[sdiy] MIDI notes filter which is also a MIDI converter for Junos ?

[rsdio at sounds.wa.com]

I hope that continuing this discussion is of interest to the list. As  
a synth DIY group, I imagine that these are important considerations.


I will concede the point about the opto-isolator. In fact, I had  
already considered that my point might be moot.

If you had any audio inputs or outputs, then I would disagree,  
because you'd need to isolate the ground. However, since your MIDI  
processor handles no audio at all, then it's safe to skip the isolation.

In addition, the only case I can think of where lack of isolation  
would cause a ground loop is where the receiving synth also skipped  
the opto-isolator, in which case the situation would be just as bad  
without your device in the MIDI chain.


Since you agree about the self-powered issue, I may not need to point  
out the following: It doesn't seem possible for your circuit to  
output the full 5mA while only receiving 5mA on input, unless your  
CPU is micropower (MSP-430) or you thin out the MIDI stream so that  
fewer bits are transmitted than received. There will be some loss as  
the CPU uses up current. I guess the diode drop from incoming 5V to  
outgoing voltage would allow the current to be maintained, but just  
barely.

Technically, your circuit is probably pulling more than 5mA from the  
sending device, because there are two paths from pin 4 - one through  
the normal path to pin 5 for the current loop, and an unexpected  
(according to the specification) path to pin 2. This increase in  
current draw out of pin 4 from the sending device could bring its  
MIDI output circuit out of spec. Technically, pin 2 is not supposed  
to be connected on the receiving device at all, so the sending device  
expects no current to flow.

I recommend that builders incorporate external power with the  
optional LM7805 from your schematic, or some other regulator. One  
change I would suggest for your schematic is to point out that pin 2  
should be disconnected when running from external power, or at least  
some mechanism should defeat any power draw from MIDI (maybe D1 is  
sufficient, though, if you can be certain that everything is running  
on 5V).


As for commercial products which do not ground pin 2 of their MIDI  
output, I'd say they're in violation of the MIDI specification - but  
with one key caveat. My suspicion is that such devices have a filter  
network between pin 2 and the actual ground. A typical filter might  
be a 1M½ resistor in parallel with a 0.1µF capacitor. This would  
provide poor ground for power, but excellent ground for shielding  
purposes. Such a filter would prevent high frequency digital noise  
from being broadcast over the cable shield, and that's a good design.  
In an audio studio, you don't want cables broadcasting digital hash  
all over the place, at least not until such time as everyone embraces  
balanced interconnects for all audio (analog synths and such).

Brian Willoughby
Sound Consulting


On Dec 28, 2013, at 18:34, Dave Brown wrote:
> I appreciate your feedback and the time to answer.  It certainly  
> "pushes"
> the MIDI spec but I disagree on the opto-isolator point.  This is a
> "floating" processor directly attached to the transmitting device.   
> It will
> only operate if the transmitting device supports pin 2 of the  
> connector
> which is ground.  I think I have seen at least one commercial  
> product which
> did not have this connection.  I think of it as an "extension" of the
> transmitting device ... just a bit more circuitry on the output.   
> There are
> no other connections to any other ground since it is self-powered.   
> The
> receiving device will still have an opto-isolator so there ground  
> isolation
> is maintained between the two instruments.
>
> The MIDI spec does not sanction providing power; neither does it  
> prohibit
> it.  It does say that the transmitting device must provide up to 5  
> mA of
> current.  On the receiving side it says the device can consume up  
> to 5 mA.
> Since I am not driving a an opto-isolator 5 mA is plenty of current  
> to power
> the device.  The voltage level will fluctuate, hence the need for a  
> crystal
> to maintain the baud rate within specifications.  I don't remember  
> what the
> typical operating voltage is, but it is lower than 5 volts which is  
> why the
> output resistors are reduced from the typical 220R values.  This is  
> not a
> product nor was it intended to be so, so I'm not worried about the  
> MIDI spec
> in itself but whether it actually operates.  The value is in the  
> software so
> anyone concerned could augment the hardware with an external power  
> supply
> and an opto-coupler if they wish.  That is the beauty of DIY.
>
> Your self-powered observation is a good one.  MIDI Solutions makes  
> a lot of
> stand-alone gadgets that are powered by the MIDI.  I've tried this  
> on a
> number of different units and they all have worked (except for that  
> one that
> did not have the ground pin 2).  I've done some specific code for  
> it to
> implement a keyboard split for some people in the UK who have some
> multi-keyboard setups and I did caution them that there could be  
> issues but
> it worked perfectly.
>
> Thanks again for the comments.  I should have mentioned the self- 
> powered
> caveat in my original post.  This is a great forum with an exceptional
> knowledge base.  Take care.
>
> Dave