previous by date | index | next by date |
previous in topic | topic list | next in topic |
> From: eric@...Interesting. The shell is connected to the chassis, but nothing in the cable is supposed to be connected to the shell. The shield is supposed to be connected to pin 2 (the middle pin) at each end, which is connected to digital ground in the transmitter but left disconnected in the receiver. If you measure 0 ohms between the shells of both plugs, then it's the cable that's violating the spec, and tying the two chassis grounds together, which defeats the purpose of the optoisolator.
>
> Well I've figured out the "clicking" thing... and the Polaris
> isn't broken!
>
> AFAIK the MIDI spec calls for the input on a device to be
> nothing but an optoisolator conected to pins 4 and 5 of the
> DIN socket... no other connections. But some newer/cheaper
> MIDI interfaces, including the one I was using, "cheat" and
> steal power from the MIDI input by grabbing +5 volts off pin
> 4 and using the shell as negative.
>
> Well in the Polaris, the MIDI output connector shell is
> connected to the chassis, and only to the analog/digital
> grounds through a 1-megohm resistor. So when current flows
> through the shell, it energizes the chassis relative to the
> digital + analog grounds. The current flow varies depending
> on whether data is or is not being sent out the MIDI port and
> I think that was causing the clicking I heard. Just for
> reference, with nothing plugged into the MIDI port I measured
> a few hundredths of a volt between chassis and analog ground
> (at the PS board), with the interface plugged into the port I
> measured almost a volt!
>
> It seems like what needs to happen is (other than
> manufacturers making MIDI interfaces that follow the spec,
> and good luck with that) to connect the MIDI connector shell
> to digital ground. But doing that with the Polaris in the
> factory configuration would result in digital ground being
> connected to the chassis which I think would be bad for a
> number of reasons. The only solution to ∗that∗ problem I can
> think of is to replace the MIDI output connector with an
> isolated one and I'm too lazy to do that. ;^)