Toe Stud Module

"An engineer can do for a dime what anyone else can do for a dollar" - Anonymous


The board which holds the toestuds comes out as a single module, now that I've removed the extraneous wiring.

As shown here, the Crescendo is on the right and the swell on the left.  However with a 3 manual instrument, you want two boxes.  As the Crescendo doesn't work very well on Hauptwerk, and in the world of infinite memories, it seems a crutch, I'm using these two shoes for Swell and Choir.  Of course, the positioning is vital, so I created a dummy shoe at the right height as the Crescendo shoe.

Workarounds

The scanner I'm using takes 32 inputs (4 groups of 8) and 3 analog shoe inputs.  The mechanisms on these shoes are essentially a set of contacts engaged by a ground plate.

So I've worked around it by building a microcontroller which scans the inputs from the shoe mechanism, then outputs a PWM to create an analog voltage (which goes into a scanner which then converts it back to a digital signal).

Seems convoluted, but it preserves the simpler interface of a single MIDI combiner.  It also leaves all 32 inputs open for pistons.  I have 8 from the manual (see previous post), and 15 from the toe studs.  A side benefit is that because I'm filtering the output of the PWM, there is some inertia in the interface, which better emulates the real action of a shutter mechanism.

The previous picture shows the old crescendo mechanism, now the swell shoe.  The old swell mechanism, now the choir shoe, is connected to the digital to analog microcontroller blob attached to the board. I picked off 8 inputs from the crescendo mechanism (swell shoe), and 8 from the swell mechanism (choir shoe). You can also see the "crescendo" dummy shoe in place.

So the result is as follows

Labels are coming. Now the only module to do us the stops, and I'm waiting a couple of weeks for the new USB versions. 

Modular is better, I think

I have made a lot of progress on the console.  At this point, I'm waiting for parts.

Modularity

 I've decided that consoles are big because of the wiring used in old technology.  I've decided on 3 manuals for this console.  I believe I can make this console with the following relatively easily transportable modules:

1) Keyboard stack - This includes the 3 keyboards.  Outputs here are 2 * 4 pin cables for the keyboards, and an 16 conductor ribbon cable for the pistons.  The 4 pin cables are arranged  one for the bottom two, one for the top one because of how the scanners are designed, using the BBS-1K scanners from www.midi-hardware.com.  I've "prepared-for" the fourth manual.  installing all of the ribbon cables but not the scanner in the old Zimmer SW manual.  There's a big scratch on the keys in the center of the keyboard, so it's less than desirable.  BTW, I think "prepared-for" is organ builderese for "You don't have the money for this, and you probably never will, but I'll put this in to make you fell good about it".  So essentially, this module is done.  I have put handles on the side of the stack to allow for easier moving.

This is built such that the keyboards flip up for contact maintenance and repair 

2) Pedalboard.  The output here is a MIDI connector, and a 5V power cable.  This module is done.

3)  Pedal piston module.  I'm planning on this to have all of the pistons (10 generals and PREV on the left, 3 Ped divisionals, NEXT and Zimbelstern on the right).  Outputs here are 2 16 pin ribbon cables for the 16 pistons, and 2 2-pin cables to interface to the shoes.  Also here would be the expression pedals.  At this point, I'm thinking of just a Swell and Choir shoe - in fact, one shoe with both on it should be sufficient.

4) Stop tab rail.  Given that I only have 26 SAMs, and new ones are $33 each (plus $15 engraving), the <$12 all-in per stop cost for lighted rockers makes way more sense.  (I'll probably develop the MIDI driver circuitry just for experience)  I'm planning on 80 lighted stop tabs from www.midiworks.ca.  Two rows of 24+16 (SW, GT/Solo) and (Ped, CH).    This configuration will fully support the CUI Casavant.  They are coming out with USB interfaced modules, so I'm waiting on those.  If they get delayed (I'm an engineer, I know how this works), I may just go back to the MIDI interfaced ones, though the USB ones would make the interface a whole lot cleaner.

4a) The music desk will be attached to the stop tab rail on the top.  LED lighting strips, with separately controlled brightness, will be attached to the top (music) and bottom (key desk) of the desk.

5) A collapsable, or at least disassemble-able table with alignment pins for the keyboard stack and the pedalboard.  I'm deciding if I should hack apart the Zimmer console to use the wood, or design from scratch.

