To the best of my knowledge, arpeggios are attached to voices, not waves. So your best bet is to just define a voice that uses the wave, and then attach the arpeggio to the voice.

To expand on what I said: If you have a wave, all you have is a wave. Now what happens if you try to attach a arpeggio to it? Let's say it's a really simple one like "play the note 3 times". The arpeggio triggers, goes to play the note the first time and...

What does it use for the envelopes? What's the attack speed? Is it velocity sensitive? All that sort of info is attached to a voice, not to a wave.

Let me re-describe what I want to do. Please let me know if anything is wrong.

If I look at 'mega steel' voice, it has 5 arpeggios (arp1 - arp5). And its arpeggios have the following elements (copied from this blog post, https://www.yamahasynth.com/apps-2/guitar-arpeggio-study-moxf).
Element 1 - Open; the main body of the guitar and the principal strummed notes of the chord.
Element 2 - Dead-note sound. (This is not used in this phrase).
Element 3 – Mute; a muted tone and adds an articulation to the phrase
Element 4 - Hammer-On; articulations in the Phrase; isolated you hear them quite easily
Element 5 - Pitch slides that occur in the Phrase; again isolated you can hear them easily
Element 6 – Harmonic; articulation (it is not used in this phrase, at all)
Element 7 - FX; the raking noise of the strings being strummed.

Each element has a wave not a voice and its MIDI notes seemed to be recorded in each track of song or pattern according to the blog post.

I'd like to add arp6 to Mega steel. I tried all hundreds of GtMG types but I couldn't find rhythm I want. So I decided to create my own user arp type. According to another blog post, I need to record one bar in song or pattern, then converted to arpeggio. Here I have my question, 'how to record one bar song'. To confirm recordings make right sound, I need to set up right tracks but each track is required to be mapped to each element of mega steel voice if I understand correctly.

OK, then you actually have a different question - "How to create a user arpeggio". Once you get it created, you can just add it as arp6 to the voice the same way arp1-arp5 are attached to it.

Yes, sorry for confusing you.

https://yamahasynth.com/resources/content/95-moxf-1 - top 10 articles on moxf. There's at least 2 about the arpeggiator, hope that helps....

Now I figured out how to record with mega voice. I thought that each element must be recorded in its own track. But elements are split by velocity ranges and all they can be recorded in one track. Thanks.