MIDI Conversion of Three Pipe Organ Keyboards

My search for better keyboards eventually led me to consider purchasing some used pipe organ keyboards and converting them for MIDI use.  I’ve never had the opportunity to try some of the commercially available keyboards, and budget constraints relegated most of them to the ‘too expensive’ category in any event.

The three keyboards that I finally purchased were offered for sale on eBay by a pipe organ company that had removed them from an existing organ and stored them for many years.  They were in generally good condition and ready for use in a tracker action organ, but in my case needed MIDI electronics. I did clean and refinish the surfaces of the keys and replaced some of the felt which was worn and compressed.  I was especially pleased to find these keyboards because they were reversed color (black ebony naturals, cherry sharps capped with white).  For some reason, I have always been attracted to the reversed color keyboards, perhaps because the best organ I’ve ever played had reversed color keyboards.

I have almost no photos of the MIDI conversion process, perhaps because it proved to be much more difficult than I had anticipated. My intent was to use magnets and reed switches to make the MIDI conversion, but I should have taken more seriously the few comments on the Hauptwerk Forum that suggested that this might be difficult to do well.  The absence of any post showing a complete keyboard conversion using magnets and reed switches should have been a warning…

To complicate matters, I definitely wanted to include some kind of tracker touch in the conversion process (for which there were a number of helpful posts on the Hauptwerk Forum and other forums.)

For the sake of other VPO builders, I’ll briefly chronicle the initial and failed efforts that consumed a totally unreasonable amount of time and caused much, much head scratching and frustration.

My initial attempt was to use 1/4″ cylindrical magnets mounted on the rear of each key which would be near and attracted to a steel bar to provide some tracker-like resistance.  Once pressed, the magnets were to move up toward and activate a reed switch.  The tracker touch was excellent, but the magnets were in such close proximity that moving any single key would almost always activate the adjacent key as well.  I partially overcame this problem by reversing the polarity of every other magnet (to limit the building of a larger magnetic field).  This helped, but not enough and after numerous tries I decided that the 1/4″ magnets and their attraction to some steel bar or perhaps a steel screw would have to be separate from any magnets used to activate the reed switches.  First attempt: Failure.

It is worth noting that there are many possible ways to mount the reed switches and magnets which can make it very hard to determine which way is best and most reliable in any given situation.  I used a technical guide mentioned on the Hauptwerk Forum to help make decisions on how to best configure the switches and magnets.

Eventually the 1/4″ magnets were moved several inches toward the front of the keyboards, mounted on a piece of angled steel under the keyboards, and allowed to attract steel screws on the bottom of the keys.  The photo below shows this arrangement.

The screws under the keys were initially adjusted to be about 1/16″ above each magnet.  Finer adjustment could be made by simply pushing the magnets further in or out since the magnets simply sit on the steel angle iron by magnetic attraction.  I found that the polarity of every other magnet had to be reversed, otherwise the entire angle iron began to act like a huge magnet that could affect reed switches several inches away.  (This concept of reversing every other magnet would prove critical in future attempts using reed switches.)  The tracker touch from this arrangement was excellent and could be adjusted from light to heavy touch without much difficulty.  I opted to adjust them for a relatively light touch.

Having obtained the desired tracker touch, I purchased some smaller magnets (1/8″ x 1/8″ and 1/8″x1/4″) to try with the reed switches.  The smaller magnets on the rear of the keys were better, but still resulted in frequent crosstalk between the keys.  I tried a physical re-arrangement (for which I have no photos) of the magnets and switches to keep them further apart from adjacent keys.  This was workable, but needed frequent adjustments, and sometimes left me with a single note sounding continually – clearly unacceptable.  Second attempt: Failure.

At this point I decided to try an approach mentioned on more than one forum in which the magnets and reed switches are permanently located close enough to each other to keep the switch activated.  A piece of sheet metal attached to each key is positioned between the magnet and switch to interrupt the magnet field.  When the key is pressed the piece of sheet metal moves up enough to allow the magnet to re-activate the switch.

The next photo shows this arrangement.

Amazingly enough, this arrangement did work…well, sort of.  For the most part, the keying was reliable, but whenever there was crosstalk with an adjacent key, it proved to be very difficult to position the sheet metal piece to eliminate it and at the same time preserve the firing point of the key so that it was similar to other keys.  I used this arrangement for a short period of time but found it frustratingly unreliable.  Third attempt: Failure. (I do, however, think that with more work, this might be a workable system.)

I was almost ready to give up on the idea of magnets and reed switches for the keyboards and default to some mechanical electrical contact (or perhaps an optical system, or hall-effect system which many others have used successfully.)  However, I found a photo of an Allen organ keying system that was remarkably close to what I was trying to do and which had to be reliable – otherwise Allen Organ would not be using and promoting it.  There was always the possibility that Allen Organs reed switches and magnets were custom made to work together, but still, it seemed like the concept should be workable with my magnets and switches.

After some experimentation in the shop I came up with the arrangement shown below.

Each 1/8″x 1/4″ magnet is located just below the corresponding reed switch and is isolated from the adjacent magnet by the metal ‘tabs’ on either side.  The screw nearest the switch allows for fine adjustment of the magnet up or down to set the firing point for each key.  The polarity of the magnets are all the same (but read further below…)

Each keyboard proved to be easily adjustable for the firing points, and 100% reliable in use.  Reliable, that is, until I put the entire keyboard stack back together.  When the keyboards were stacked the firing points changed unpredictably, and worse yet, some keys were likely to sound continually – a completely unacceptable situation.

Prior attempts had alerted me to the possibility that when the keyboards were stacked there could be some magnetic interaction between the keyboards because the ends of the keys are only about 1″ apart vertically.  All told, both the tracker and keying magnets were uncomfortably close to each other horizontally and vertically.

Initially I thought that the tracker magnets on the next lower keyboard might be affecting the keying magnets of the above keyboard, but a few experiments indicated that the effect between keyboards was limited to just the keying magnets.  Eventually I realized that I should reverse the polarity of all the center keyboard keying magnets to help cancel any interaction between keyboards.  This, fortunately, was completely successful and restored each keyboard to 100% reliability when stacked.  This arrangement is recent enough that I can’t judge its long-term success, but given the lessons learned from the failed attempts, I think all of the issues have been dealt with and I expect long-term stability and reliability.  Fourth attempt: Success at last!

I’ve not shown the wiring on the rear of any of these keyboards because each was nearly identical to that used on my pedalboards using Midi-Gadgets components.

Would I use magnets and reed switches again?  Had I known how difficult and time-consuming this process would have been from the beginning, I probably would have not have chosen this route.  However, having worked out all of the issues, I probably would attempt it again because it seems like there is little that could go wrong in the long run.

If I had been required to pay myself for all of my time, any of the best commercial MIDI organ keyboards would have been a clear choice, hands down!