You may (eventually) be able to download from here: http://soundmondo.martintarenskeen.nl/ and use a standard tool for sending the sysex to your instrument. The last update at this site was Dec 07 and missed the boat of my upload - so you'd have to wait for the next update of this site.

From what I can gather, you've made a demo attempting to show that you know how to use the Motion Sequencer to control something without showing the problems with controlling that something.

Have I got that right?

BTW: the title of this topic and problem statement says one thing and I'm not sure if you've moved to more of a focus on the frequency modulation response itself. Has motion sequence as a mechanism been cleared of any contribution to your issue? Meaning that motion sequence is reset properly by keypress and outputting the right lane values at the right time for what you've programmed and the problem is now with the sound you get when the LFO2 FMD modulated by this stimulus?

Or do you think that MSeq as a mechanism can go off the rails by the modulated parameter (tail wagging the dog issue)?

For complexity - do you have specifics to suggest that I should be trying?

I think both of these things are problems.

But you'll have to learn to look for problems and exaggerate them.

Long ago I suspected there was a problem with rapid, single cycle, single run Motion Sequences being retriggered by keypresses properly. Now I'm certain of it.

And there's a compounding issue with complexity of FM-X sound design and this particular "effect" adjustment. I got lucky, a two-for.

Ok - beyond a rate of 127, each-key reset, the default standard (rising) unipolar pulse on MSeq, and rapid retrigger using multiple fingers and "drumming" techniques on a single key - what else specifically should I do to attempt to reproduce the problem you see with MSeq? Is the problem occurring with more than one Part running a motion sequence? Is there an arpeggio running at the same time? I cannot get MSeq itself to fail even when driving LFO2 FMD on an operator stack with parallel and series modulators.

By the way - I have professional experience with finding issues. A period of my career was as a test engineer when success meant finding the problem. I carry those skills forward with me. To that end, my previous testing with MSeq was that the 127th "tick" of the clock for that single output value was skipped (boundary condition). I haven't tested lately to see if that was fixed or not and it's not related to your issue - but I dig deep and see finding an issue as success. Not attempting to show alternatives or workarounds as success and covering up problems due to avoidance.

In this case - you have the keys to the city already. There's no great reason for me to do anything other than follow a set of instructions you provide starting with some baseline (like a Preset) and tweaking the values to match the case sitting in front you showing the issue you see. Or, at least, communicate some set of Performance parameters you think are key to contributing to the issue and outlining how those should be set.

It is as described above. It is unfixable, by us.

It seems incredibly unlikely that it will be fixed, as it's been like this for quite sometime.

You'll probably need a MODX to experience this, I suspect the superior CPU power of the Montage largely prevents this from happening, or at least makes it less noticeable.

Another compounding problem:

The 2nd LFO, on FM-X parts, at a value of 0 in either extended or normal operation, is not zero.

It's slower in extended mode, but still not zero.

Seemingly, the articulation of the LFO2 FMD value by the Motion Sequencer is also articulating/accelerating or otherwise impacting the pacing and position of the 2nd LFO. The faster the Motion Sequencer is running, the more this is noticeable.