven too hard, so the engineers simply left all the pads switched on all the time since then, to prevent possible future problems.
If the input gain trims on the board can't be set to where they are overload-proof for a given microphone, then the best remedy is a resistive pad (attenuator) at the board's mike input. Such pads reduce the noise from the microphone's circuitry to the same degree that they reduce the wanted signal levels, so they prevent overload distortion without making the recording that much noisier. Resistive pads are inexpensive and reliable, they don't color the sound, and they don't interfere with phantom powering. Every engineer who does live recording should have some on hand, both as trouble-shooting devices and as real-world problem solvers. And then they can leave the microphones' pads switched off when their use isn't strictly necessary.