Articulation problems in Iconica

I use Cubase Pro 11 with a 128GB RAM 32 core PC. I’m experiencing articulation problems in Iconica. I’ll describe how I experience them at this moment, which is in the string section, but I’m not sure if similar problems couldn’t occur in other sections. I wonder if I’m doing something wrong or overlooking something. Please help, when you can.
Here are the problems. I often combine two articulations, typically Marcato and Legato. The purpose is to give each note a firm start (which can be dosed by means of the velocity) and still add the possibility to make it a long note (where I dose the playing volume with the modulations curves). I find that combining those two variable articulations on each note (Marcato and Sustain - or Legato, Tremolo, Sustain Vibrato…) offer tremendous expression possibilities. But what happens is that, when having multiple string instruments in my section (at present, when the problems occur, there are 6 tracs in the strings section playing simultaniously: 2xViolin1, 1 Violin2, 1 Viola, and 2 Celli), sometimes only the Marcato is performed while the second articulation (Sustain) is skipped, which makes the note stop much too soon. This doesn’t occur with all instruments, just one or a few, but it seems to vary at random. Today it occurred with the Violins1 at one time and with the Celli at another time - all within the same phrase of two bars that I was looping for detailed work. Typically, when trying to find out what was going on, I muted some of the instruments, to isolate the “problem”-instrument. Doing that, the problem disappeared, which made it as far as now impossible for me to diagnose. Anybody any experience? Any advice?
Thanks on forehand!

One possible reason is that the Max Voices value is not enough.
https://steinberg.help/halion_sonic_se/v3/en/halion/topics/_shared/options_page_r.html

Thank you so much, Accel. Thinking of what you said, and planning to carry it out, I wanted to clean up the string tracks first. (This was something I had to do anyway, but had postponed until now.) I guess that, by doing this, I reduced the voices already sufficiently, because the problems are over. Imagining how it is described in the manual, on the place that you indicated with your link - how voices are being “stolen” when there are more voices used than provided - I realized that this is the reason why the effect appeared in a (seemingly) random way. So again, thanks a lot!