Grumpenspiel - Micro PT

A MIDI THRU 1x2 splitter that also provides 5V DC power.

Overview

Micro PT is a simple MIDI THRU box that also provides 5V DC over the unused pins of the standard MIDI DIN connector. This project is a prototype of the larger Power Thru which is the communication hub and power supply for Grumpenspiel.

Don’t be fooled by the label “prototype.” Micro PT is a fully functional MIDI THRU device that splits a single MIDI IN signal into two identical MIDI OUTs. It is MIDI compliant and can be used with any standard controllers or synths, passing all MIDI messages thru (including SysEx and CC).

The communication signal is live any time Micro PT is plugged in. That part does not switch on/off. Each MIDI port has a power switch, used to connect pins 1 and 3 to GND and 5V DC. When switched on, the port still sends the data signal in addition to the power supply. Grumpenspiel devices are built to use this power connection, but standard industrial MIDI devices are not and may not like having this power hookup. The beauty of DPDT switches allows the power feed on each port to be totally disconnected.

Not the most exciting video by any stretch, but represents a happy moment in the life of a robot maker when all the modules in a system talk to eachother nicely.

Inspiration

Micro PT’s circuit is inspired by Morocco Dave, specifically his Midi Thru v2 project. His thru box features more ports than Micro PT, and a very nice looking and compact enclosure. Thank you to him for providing a good description as well as schematic and veroboard layout example.

MIDI in Action

Jumping to the end, an action shot felt better up front here. Details about construction and electronics follow below. This shot shows the glory of Micro PT’s final test. This round of testing proved MIDI IN properly split to both MIDI OUT ports. Power can independently switch on/off for each port, a device that provides its own power is fully compatible. Happy robot builder!

• Upper Right - Micro PT - providing MIDI communication and power.
• Upper Left - MIDI Mud - MIDI controller sending notes into the system.
• Lower Right - Demonic Tapioca - MIDI activity light proving OUT_1 works.
• Lower Left - Simple MIDI LED - prototype circuit with activity light proving OUT_2 works.

I love the timing of this picture, it sent me digging deep into the workings of all four of the modules shown. Looking closely at each device, we see:

• MIDI Mud note indicator is on note 7 (the red LED row at the top).
• MIDI Mud blue instrument shows active for note 7.
• Micro PT MIDI activity indicator is off (the blue LED just to the left of the black MIDI cable, not illuminated and therefore very dark in the photo).
• Tapioca MIDI activity indicator is off.
• Simple MIDI LED activity indicator (red LED at the bottom left, not illuminated and therefore pretty dark in the photo).

When I first saw this, I was alarmed at what seemed to be a high latency of the MIDI signal. This is surprising because the basic MIDI circuitry is built entirely with passive components and very short runs of high quality cables. Did I really capture the exact moment that MIDI Mud arrived at Note 7 and sent the MIDI OUT note, but the communication signal has not yet reached the devices waiting for the incoming data?

Spoiler alert - nope, there’s barely any latency in the system at all. This picture actually captures a moment after the NOTE ON message was sent for Note 7, but before the NOTE OFF signal gets sent as we transition to Note 8. Makes much more sense when seen from that perspective, but it certainly did give me a fun little exploration exercise to find lag that does not exist.

Construction

Micro PT enclosure is made out of poplar wood. Stained with Minwax Jacobean (2 coats), finished with Polycrylic (4 coats), artwork painted with Testors Enamel.

The enclosure was designed in Fusion 360 and carved on the CNC mill. 3D model allows for precise layout and positioning of the components and artwork. This was my first test of Fusion’s V Carve feature for text and lines. Did a test carve to ensure I was generating G-Code for the vcarve properly. Came out fantastic, both the test cut and the final result.

One of these days I’ll get my enclosure side holes right without blowing something up. But until then, I will keep improving my repair skills. The enclosure is made up of three layers of wood, with grain going in perpendicular directions, a bit of a make shift plywood. Top layer and middle layer are glued together, then that assembly connects to the bottom with wood screws. The 5.5mm barrel jack passes through the middle layer, and requires a wall thickness no greater than 3.5mm for the retaining nut to engage the threads. 3.5mm of end grain is not very strong and cracked while drilling. Reinforcement is required, since this location will receive the force of barrel jack insertion and extraction. Clip and sand some craft popsicle stick to perfect size and wood glue to the rescue. Nice and strong.

Dogs

We adopted a new dog in the middle of Micro PT’s construction. She quickly made friends with our first pup. She also quickly discovered that when daddy is soldering it makes for a great opportunity to take a nap.

Electronics

Eagle .sch schematic file.

Each project presents new challenges and new opportunities to learn from mistakes. The first step for my circuit development is to make a solderless breadboard prototype. Quick and easy to prove the concept and ensure the basics of your electronic design are sound. After that is done, build a more robust version of the circuit on a permanent board. Once the perma board is done, test it out to ensure you got it right before you permanently mount it into the enclosure.

When powering up the first permanent board for Micro PT, no MIDI activity light. Sad robot builder.

Test each sub-module of the circuit. Power to each MIDI port. Power indicator LED on each MIDI port. Tested the MIDI OUT function by patching in the signal from the solderless breadboard.

This narrowed it down to an error in the MIDI IN part of the circuit. Reexamined the circuit dozens of times, comparing to schematic, comparing to prototype, testing continuity. Everything is good.

Clipped/drilled/desoldered the optocoupler and replaced it with one I knew to be functional. No change. Built a brand new solderless version to make sure I wasn’t crazy.

Cut your losses and build a brand new permanent circuit from scratch. Power it on. SAME RESULT!!! A true test of patience and remaining calm, cool, collected. Walk away for a bit, come back with a fresh and clear mind. This one had a nice and easy and obvious flaw, a failed solder joint. Heat her up and resolder. Bam, good to go.

Power Requirements

Micro PT supplies power, but it also uses some for itself. MIDI activity indicator LED, up to three power indicator LEDs, plus the passive components (optocoupler, hex inverter, transistor).

Technically, Micro PT does not supply power, it passes power through from its own power supply. 5VDC in, at whatever current rating the wall wart provides (my typical one provides 2A).

This calculation does NOT take into consideration any load drawing from Power Thru, it is the power Micro PT itself uses.

13 mA maximum current draw with all three power connectors on.

Component	Current	Notes
Power Indicators	3 mA each, 9 mA max	direct measurement
MIDI activity indicator	3 mA	estimate. this depends upon incoming MIDI messages, made an estimate.
Passives	0.5 mA	datasheets