It's all about ground

The keyboards from Ryan Ballantyne feel nice, however once I started to use them, the contacts were intermittent.  Given that I put multiple contacts in parallel, it seemed incredible that they wouldn't work.  So I played with cleaning the contacts, to no avail.  Then I looked at the grounding.  The key bars are grounded through a ground wiper in back, and through the return springs.  The return springs are soldered on the bottom side, but not on the side on the key bar.  For the intermittent keys, all I had to do was solder the springs, and instantly the keys worked reliably.  I only soldered the ones that were problematic.  I may regret this later, or I may just go back and solder the rest.  Keys in soldering these 1) use flux, 2) protect the felt pads from the hot key bar - I used a small piece of wood.

Cable Dressing

As in everything else in this project, the second one is better than the first one where you are trying to figure out how to make it work.  Guess which one I did first?

Pistons

Given my experience with my existing console, the only pistons I think are necessary are generals, pedal divisionals (just a few), and NEXT.  Other controls, not necessary to be on the keyrails which are necessary are Memory 000, Memory +10, Memory -10, Memory +100, Memory -100, and system power off.  I will also add lighting controls Music Desk +, Music desk -, Pedalboard+, Pedalboard -, and Keydesk+, Keydesk1.  So to that end, I built two piston keyrails using the pistons from the Ryan Ballantyne keyboards.

Choir piston keyrail: Set, Next, General Cancel.  The good thing here is that the keyscanner that I used from www.midi-hardware.com has 64 inputs.  61 are used for the keyboard, so 3 are left over and can be used for these three pistons.

Gt piston keyrail: Generals 1-8, Next.   I'm not doing 1-10 - I'll put 1-10 on the pedal pistons.  8 pistons just look better in this group.  Next was put into the keymatrix, and the 8 generals go out to a separate scanner through a ribbon cable.

This weekends work will be in finalizing the key stack, and working on the pedal piston/shoe module. 

2 manuals, three or four?

I was able to complete the electronics to drive the stop action magnets - somewhat.  The drivers work, but the MIDI logic doesn't.  My adaptation of the Midi-Hardware LED lit switch driver to this structure didn't work.  If I move forward with this, I'll probably do the entire structure - from MIDI to SAM drivers, without a third-party MIDI interface in the middle.

As well, there are only 26 SAMs.  26 tabs is barely adequate for a 2 manual instrument, and will be entirely inadequate for a 3 manual instrument.  So I believe for now I'll abandon using the SAMs in favor of lit rocker tabs.  More on that later.

This weekend, I dropped off the pipe driver parts and electronics with Ryan Ballantyne, and was able to pick up 2 keyboards from him that he wasn't using.  The plan was to put one keyboard onto the two I had, however these keyboards feel WAY better than the plastic Zimmer keyboards that are on the console, so I want to use both of the new keyboards, and 1 or 2 of the zimmer keyboards.

But what would I do with one leftover keyboard?  Should I build a 4 manual console in this tiny shell?

They'll fit, but stoptabs will be another problem.  Options here are the sides (which have room for rows of 8 across) and a box built on the top just below the music desk.  Still working on the design here.

It's amazing what electronics have done to how the organ is built, even pipe organs.  This organ build who knows when, has 3 sets of contacts for the top manual and NINE sets of contacts for the bottom manual.  The contacts are moved in and out of connection with a key bus bar by pallet magnets, and there is a GIANT bundle of wires coming out. 

Fortunately, all of this can be removed, and replaced by scanning electronics which reside at the keyboard, allowing a cable of 4 wires to carry all of the relevant information from the keyboard to the rest of the system.  I forced the contacts in the "ON" position, and wired across 3 in the top manual and 4 or 5 in the bottom manual for redundancy.

It only takes 2-3 hours to wire up the ribbon cables, and results in a fully self-contained keyboard.  No giant cables running from it, and it can easily be unplugged and removed for maintenance or whatever.

Why am I doing this?  My wife wants to know.  Of course, she's amazed any time someone buys my Hauptwerk sample set.  It's a weird and esoteric hobby.  Maybe not as esoteric as something like making your own paper, but it's up there.

Pedalboard and Reed switches

Reed switch are a pain.  Though they create a reliable connection, the switches themselves are delicate, and it seems like magic as to what will and won't trigger them.  Sometimes, magnetism seems kind of magic anyway, so, no surprise.

