[I am writing this when I am very tired. It will be terse.]

Now that the upper manual is basically controllable by MIDI, it’s time to look at the lower. The lower manual is more complicated than the upper. It has two sets of contacts for each key; one of which fires off the bass and chords, and one of which is ‘accompaniment’.

Essentially, then, the organ needs two sets of opto-isolators, one for each set of contacts. This means that each pin on the MIDI board needs to switch two LEDs; and for sanity’s sake, it’s probably a good idea to stick a tolerably heavy-duty diode (I have obtained industrial quanities of 1N4003s) along with the LED to make sure that current is only going where we think it is going.

First: looking at the ‘bass and chords’ contacts. These are in a scanmatrix arrangement; and so need a diode next to each key to avoid squirting current back up the scanmatrix pins on the mystery chip which seems to be running all this. (The diodes can be seen on the wiring for the lower keying; pps 8-9 in Part 2 of the TG-44 service manual, which I linked to previously.)

This means that, instead of our previous circuit, for each key, we have:


This is the basic outline for relayboard version 2b (2a is identical to 2 except that the two rows of chips can have different Vccs). This /should/ be interchangeable with v2 (so long as I’ve understood the organ correctly). In terms of board layout, it makes things a little denser; although the diodes on the MIDI side can sit right next to the resistors, the diodes on the other side will necessarily stretch things out a little. (I will post photos when the diodes arrive and I’ve had time to fiddle).

Each scanmatrix pin on the magical chip of doom deals with one octave, or twelve notes. It’s easiest then to match each octave to one row on a relayboard.

Plan for the ‘accompaniment’ keyswitches to come later.


Pruning the electronic forest

The version 2 relayboard is on the right; it uses the solid-state relays mentioned in my last post to control keys. It works. I’m making another one now; after that I will have enough to MIDI-ify the whole upper manual.

(This is a nerdy update at least as much for my own memory as for anything else.)

To try to get around some of the issues I’ve raised in the last post, I’m going to try a new ‘relayboard’ using Opto-Isolators.  An opto-isolator is effectively an LED and a light sensor (a phototransistor) in a sealed, dark box.

It should be possible to wire one of these up like this (ish):

I’m looking at a Toshiba quad-opto-islator chip that will switch up to 55v at up to 150ish mA (datasheet) – and given that the existing diodes in this beast seem to be rated at 100mA, it probably won’t go toasty.  The voltage on the organ side is somewhere less than 23 volts, according to the service manual (schematic 5; page 18 of part 2; the left hand side – point BB is the pink common wire to all the keys on the upper keyboard)

So it should be fine to switch it through the transistor side of the opto-isolator.

Still need, though, to find the value of the resistor that protects the LED: according to the datasheet, the LED will eat 1.15 volts at up to 50 mA (0.05A).  The power supply is at 12v.  For safety’s sake (and to avoid warmth), we’ll stick to pushing 20mA through the LED, giving a resistor value of:

So a 620 ohm resistor should do the job.  Next step: prototyping this…

“no, no, make it stop, please at least use cable ties…”

There are now 20 notes active on the relayboard.  This is all the notes I can get working without going back and fixing some dud soldering.

What’s good about the relayboard design?

  • It works, and it works reliably. It is odd playing the keyboard and having the organ go off.
  • It is electrically very simple, and within even my questionable powers to assemble.
  • It is unlikely to explode without warning

What’s not so good?

  • It sounds like a room full of people knitting. Sound will follow.
  • It gets unseasonably warm. Not too warm, but more so than I am entirely happy with for something that performs such a simple job.
  • Corollary: it is probably drawing more current than I’d like.

I’m still alive, and so is the project!

Relayboard no. 1 is finally complete; this board, to remind any imaginary readers I may have, will electrically replace about half the keys on the top keyboard.

I wonder which species of bird is nesting on my sofa

For the non-electricians: a relay consists of an electromagnet and a little switch in a box, so that when the electromagnet is turned on, then so is the switch; and when the electromagnet is turned off, the switch is turned off.  This is nice here, because I don’t trust the electronics inside the organ to be predictable – so I can control the “keys” (the “switch” side of the relay) via another circuit without actually passing any electricity between them.  So far, so good.

I have now tested the control side of the board, and it more or less works; next up, testing the switch side of the board.

… the badgermin. (via arbroath).

Video of this horrific … thing:

(A project note: I’m waiting for my first MIDI decoder to arrive, and so there’s not much to report on the organ front at the moment.)

… and the barman says “I hope you took the bees out of that first”.

Which is pretty much what happened, with the exception that the barman needs to be replaced with the taxi driver.  I’m not sure I understand…

Sorry for the phonepic, but I wasn’t carrying around a camera. For obvious reasons.

These keyboards are the pleasantly repunit-y age of 111 years old – built in 1901.  They are made of solid oak – a fact which, having carried them all the way back from Lincolnshire on the train, I am eminently qualified to appreciate.  They came from an organ which had a pneumatic action – which means that they will need a bit of adaptation.

I will need to brace the backs of the manuals, attach springs to each key, and adjust to taste.  Which all sounds much easier than it is going to be, given that my woodworking skills are at best mediocre and at worst actively dangerous to myself and everyone around me.  So, this sounds like it might be fun…

As regards actually getting signals out of the keys to the MIDI encoder, I’ve really got two options – either attach magnets to each key and fit reed switches on a bar above the keyboard, or use the existing copper contacts on each key.  The former sounds easier; less wiring to get entangled around the backs of the keys sounds promising.  But this will require research!

There is also the slight issue that they are too deep for the chassis, but I will burn that bridge when I come to it.