Video on my progress

January 28, 2010

All keyboards and pistons work now.  I’m working on a post describing my trials at sysex uploading, but that is for another time.

Next step: connecting the stop tabs and the swell pedal.


Connecting the key wires

January 18, 2010

Last Saturday I completed hooking up the upper manual key wires and the pedal key wires to the midibox.

Connecting wires from keyboard to ribbon cables attached to MIDIbox

I used crimp connectors after tinning the stranded wires of the ribbon cable for the upper manual wires.  As I said earlier, they are color coded by note, so it was not too difficult to keep track of each note, especially because I preserved the linear sequence of the snipped wires.

Even so, I missed a wire in mid-range.  I didn’t notice it until I was finished and had an extra input unconnected.  I just connected the missed wire to the last input and decided to edit the midi control system file so that it would play the right midi code anyway.

The pedal notes were easier.  I cut up a circuit board (one of the percussion matrices) that had plenty of prong-type connectors on it.  Ribbon cable wires were direct-soldered to individual prongs, and the pedal wires, with their press-fit connectors still on, slipped onto each prong neatly.

Although it all is very straightforward, it actually took the better part of a day to get this much done.

Next, the test:  I plugged the USB-MIDI interface into the computer, fired up the Miditzer program, checked to make sure of the midi source, and applied power to the midibox.  Then I played some notes.

They worked, except for a low G natural.  I checked continuity and isolated the problem to an out of place whisker wire.  Once slipped back in place, the keyboard played perfectly. 

Same with the pedalboard.  One note was missing–it turned out to be a whisker wire too. 

Now I have a playable organ, but the lower keyboard still needs to be hooked up.

I’ve stopped at this point because I ran out of IDC pin header connectors. These press-to-fit connectors make quick work of connecting the ribbon wire to the small junctions on the midibox.  I didn’t order enough, so I’m waiting for the next shipment. 

The alternative is to hand solder each of the 61 wires to their respective places on the printed circuit board, plus hard-wire the various leads from pistons and stops.  It is not very enticing.  Besides, being able to easily connect and disconnect the midibox makes troubleshooting much easier.

Building the MIDIbox

January 9, 2010

Two DIN boards chained to a Core board on the right, MIDI cables go to computer

The brains of the system are a combination of hardware and software that converts the key signals into MIDI signals that can be rendered by a computer to produce sampled organ sounds. The approach I decided upon was the MIDIbox.

MIDIbox is a do-it-yourself system developed by Thorsten Klose and supported by a hobbyist community.

There are commercial MIDI controllers available and they are apparently easy to use. They do cost more, but mostly I wanted to build these things to learn about them. There are no easy instructions, but careful study of the schematics and documents on the forums linked above gave me enough information to build, program, and configure the system.

Basically, you have a Core system for every 4 Digital Input modules (DIN). Each DIN can receive up to 32 digital signals (in this case, on-off signals from the keys and other signals from stops and pistons). This means that one Core can handle two 61 note keyboards with a few inputs left over.

But because I also have 32 notes from the pedalboard, and some stops I’d like to use, I needed at least two more DINs. This required an additional Core module, which can be connected to the first Core to merge the MIDI signals from the pedals and stops.

I’ve built one Core and four DINs so far. The Core needed an operating system installed on it (called MIOS, developed by Thorsten Klose), and that took some figuring out. Then I needed to install the application that turns switch functions into MIDI signals (called MIDIO128–downloadable from the MIDIbox website). This required learning about PERL and changing an edited *.ini file into something that can become an executable file. I actually managed to get all this done, after a few fits and starts, by following information I found on the web. Later I will try to post links to the most helpful documents.

After loading the operating system and application, and after a bit of tweeking, I was pleased to hear organ sounds by shorting to ground the various keyboard inputs on the DINs. This worked with all three of the organ emulators I have loaded: Miditzer, jOrgan, MyOrgan (more on these later).