As with most things, experimentation and making mistakes teaches us the better way to do things.   Initially, I set the switches crossways

And measuring carefully, just set the magnets into the pedals

But as you can see from the ones with 2 and 3 magnets in them, there's more magic in them than I at first thought.  So I found some great double sided tape in the garage (finally, a use for the random stuff that I've collected over the years), which allowed me to move the magnets until they worked, then glue them in place.

The next thing I tried was to set the switches 90 degrees from where they were.  This worked WAY better, as they were further apart from each other and had fewer interactions to worry about.

Whenever you use reed switches
1) Buy extras - the glass tubes are fragile
2) Figure out a way where the reed switches - all parts of them - are as far away from each other as is possible
3) Figure out a way that you can move the magnets around until they work well.

Video of the pedalboard in action to come soon.  It's a busy week coming up.

2 manual Zimmer console conversion - Initial steps

I was fortunate to be able to acquire a 2 manual console which I'm in the process of converting to a Hauptwerk controller.  St. Paul's Lutheran Church had had their 16 rank unified Zimmer in storage for 2 years, having left their building in Laguna Beach in search of a place in Irvine.  With no end in sight with respect to a new permanent home, they decided to reduce their storage costs and give me and my friend and organ builder Ryan Ballantyne (http://www.rmballantyne.com/) the entire organ.  The pipework and chests were taken by Ryan for repurposing, and the console was left with me, as Ryan didn't feel it appropriate to use it in a church installation.  Two of the ranks from the St. Paul's organ are already in place at St. John's Lutheran in Orange.

It turns out I picked up a 2-manual Moeller console on the same day, but we felt due to the quality of the console, it would be a more appropriate console to repurpose for a rebuilt pipe organ, so I ended up with just the Zimmer console.  My wife was glad.

The first step is to clean out all of the obsolete and superfluous circuitry.  As all of the memory and combination system will be taken care of by the computer under Hauptwerk, any of this can be removed.  The only objective is to retain all of the switches and wiring to the Syndyne Stop Action Magnets.

 The rack of equipment containing the combination system above the crescendo was removed, along with all of the Peterson coupler action pieces, and the power supply.  The plan, since the power consumption is orders of magnitude lower, is to simply use small power supplies for the MIDI equipment.  Doing the math on the SAMs, there are 28 of them at around 30 ohms.  At 12V, it's 1 ohm if all of them are on at the same time, so it's easy to find a 12A 12V power supply, if no other place, than with a standard 300W PC power supply.  This supply is far more efficient, and lighter than the goat anchor here in this console.  I think by removing all of this equipment, the console is half the weight that it was before.

 I'm using MIDI scanners from a company I found online called "MIDI Hardware" (http://midi-hardware.com/).  Roman Sowa runs the company from his home in Poland.  The circuits are small yet robust, and very inexpensive compared with other alternatives. 

After removing superfluous wires, I wired up the keyboards.  The console was built with pins for each switch and a common ground.  This made it easy to wire ribbon cables up to them, and then run them to the scanner hardware.

Fully wired keyboards are run to the Midi-hardware scanners through the ribbon cables SW on the left, GT on the right.  I'm planning on adding a 3rd manual, since hardware is so compact now.

The pedal was a different story.  The original design of the instrument had the switches for the pedals wired into the console.  This is heavy, bulky, and limits flexibility.  I decided to make the pedal a completely stand-alone MIDI pedalboard, with its own scanner and MIDI output.  This required some rewiring of the switches from a common ground system to a matrix system.  Diodes were added in order to support the inexpensive scanner design.

MIDI output for the pedal board, along with a USB connector for power.  Better would be to have a USB MIDI system, but that hardware is way more expensive.

The switches had been sitting unused for 2 years.  These switches are 2 wires wiped by a copper bar.  After completing the installation, I found that they don't work very well.  I'm going to have to replace them with reed switches before the instrument is usable.  Even after cleaning the contacts, they don't work very well.  I suspect it has a lot to do with the fact that the original system actually drew quite a bit of current through the switches, which tended to hide a lot of keybounce issues.  With modern systems, they really don't work well.

Then there were the Stop Action Magnets (SAMs).  In this system, the reed switches were reference to 12V and the magnet coils were referenced to ground through big bus bars across the top.

