There are many ways to proceed. First, it is important to realize that you do not have to record everything as audio all at the same time. The fact that you are using a Windows computer should not prevent you from getting the most out of your recording experience.

True, you can only connect and record audio from a single audio interface device at a time when using a Windows computer as your recorder. This does not prevent you from adding audio from a different audio interface at a separate time. Just pay attention to sample rates.

Use whatever audio interface that allows you to setup and record your eight drum microphones, as you require. For example, the Steinberg UR816C, or whatever. You will be recording the drum audio direct to your DAW on the initial session - using that audio interface and its Driver. The MONTAGE will simultaneously be recorded as MIDI data, either to your DAW, or to its own internal Performance Recorder. Your choice!

Depending on the complexity and what you want to do to the composition in post-production, you may want to setup the DAW as master clock and use it to provide conductor tempo... doing so will make it possible to use clock synchronized effects, LFOs, Motion Sequences, etc. The DAW metronome can be fed to both musicians during the “live” recording session.

By initially recording the drums as audio and the MONTAGE as MIDI, will allow you to concentrate your audio energy on getting the drums recorded properly. By setting up to record every nuance of the MONTAGE performance as MIDI, you can render this MIDI Data to audio in a separate session, later... when you are ready to render the MONTAGE as audio, set it up as your computer’s audio interface. The MONTAGE will be able to playback your recorded drum audio tracks, while you setup the MONTAGE to render the keyboard performance to audio Tracks.

You are stuck because you believe all the audio rendering needs to be done at a once, in a single pass. It does not. By recording drums as audio, MONTAGE as MIDI, you can render the MONTAGE MIDI to audio tracks, separately (at a later time), using its 32-bus audio features.

Simply setup your DAW or your internal MONTAGE Recorder to capture the synth performance as MIDI in sync with your drums.
You find that playing back recorded audio can be handled by any audio interface, not just audio that it records. You will be playing back the drums but you will be setting up to render the MONTAGE Parts to the DAW as audio.

Just prepare the initial audio Record at the same sample rate you intend to use for the MONTAGE audio (that will just save you time and having to convert it later).

Once you have rendered the MONTAGE Parts to separate audio tracks as you desire, you can again use any audio interface to play them back through. You could again setup your other audio interface... the device acting as the audio interface is the one sending and receiving audio to the DAW and is responsible for routing incoming audio to the speaker system.

Extra Possibilities:
This approach let’s you get the most out of each device. Just because you have 8 MONTAGE Parts does not mean you will wind up with just 8 audio tracks. To get the most out of the 32-bus audio capability you have think differently. For example: A Single Part Drum Kit could generate separate Kick, Snare, Hihats (3), Toms, crash, ride, cowbell, and individual percussion sounds and effects etc., etc. — each could be assigned to a discreet audio bus output or routed together in groups.

Don’t limit your thinking that one MIDI Track always equals one Audio Track... that does not necessarily have to be the case with MONTAGE. In general, you can take an individual Output for each sound or sound effect you generate within the synth engine. The benefits of using the MONTAGE’s powerful audio engine (interface) are often specifically designed to enhance the MONTAGE experience. And since you have the luxury now of rendering audio separately from the task of recording the 8 drum mics, you have the time to get creative with how you render the MONTAGE data.

Hope that helps.

You will need an auxiliary software for to route signals from different hardware or software sources to a common output like the SP/Dif conector of the computer.

Don´t expect detailed explanations about your need in this forum due it is exactly the type of features or usages that Yamaha / Steinberg reject to comment or to implement.

These two programs seem apropriate for your purpose:

DS WASAPI ASIO Router Mixer

VB Audio Voicemeeter Banana

Also an ASIO multiclient driver can do the job

In the past there was a Steinberg 32 bit ASIO Multiclient driver (in facts it is actually an ASIO Multiclient Wrapper) designed by a company called Vidance that you can find and download at "http://www.vidance.com/asiomulti/" but it only works with 32 bit systems.

The german factory RME use to develope audio hardware with multiclient drivers .

You also can do what you want using independent audio boards that don´t use Windows drivers, like those that are based in DSP chips and are complete audio systems in themselves (very rare these days), under the condition of to have a computer motherboard that can support it.

Universal Audio and SonicCore have systems of this type but be aware about the tech support you will receive due it is a lonley territory

Sorry but I can´t post the web addresses due my post is rejected as spam, you will need to search in the web by yourself.

Additional considerations about multi board systems (of different marks):

Combining different audio systems in a single computer is not so simple and several problems can happen.

1) The systems can have different latencies.
2) The systems can conflict because they try to access to the same hardware or software resources.
3) Most of the DAWs are not designed for to work with multiple audio systems, particulary true in ASIO drivers and since Windows Vista.
4) Don´t expect to have the same efficiency than with a single audio system.
5) Everything should be tuned very carefully and this takes time.

thank you mister bad and Pablo for your answer; yes record in midi is the solution, capito:) . and for monitoring, how do you proceed? parrallele monitoring, one for the montage audio and one for the computer onboard card?

Suposing that you monitor in stereo, all your sources should converge to the same pair of speakers thru the most neutral output you could have in your system.
Usually, for monitoring, is used the digital stereo output of the system (SP/Dif or AES/EBU) due is the most neutral output.
Of course you need also a good pair of powered speakers with digital input compatible with your output.
How good or how bad will be the sum of signals is impossible to predict without having in front the concrete combination of devices you will use.
Consider that the worst problem that happen combining digital signals is related with slightly phase offsets and jitter, so the use of a master clock is a must if you want professional and on phase quality.
With the exception of very rare and expensive servers, the clock of the computers are not enough stable for mixing succesfully large amount of tracks.
Remember that the stability of a digital clock does not depend only of itself but of the auxiliary circuits that provide power and interface with the rest of the machine.

thank you Pablo for informations.