Hi Jorgen,
You have a very good understanding of the situation. We will try to fill in some of the information... hope it is helpful.

First, the driver for the MX-series and the AG-series of mixers is the same, Yamaha Steinberg USB Driver. It will identify the connected product(s). This will be able to handle both products, don't let that confuse you - each product speaks the same language. The driver "knows" what Yamaha and/or Steinberg product is addressing the computer. That same driver handles USB/MIDI from an entire list of both Yamaha and Steinberg USB products... from simple stereo through 34 audio channels.

You simply need to download and install the proper one for your computer and operating system. In your DAW software you determine which one of your devices is going to act as your *primary* audio interface. The *primary* interface is the one that connects to your monitor speakers. It will be the one distributing the audio.

The type of computer (you didn't mention) will determine, to some degree, your options.

If you have a Windows PC you can only address one of the two devices as an audio interface at a time. This means a workflow where you record your music tracks using the MX as your audio interface (as you mentioned you've done before getting the mixer). Speakers connected to the MX. Then, later on, you reconfigure your setup to use the AG mixer to record your vocals, etc. in this configuration you would switch to the AG as the primary audio interface. This means connecting your speakers to the AG, as it becomes your *primary* audio interface.

If you have a Macintosh computer you can construct an aggregate audio device that includes digital (USB) input from both the MX and AG simultaneously. This means a workflow where you record your music tracks using the MX (as you've always done) but you can also simultaneously record vocals via the AG mixer. The 'aggregate' device simply allows both USB audio devices to send audio into your DAW simultaneously. (You may or may not need this functionality).

In both cases, it is important to realize that when recording (particularly with an open microphone) that your speakers may need to be turned off while the microphone is active. Therefore, hearing the keyboard and the vocal mic (mixer) must be achieved by connecting the analog Output of one to the other... for monitoring purposes. Both devices need to feed your headphones... so one must be connected to the other.

Basically, only one device can address the speakers (phones) directly.

Say you are overdubbing vocals... a mic connected to the mixer is routed digitally to the computer via USB. The music tracks in your DAW, if still MIDI data, will need to be heard... the MIDI data returns to the MX, the MX audio output will need to be heard. You could connect the analog Output of the mixer to the Aux In of the MX (with appropriate adapter), or connect the analog Output of the MX to Inputs of the AG mixer.

You have to connect analog audio between them!

If you have already converted (rendered) your MX tracks to Audio renderings, you will not have to worry about this interconnection for monitoring. When you mixed down your MX tracks as audio, it becomes a simple matter to monitor them (even if the MX is not powered On) because they exist as digital audio in the DAW, and can be routed via USB to the AG for monitoring.

Which is Best Practice?
It's a great question...
Some people worry about losing audio quality.
Some people worry about the hassle of having to switching primary audio interfaces (which means physically changing the cables going to the speakers).

Which is better? Let's consider the question...
Losing audio quality. Is a direct digital connection between MX and computer better than connecting analog out to the AG first? I'm sure it could be measured. Whether or not anyone would ever be able to hear the difference, is the question. I say, try both workflows. See just how much loss of quality there is. See if anyone can hear it.

In any analog transfer... There is a potential for error because *you* have to gain stage the signal properly (there is this human factor) - but assuming you properly set the levels... just how much difference is there? I recommend you do that experiment. Bring in a friend to listen to recording done each way... see if they can tell a quality difference. (it's small). Do a true blindfold test, don't tell them which is which...

This will put it into perspective.

Switching Outputs between audio interfaces, is this better?
Sure, I think so. But this for some is a logistical nightmare. It means powering down the speakers, perhaps moving gear, disconnecting cables... etc. for some users they would rather not. (me, I'm an engineer, I look at it as "I setup for what I need to do" - I've accepted that a computer is a device I *can* configure to record, but the key point is ... it wasn't designed specifically for the task, so reconfiguring is apart of the ticket I purchased when I bought in on using a computer.

But I do understand (and I am sympathetic) when I hear from users about what a major hassle this is... and it is if you don't prepare for this - and this is why, your question is a great question.

I always setup my studio with this possibility in mind! (Plan ahead). Either way is a reasonable workflow. A lot depends on how you go about recording. In a one person studio, you gain nothing from having both devices combined... when your workflow is music tracks first, then overdubbing audio, you can design a workflow and configuration that makes you comfortable. Because you've worked out already how it is to work with just the keyboard and music tracks (without the AG06 mixer), adding audio is not going to complicate things to much.

Thank you Bad Mister. As always, you provide us Yamaha synth owners with detailed and easy to understand explanations.
Your support is highly appreciated. When I had the MOX many years ago, I had great help with your video tutorial that came with my MOX. Made a lot more sense to me than reading through all the manuals.
I do not think I will mind to much, reconnect the cables between the MX and AG06, so I will probably use this option, but I will try out both alternatives and see if I can hear any difference in quality.
Respect!!!
/Jorgen