The diode array coming in the back is for the SFORZ action, which is unnecessary in the new system.  Ground is on the right, connected to one end of the magnets, and +12V on the left, connected to the switch.  Conveniently, though it's very straightforward to reverse engineer, Syndyne gave the patent number 3832658 which provides some details on how they are hooked up.  I've built a circuit which will drive these magnets, since you want momentary actuation of the magnets.  This simple circuit will bring in the MIDI state change for the stop, and look at the edges and create a short (probably 200ms-400ms) pulse to energize the ON or OFF magnet.  However, it's way more convenient to have the switch referenced to ground so you can use a ground referenced scanner to determine the state of the switch, and the magnet coils referenced to 12V so you can use N-FETs to drive them on.  Fortunately, there are no active circuits on this board, so all I have to do is swap the bus rails and drive the right one to the coils at 12V and the left one to the switch at GND.

So left to do is

1) Reed switches for the pedalboard
2) SAM drivers
   a) write microcontroller code
   b) Test and install using the Midi-Hardware LITSW stoptab drivers
3) Wire up the peripherals to a scanner (Pistons, Toe Studs, Swell mechanism, Crescendo mechanism)
4) Add the 3rd keyboard

Custom work done on the Midiworks.ca 4 manual console

Building the Lighted Music Desk.

I was able to finish the last major piece of construction on the organ today, the building of the lighted music desk.  I had considered a number of things for lighting, including using some LED structures I had built about 8-9 years ago for garden lighting.  I determined that they weren't bright enough in aggregate, and they got too hot.  While they worked for garden lighting, for this application they aren't so great.

 


I have been monitoring LED shipments into the home stores.  Generally, my opinion is that LEDs will change the nature of architectural lighting fundamentally, so that structures and the way that lighting is done will be very different from what we are used to now.  We're already seeing this in automotive lighting, where industrial design cues are now informed by LED lighting, and would have been impossible with standard incandescent or fluorescent lights.  There's a great "new" product (it isn't really that new, just now relatively inexpensive and available at the home stores) now on the shelves at Home Depot.  After considering a lot of options, I decided to use this.

 I have a plethora of 12V power supplies around, and even if it weren't there, I could pull 20W off of the PC power supply without an issue.  This product has 12 feet at 25W, or about 2W/foot.  I'm using almost 8 feet, 4 feet on the top to light the music and 4 feet on the bottom to light the keyboards.
 

 I used the hardware that was shipped with the old music desk and built a new wooden structure.  While I could have better matched the wood on the rest of the console, at the request of my wife, I was only to use wood I had available in my shop.  So a left over piece of 3/4 birch plywood for the bottom vertical part of the rack and a nice piece of old pine from some ancient bookshelves for the horizontal section sufficed.

 I routed a thin groove on the top and on bottom of the horizontal section and placed the LED strips therein.  Then the strings were wired together using my trusty soldering iron.

Mounting the desk provides a lot more space for music, and I've eliminated the requirement for the overhead lamp.  And so the project is finished.  I'm sure there will be tweaks in positioning of the stop tabs and the like, but I'm pretty pleased with the instrument now.

I had considered building a music desk which incorporated the two touchscreen monitors.  Theoretically, one could have music displayed on the monitors.  However I'm so much more comfortable with paper at this point, and annotations are much easier.

Building the Stop Tab bolsters

This took the longest time.  I've been designing this in my head for a very long time, and it was nice to see it come to fruition.

The stop tabs are nice, Schlicker-style lighted stop tabs.  Moving ones are always an order of magnitude more expensive, so lighted are a good compromise.  The tabs are really nice in that one direction sends a different message than the other, so that ON is ON and OFF is OFF, and there is none of this toggle nonsense that is in most lighted stop tabs.

The construction of the bolsters was rather complicated.  I wanted the angle to match the angle of rise of the keyboards, and sweep out from the console for easy access.  This resulted in a complicated trigonometry problem I could never solve.  So I just built it.  It looks great now, but I know that there is a couple of degrees of error in a couple of places.

Here's the bolster partially assembled.  It took two full days in the shop on a weekend to build them.  I used left over wood - old walnut plywood cabinet doors.  Ends up looking great, and no new wood to buy!

The completed bolsters with the stop tabs installed are here.

It took another week of evenings to build the cables, and a few evenings to debug them.  Eventually, all was figured out.  I was able to use ribbon cable for most of it, which really reduced the construction time of the cables.  Only the ends that went into the controller were individual connection IDC connectors.  This was good, in that I had to swap a few connections around when debugging.

The master controller in the lower left can control 128 stop tabs.  There are 64 on each side, and I had to get the cables from one side to the other.  I ran them under the table above the pedalboard so that the runs would be as short as possible.  To hold them in place and cover them up, I hogged out a channel in a piece of oak moulding. 

