I have just had exactly this problem. I've been advised that the symptom usually points to a short circuit in the filter chip that sits on the logic board just behind the internal floppy port. It is clearly marked FILTER. If shorted, this results in the pin that selects the external port remaining permanently high, so bypassing the internal disk.
I've yet to do this, but the advice is that if you remove that filter chip, the internal drive will function but the external one won't. On my Plus logic board, the filter is a Bourns 4120R RC network, 20-pin DIP. You can replace it, or build yourself one on a 20-pin DIP socket following the schematic available on the Bourns website, using sixteen 25R resistors and eight 200pf capacitors. I think the replacement is available from serious electronic stockists.
FWIW I think I shorted this filter by (experimentally) attaching a 3.5" Superdrive to the internal connector. Pin 9 on that connector is -12v for legacy drive pullup purposes, and should not be connected to a Superdrive, which will ground it. I suspect this drew a serious overcurrent through the connector. The logic board generally was protected by the analog board "fluffing" overcurrent circuit-breaker, but the filter is likely to have failed.
I'll get to removing the filter in the next few days to see if this advice is correct, then check replacement options.
EDIT: Removing the filter has not solved the problem for me - more to study about where signals are going between the drive and the IWM chip.
Rick