Wiring the keyboards

January 9, 2010

I finished preparing all of the keyboards for the MIDI conversion. From a technical standpoint this turned out to be a very trivial matter. Each keyboard involves at least one set of direct contacts made by a silver wire on a key jack contacting a silver bar. Usually either a resistor or a diode separates the point of contact from the terminal from which the key wires proceed. It is a simple matter to attach a jump wire to short out the resistor or the diode. Still, this required a fair amount of work. There are 122 keys and 32 pedals. This allowed me to use the existing color coded key wires to make connections to the digital input (DIN) modules of the MIDIbox system.

Keyboard jumpwires bypassing resistors

Pedalboard jump wires bypassing diodes

A few observations so far:

First, the silver contact bars and the silver wires do not look silver in color. They are black. I am pretty sure this is just a matter of silver tarnish, which is silver sulfide. Silver sulfide is reasonably conductive. After measuring the resistance on the contacts, I decided there was no reason to clean the contacts because the resistance was on the order of 10 ohms or less. Because the key contact signals are routed through a 10,000 ohm resistor on the DIN board, I decided that an additional small amount of resistance is of no consequence. I hope that is the case.

I have read on forums that some people encountered, on the older Baldwins, a black goo-like substance on the pedal board contacts. I did not see this on the Cinema II, but I suspect earlier versions (pre-1972) may have used elastomer contacts. I get this from the service manual, which states that manual keyboards use elastomer contacts for playing stops. These elastomer contacts apparently allow for a varying resistance as the pressure increases on the key. However, for the percussion contacts, silver direct contacts are used.

The silver contacts seemed to be the best approach for signaling MIDI inputs.

Initially I thought I would just remove the resistor or diode that was found between the contact and the wire terminal. I quickly found this to be a mistake. The reason is that the diode or resistor and the whisker wire (the silver wire pushed up by a key jack to make contact with the silver bar) are held in place on the circuit board simply by a drop of solder. If you try to solder a jump wire directly into the hole holding the whisker wire, the whisker wire falls out. To fix this requires unscrewing the circuit board that holds the whisker wire, flipping it over, and soldering the whisker wire back into place while also soldering the jump wire. This also requires that you align the whisker wire properly so that the jacks will push it against the contact bar when the key is depressed.

I made this mistake on the first five pedal contacts. After reinstalling whisker wires, I decided instead to leave the components in place and simply solder the jump wire to short the resistor or diode that was there. This kept the whisker wire solder joint from melting. This minor adjustment saved hours of frustrating reconstruction.

Underside of pedal contact board where I had to reinsert whisker wires

Another observation: careful tracing of the key wires prior to snipping allows you to maintain an orderly sequence of wires. On the Baldwin Cinema II, the key wires are color coded by octave. By careful location of the snipping points, you can even keep the octaves themselves in sequence.

For instance, the upper manual percussion signal wires are connected, note for note, directly to the lower manual percussion contact buss. Once I understood that, I could snip each wire at each lower manual contact terminal and end up with a nice linear pattern of color coded wiring coming from the upper manual. It was an easy matter to continuity test each key contact by moving down the line.

For the lower manual, however, I snipped the entire cable inside the organ case. I did this because it seemed too difficult to remove all the filter boards they connected to. I later realized that this was a mistake, and I will now have to segregate each octave before connecting to the DIN module. Still, the color coding helps. I only have to identify each octave rather than each note.

The pedal wires were the easiest of all. They routed through the bottom of the console and attached to a PC board by simple push-to-fit connectors. It was nothing to pull each wire off its terminal with needle nose pliers. They ended up in a nice linear sequence. I was able to continuity test all the pedal wires in less than 10 minutes.

After continuity testing of the two keyboards and the pedalboard, I bolted them back into the console, routing the wires into the back space for connection to the MIDIbox system that I am still working on.

