How that could happen is beyond my understanding. The thing is truly a piece of shit. The only part of it that is of any value is the keyboard module. I've removed it and made a MIDI controller out of it, but the rest of the device has collected dust for years, and despite being almost unused, almost every part of it is now broken. The volume knob produces static when turned, when some of the buttons are pressed it behaves as if other buttons are pressed, the pitch wheel is also defective (despite that I never use it because I think pitch wheels are stupid), the power jack connector has worn out, and even looking past all of the hardware faults, the device itself is rather lame in that, while it can technically produce the sound of multiple instruments at once, it doesn't do so in a way that resembles how MIDI devices usually operate, and even if it did, the sound of most instruments programmed into it is greatly dependent upon the effects processor, of which there is only one, and so when using multiple instruments at once, you can only apply the proper effects of one of them, and so all of the rest sound quite lame.
The thing truly is a complete piece of shit. Do yourself a favor and fall in love with something else.
> I'm curious whether or not you have anytime to discuss potential solutions
There is a wonderful application called ZynAddSubFX. However, every time I've tried it on Windows, it's had significant lag between note-on events and the actual sound being produced. The lag can be reduced to almost nothing by running it in Linux, but then you have to deal with Linux sound issues, resulting in it being unreliable, only working when the phase of the moon is just right.
Unfortunately, while every game in the days of DOS was able to figure out that you can get real-time response out of audio simply by writing into the DMA buffer, I've discussed the issue with some Linux people and they're strangely under the impression that writing to a DMA buffer that is in use might cause the computer to crash or something. So they're unwilling to do that, and so instead the approach taken in Linux is to minimize the size of the buffer, causing the whole experience to be a trade-off between too much lag and too many buffer under-runs.
I expect the only solution I'll ever find is to create my own audio device and write my own audio software to utilize it, but as that is a significant undertaking, I wouldn't recommend you hold your breath while waiting for my solution. Hell knows if I'll ever do it.
Probably you'll just have to wait until the year 2100 when computer science has improved to the point that people have some fucking idea what the hell they're doing. Perhaps then something as simple as real-time synthesis of audio will be possible on a PC.