[sdiy] MIDI phantom power...over 5 pin MIDI connector ?

[rsdio at audiobanshee.com]

On Sep 10, 2015, at 1:29 AM, Ove Ridé <nitro2k01 at gmail.com> wrote:
> On 10 September 2015 at 09:40,  <eidorian at aladan.net> wrote:
> > I just re-read the original post, and it might be that JP isn't asking how
> > he can get power from a MIDI cable, he might actually be asking a subtly
> > different question: how can he *send* more power over the two MIDI pins
> > (presumably without affecting the ability of those pins to transmit MIDI to
> > older devices at the same time).
> 
> The MIDI-DIN states that pin 2 should be connected, if nothing else, through the shield. Are there really cables around that have pin 2 completely unconnected?

It's not the MIDI cables that have pin 2 completely unconnected, but the MIDI Out circuit that uses a separate ground for pin 2 than for the signals.

If you look at the official MIDI circuit, you will see two different ground symbols. The hollow triangle is used for the opto-isolator ground (and, presumably, also for the inverters/buffers). The pseudo Earth symbol with three parallel lines of different lengths is used for both outputs (MIDI Out & MIDI Through). There is no guarantee that these are connected together with a low impedance such that full power supply current could pass between them.

In fact, for best noise protection in analog synths, the MIDI cable shield should be filtered before connecting to ground, so that the MIDI cables don't act like an antenna to pick up noise and inject it into ground. A good filter would be a 1 Mohm resistor between the two grounds, with a 1 uF capacitor in parallel.

Considering that keeping noise out of a synth is far more important that providing power to an external device that is bending the spec, I'd actually prefer the designs that don't connect pin 2 directly to audio ground.


> > But yeah, there's no fun to be had doing this.  Any device that requires
> > your "high powered" MIDI-out ports for power is going to damage your
> > "normal" gear if you ever plug it in accidentally :-(
> 
> I don't see how. A compliant, regular output device should have 220 ohm resistors in series on both pin 4 and 5 to limit the current and should handle even being indefinitely shorted. Likewise, a regular input device will have 220 ohm in series with the optocoupler. 5/220=23 mA and that's being really pessimistic and ignoring the LED forward voltage drop. 

In case of a short, the 220 ohm resistor would only limit the current to 22.7 mA (23 mA), but that doesn't mean the internal 5 V supply was designed to handle more than 5 mA. There are many examples of analog synths with wimpy 5 V regulators that cannot handle the full synth demands. Another 15 mA might not really cause damage, but it can make a bad situation worse. I'm thinking of the Prophet 5, which was notorious for having the 5 V regulator burn out easily. That was even before MIDI, I think. I'm talking about damage to the sending synth.

I think "eidorian" was saying that a turbo-charged MIDI output would damage the MIDI input of a standard device.

Brian

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20150911/50d37284/attachment.htm>