View Single Post
Old 29th May 2020
Lives for gear

The MIDI signal is actually more of a current loop than a voltage. It's designed to drive an LED (inside an optopcoupler or optoisolator--different names for essentially the same thing) with a 220 ohm resistor in series at around 5mA. While many do this from 5V rails (and so when unloaded, such as when measuring with a voltmeter, you'd see 5V), it is possible to do it at higher or lower voltages as well, as the FAQ from MIDI Solutions mentions somewhat obliquely.

The center pin (pin 2) on the MIDI connector is for the cable shield, and is not part of the signal. On properly implemented MIDI inputs, it's not connected to anything; so having it connected atypically on the output port causes no general problems for compliant inputs (but, if it's not connected to a ground, does lessen the effectiveness of the shielding in the MIDI cable). The MIDI data is on pins 4 and 5 (the two next to the center pin--pin numbering on the DIN connectors is unintuitive).

MIDI Solutions seems to me to be playing pretty fast and loose with the MIDI specifications. MIDI was not designed to provide power to devices, save for the one solitary LED in the optoisolator of the input. The MIDI Solutions box is assuming that the devices are using the same ground reference for the shield as for the MIDI data output, something that the specification doesn't specifically require (even with their sample circuit). They are also specifically and explicitly ignoring the requirement that MIDI inputs be optoisolated, since doing so would mean they couldn't steal power from the MIDI input.