Note also the LED strip light.  Gives great light for the pedals. I really need a structure for the music desk and keyboards.  I could buy commercial like this, but I bought a bunch of high output LEDs for a previous project (LED garden lighting) which I might as well use.  Of course, it will be a development effort...  This will be in conjunction with building an oversized music desk as I had built for my old Johannus.

After debugging the wiring (with great instructions from Atilla at MidiWorks - big shout out) through the built in self test, there was the challenge of getting all of the tabs to light up.  Another big shout out to Darryl at MidiWorks for his hints in configuring Hauptwerk.  A couple of key things here.  Firstly, you need to check the box that prevents Hauptwerk from checking for MIDI loops.  Secondly, the pedalboard needs to be on a separate MIDI controller.  For some reason, the toe studs which run through the pedalboard have a MIDI feedback which messes up the rest of the system.  So on my system, I used my old M-audio UNO interface to receive only MIDI IN from the pedalboard.  The rest of the system - 4 keyboards and the stop tab controller - go through the USB MIDI interface embedded in the keyboard.

So the (almost) completed system is here

Still left to debug are
  1) Getting the lighted pistons to work.  The pistons operate, but the lights don't work yet.  The message is getting there, but either it's the wrong one, or it's not being properly operated on.
  2) I had a display from midi-hardware.com (see the lower right of the left bolster) which I'm integrating.  This display will be to show the current bank of memory in play, and the loading percentage of the instrument.  This will be necessary to do any headless operation.

Still left to build is the large music desk with integrated desk and keyboard lighting.
 
For the stop layout, I borrowed a lot of ideas from Cameron Carpenter's Marshall & Ogletree design http://www.marshallandogletree.com/.   Pedal is White, Swell is Red, Great is Blue, Positive is Green, and Solo is Yellow.  Would be cool if I could get LEDs to match, but that's another project. 

Within the color for the division, the color of the text indicates the type of stops.  Black for Principals, Red for Reeds, Blue for Flutes and Green for Strings.  Couplers would be solely color based.  Because of having multiple Hauptwerk instruments, I chose to name the stops generically rather than by the specific pipe construction which would be specific to an instrument.

There's still some changes to be made.  I used removable inkjet printable labels so that I can modify and change what doesn't work and improve the layout.

It's been a fun project.  Anyone else want one?

Lighted Music desk for my Hauptwerk Console

I was able to finish the last major piece of construction on the organ today, the building of the lighted music desk.  I had considered a number of things for lighting, including using some LED structures I had built about 8-9 years ago for garden lighting.  I determined that they weren't bright enough in aggregate, and they got too hot.  While they worked for garden lighting, for this application they aren't so great.

 


I have been monitoring LED shipments into the home stores.  Generally, my opinion is that LEDs will change the nature of architectural lighting fundamentally, so that structures and the way that lighting is done will be very different from what we are used to now.  We're already seeing this in automotive lighting, where industrial design cues are now informed by LED lighting, and would have been impossible with standard incandescent or fluorescent lights.  There's a great "new" product (it isn't really that new, just now relatively inexpensive and available at the home stores) now on the shelves at Home Depot.  After considering a lot of options, I decided to use this.

 I have a plethora of 12V power supplies around, and even if it weren't there, I could pull 20W off of the PC power supply without an issue.  This product has 12 feet at 25W, or about 2W/foot.  I'm using almost 8 feet, 4 feet on the top to light the music and 4 feet on the bottom to light the keyboards.
 

 I used the hardware that was shipped with the old music desk and built a new wooden structure.  While I could have better matched the wood on the rest of the console, at the request of my wife, I was only to use wood I had available in my shop.  So a left over piece of 3/4 birch plywood for the bottom vertical part of the rack and a nice piece of old pine from some ancient bookshelves for the horizontal section sufficed.

 I routed a thin groove on the top and on bottom of the horizontal section and placed the LED strips therein.  Then the strings were wired together using my trusty soldering iron.

Mounting the desk provides a lot more space for music, and I've eliminated the requirement for the overhead lamp.  And so the project is finished.  I'm sure there will be tweaks in positioning of the stop tabs and the like, but I'm pretty pleased with the instrument now.

I had considered building a music desk which incorporated the two touchscreen monitors.  Theoretically, one could have music displayed on the monitors.  However I'm so much more comfortable with paper at this point, and annotations are much easier.