'MIDI power stealing'
I'm working on a MIDI project.
For those of you not familiar with MIDI, it stands for Musical
Instrument Device Interface, and it's a simple serial protocol at
31.25 Kbaud for communicating between musical instruments.
My project is a simple one: part of the information in a MIDI "key
on" message is the velocity with which the key is pressed. A friend
has a keyboard he really likes, but which can't sense velocity, so
it sends out a fixed value for the velocity. He wants this replaced
with a larger number, so that the synthesizer receiving the data
will think that the keys have really been hammered. Loud Is Good.
The project is done, but I'd like to make One More Improvment:
eliminating the need for an external power source.
I know this is possible, the friend I'm doing this project for has
a MIDI filters that doesn't have batteries. Looking at the MIDI
spec, I should be able to extract +5v and GND from pins 5 and 2
(although I'm violating the spec and risking introduction of a
ground loop by doing so). There should be about 22mA available.
I'm running a 16C63 at 2MHz (should only need <5mA), a 74ALS04
(should only need about 5mA max), and the MIDI-required optoisolator
(don't have a spec sheet handy, but it's under 5mA). I have one
blinkenlight using about 1mA.
It doesn't work. As far as I can tell, the PIC continually resets
itself. And this is without anything attached to the MIDI-OUT and
MIDI-THRU sockets, which would result in even more power drain.
I'm just on the verge of making this thing use batteries instead.
If anyone has solved this problem (or knows that it's insoluble),
I'd be grateful for any suggestions you might have.
I have a Portman PC midiman , which is an external MP401 midi controller
that connects to a PC LPT port. It has a selector switch to use it with or
without an additional power supply. It works fine for me with out a power
supply. I opened it up and it appears to be drawing power from the LPT port.
This is not much help, however along with about 6 IC's there are also 2
At 08:36 AM 7/17/97 -0700, you wrote:
>I'm working on a MIDI project.
>For those of you not familiar with MIDI, it stands for Musical
>Instrument Device Interface, and it's a simple serial protocol at
>31.25 Kbaud for communicating between musical instruments.
>I'm just on the verge of making this thing use batteries instead.
>If anyone has solved this problem (or knows that it's insoluble),
>I'd be grateful for any suggestions you might have.
It may be possiable using a diode and capacitor as a charge pump (Its common
on rs232) this would only violate the recption end and you could do away with
the recpition opto coupler and replace with a general pourpose transistor.
This would reduce the ammount of consumed current taken by the reciver led.
The capacitor would be required to hold sufficent charge to enable opperation
of the downstream LED
one minor snag if midi transmission device complies properly then a 220R
resistor is in series with both pin 2 and pin 5 before the oto connector this
limits the current available to a short to 5v / 440R (11mA) the same thing
should be presant on the reciving diode further reducing the LED current to
5.5 mA assuming the LED has Vf of 0v in practice its less than this.
I am not good at drawing circuits in ASCII but here goes (untested but the
idea should work)
midi in hi_____R_____D_________________________C+ power +ve
l_____b l_________________Pic input
midi in lo_l____________________________________C- power -ve
The diode should be a fast one or shocktty due to data speed (11DQ04) or
equiv the NPN could be 2N2222 and the base resistor and emmitter resistors
about 10K a 100R powersupply feed resistor could be used but this may need
adjusting to suit. The value of capacitor would need careful selection to
avoid overloading the output but if it is too small there would be too small
a charge to power the output LED for the duration of the transmitted signal.
Some untested food for thaught Steve......
> From: Brian Scearce
> I know this is possible, the friend I'm doing this project for has
> a MIDI filters that doesn't have batteries. Looking at the MIDI
> spec, I should be able to extract +5v and GND from pins 5 and 2
> (although I'm violating the spec and risking introduction of a
> ground loop by doing so). There should be about 22mA available.
> It doesn't work. As far as I can tell, the PIC continually resets
> itself. And this is without anything attached to the MIDI-OUT and
> MIDI-THRU sockets, which would result in even more power drain.
Pin 2 (assuming a "standard" 9 pin) is Receive Data and Pin 5 is
Signal Ground. Wouldn't trying to swipe power from a data line
either bugger the data or give you an intermittent power source.
Since what you're describing sounds like an intermittent power
source, I wondering if you could swipe the power from pin 6
(Data Set Ready). On some systems DSR is simply pulled
high and left there.
Image Logic Corporation
Byron A Jeff
While MIDI is a serial protocol. It's completely unlike standard RS-232.
To prevent ground loops each connection is optoisolated. Typically this
is done by passing a current limited +5 and wiggling the other line
between Ground and Open.
Brian, it could be just a simple as the fact that GND isn't passed though
pin 2. The MIDI projects I've done didn't have ground passed on the shield.
Since you can't gurantee it you may want to look elsewhere for power.
More... (looser matching)
- Last day of these posts
- In 1997
, 1998 only
- New search...