Lower Keyboard contacts and pedal switches

December 28, 2009

Lower Keyboard Switches

I lifted the upper manual to take a look at the lower manual.  One interesting thing is that the wires from the top row of the upper manual contacts connect to the top row of the lower manual, note for note.  When I looked at the schematic, I realized that because this was what triggered the rhythm circuits, the manuals were permanently coupled for that purpose.  No real problem–I plan to cut the wires leading from the upper manual at this point and route the wires to the inner case for attachment to the DIN units of the MIDI controller (more on this later).

The lower keyboard was slightly different, but still had an easy way to bypass the resister and diode circuits so that direct on/off switching was possible.  Again, the upper contact was a whisker wire on a silver bar.

Lower Keyboard Contact Circuitry

The alligator clips are connected to where there is a non-resistant on/off switching circuit.  This means that a simple jump wire bypassing the diode (located just above the red alligator clip) will let the switch work through the existing wiring.

Here is a shot of the lower keyboard wires coming into the case awaiting snipping and connection to the DINs.

Lower Keyboard Wiring Inside Case (Marked by blue tape)

Pedal Switches

I was nervous about the pedal switches, having heard horror stories that Baldwin pedal contacts involved some sort of black goo.  The schematic showed a circuit that included a resistor and biased diodes.  The theory was intriguing:  the pedal 8′ and 16′ signals were always on but prevented by a diode from proceeding to the various filters and amplifiers.  When the pedal was depressed, the switch applied an 11 volt load to the diode, essentially opening it up to allow the signal through.  Very ingenious and somewhat Rube Goldbergish (look him up if you haven’t heard of him!).  I think it was a way to allow the signal to come on a bit gradually instead of abruptly.

In any event, I don’t need such complexity.  I was hoping that I could find an easy way to bypass the existing circuitry to get a simple on/off switch for each pedal note. 

Here is the switching assembly with the cover on:

What I labelled “pedal switches” above are actually plastic jacks.  The are thin, somewhat brittle, and therefore fragile.  I took care when working on this assembly to not bump them against any other parts.

For reference, the jacks are pressed down by felted tabs on the end of the pedals:

Pedal Felt Tabs

Cover taken off:

After the cover came off, I got my first glimpse of how the switches work.  There is a whisker wire passing through each jack.  When the jack is depressed, the whisker wire contacts a rail. 

Contact Points

The alligator clips locate the non-resistant switch points.  It looks to be a simple matter of jumping past the resistor below the black clip so that the existing pedal wiring can be used.  The red clip is attached to an uninsulated wire that grounds the entire pedal switch assembly.

So, this initial reconnaisance tells me that, with a little work on each contact assembly, I can use the existing wires to send switching signals to the MIDI controller.

Tracking the upper key switch wires

December 25, 2009

The Baldwin Cinema II has numerous wire contacts per key.  When a key is pressed, it lifts a jack (much like a harpsichord jack), bending various wires and forcing them to touch contacts.

According to the service manual, the stop sounds are controlled by small gold wires being pushed against elastomer contacts, while the rhythm sounds are controlled by a wire touching a silver rail.  The reason for the difference is apparently that the stop sounds, controlled by shorting a field-effect transistor, are susceptible to a “pop” unless the contact is made with gradually decreasing resistance.  The rhythm sounds are not triggered in the same way and can rely upon a basic on/off contact.

The elastomer technology is intriguing, and I imagine that there is potential for using the arrangement to control velocity MIDI inputs much like piano controls operate.  This would be useful if I were building a piano, but I’m building an organ.  The simple on/off signalling is sufficient, and, as it turns out, much easier to connect.

I lifted up the lid and took a look at the upper keyboard:

Lid up

A closer view of the top of the upper keys on the treble end:

And here is an out of focus photo showing the contact wires, jacks, the silver rail and the elastomers:

After poking around with test probes, I found what I was looking for: a simple on/off contact when a key is depressed.  Most of the wires coming from the contact assembly had bias resistance–something I didn’t want to go to the MIDI controller I would be building.  But, providentially, the upper contact before the limiting diode (shown below) showed a direct connection when a key was pressed. 


