today I made some analyses to find out what causes the OPL2LPT to output only distorted screechy sounds on some of my computers / IO-controllers. I think I came to an idea and an option to fix it, but not sure if this is the preferred way, so please let me know what you think.
I found a similar problem on my IBM 5160 and PCjr... I wrote my own VGM replayer for OPL2LPT, and initially tested it on my 286-20 and Celeron 1.5 GHz laptop. The code worked fine on both.
Then I tested on the 5160 and PCjr, and found that they only played garbage.
The way I solved it was by changing the order of the commands sent to the LPT.
Apparently the IBM printer interfaces were sensitive to this somehow. The changed code still worked fine on the other two machines.
I can build some test programs using both variations. I wonder if the issue I had was similar, and changing the commands around would somehow change something in the signals similar to the resistor you added.