The black lead above is connected to an uninsulated jump wire that connects to ground.  The entire metal key frame is grounded.  The red lead is connected to a diode lead.  If I connect to the other side of the diode, resistance is 3K ohms.

The wires from the diodes are color coded and arranged by octave.  They control the existing circuitry that allows sounds from the tone generators.  I would like to keep the wires because they are so orderly.  Each color represents a note.  For instance, C is black, and it is a fairly easy thing to keep track of each key’s wire.

The problem is that the diode presents resistance and a one-way current limit that I don’t want.  I’ve seen other organ MIDIfiers simply tear out all the wires and start over with new wiring, but I couldn’t bring myself to waste the beautiful color-coded arrangement. So my proposed solution is simply to solder a jump wire past each diode and use the existing key wires to connect to the MIDI hardware.

In the next post I’ll put up photos of the lower keyboard arrangement.


December 25, 2009

The organ is dying.  I purchased the used but well maintained Baldwin Cinema II organ, model #214DR, about a decade ago because I wanted something with a full, 32-note, pedalboard.  I saw it advertised by a widow in Seattle for a pretty low price (for that era).  Her husband had played it, but, after he had died, it was just collecting dust in the living room.

It was in good shape and sounded very nice.  I bought it and proceeded to load it into my truck single-handed.  I did have a dolly and a bunch of straps, but the thing was heavy.  It took me about an hour and a half to get it into the back of my truck.

Baldwin console

The organ as it is now

It gave me years of pleasure.  I’m not a great organist, but for a time years ago I was a performance major in organ and I learned more or less how to play.  I happily played Buxtehude, Bach, Franck, hymns, and my own works in the living room.  It wasn’t a real pipe organ, but it sounded fairly nice.

Until the tibia started cracking.  The tibia stops were the prettiest sounds on the instrument.  Without them it was just a giant oscillator.  I had a technician repair the problem and was again happy.

Until it happened again.  And yet again.  As time went on, other stops started failing.  Hissing and cracking emanated from the organ’s four speakers.  I tracked down worn out transistors, dried out electrolytic capacitors, and various other decaying electronics, but replacing these things was only a temporary fix.  The electronics, some 38 years old, are undergoing solid-state decay.  I am convinced that it is irreversible and replacing the circuit boards is not worth the money.

So I searched the web and learned that others are MIDIfying their old organs.  Essentially, this means to take an old analog console and turn it into a MIDI capable instrument for controlling a computer that has sampled pipe organ sounds on it.  The improvement is amazing, and it doesn’t necessarily cost a lot if you are willing to do the work. 

I hope to chronicle the journey here.  I have already purchased the parts I need.  Today I mapped out the keyboard wiring and am convinced that I can keep most of it.  The rest of the electronics will be ignored or removed.  I’m not sure yet.

Here are the keyboards and tabs.

The organ has 5 pedal stops, 8 Accompaniment stops, and 15 Solo stops.  Because it is a theatre organ, the top manual is called the Solo manual instead of the Swell, and the lower manual is the Accompaniment manual instead of the Great.  I cut my teeth on classical organs and probably will be calling them Swell and Great.

In addition to the stop tabs, there are some 24 other tabs for various effects, couplers, special sounds (harp, chime, etc.).  There are also a bunch of buttons for rhythm functions and odd-ball sounds like car horns.  I never use these.

Finally, in the middle under the upper manual there are 4 preset “pistons” and a cancel button.  These would play factory-preset stop combinations and were not able to be programed.  The buttons are arranged mechanically to allow only one button to be depressed at any one time.  Press another button, and the depressed button pops out.  These probably can easily be used to control piston functions on the computer.

My initial research tells me that the stop tabs can be used on some MIDI hardware to send signals to control organ stops in the software.  I’m hoping that this is the case.  I am beginning to understand that there are complexities involved, but we cross that bridge when we come to it.

Next post: tracking key switch wires.