My Virtual Organ Project

[Aduragbemi]

I've been led to believe poly can become sticky, especially when exposed to moisture.

Oops! Thank you for this vital heads-up. I
Then I'd better looked for felt as fast as possible.

[Aduragbemi]

Testing the reed switch and moving magnet:

I taped an angle bracket (inside out) to the end of the fixed pedal, and taped a reed switch opposite the pedal end with the terminals connected to a multimeter. The magnet is a small (5 x 3mm) but powerful neodymium magnet and it attracts firmly to the bracket. It's position is easily adjusted by pushing with a finger.

[Coenraads]

Aduragbemi, your creativity seems to know no bounds. Do keep us informed!

I have one concern about using such rather powerful Neodymium magnets. There is a possibility that the magnets may interact with adjacent switches. This usually shows up as follows: play a pedal note and while holding it down push down the pedal beside it. Now make sure that when either pedal is released the reed switch opens. I've had problems with the note staying on even when I just use a pair of 12 mm craft magnets which are relatively weak in comparison. One way I've solved the problem is by alternating magnet polarity. In recalcitrant cases, I cut a 4 cm square of tin plate (from a tin can or cookie tin) and mount it between the two switches so that the magnetic fields are shielded from each other. It's a nuisance but it works.

Another problem I've encountered with reed switches is that the length of the ON stroke is too short. In that case use two reed switches side by side in a staggered configuration and wire them in parallel. That way should one switch turn off, the adjacent one will stay on.

I understand that you are using the small one inch reed switches. I've used them successfully in the past but the last batch I got was a bad one. Just by looking at them they seemed to fail. I was using them on a pipe organ pedal board which had to be reliable so I ended up tearing them all out and using Hall sensors instead. They have presented no problems whatsoever and I may use them from now on especially since an Arduino can sense them directly by using the Arduino's built in pull-up resistors. The 3144 sensor does need a strong magnetic field for activation. I screwed two 12 mm Neodymium magnets (with holes) in staggered formation to an aluminum bracket.

[Aduragbemi]

your creativity seems to know no bounds.

First of all, I acknowledge and do appreciate your big contributions to this project. I can't thank you enough for your mentorship, both on this forum and in private through mails. For the records, you hand-held me through pedal midi-fication with the use of Arduino, including loading codes on it and wiring it up for the pedals (I'll be using it for the thumb pistons as well).

I have one concern about using such rather powerful Neodymium magnets. There is a possibility that the magnets may interact with adjacent switches.

Quite true! I planned to use the magnetic shutter method, but I jettisoned the idea when the shutter won't function unless it is large enough (plus, the approach will be more labour intensive). The magnetic strength may be the reason why that is happening. Now I'll be doing the moving magnet method instead.

I understand that you are using the small one inch reed switches

The switches are 14mm and they are covered with plastic. The magnets are 5mm.


I hope their small sizes will help to prevent interference. In addition, I will use the "pole-to-pole" approach (the magnet facing an end of the switch). This reduces a chance of interference further, but the risk of "push-through" is higher than when "side to side" approach (magnet parallel to the switch and approaching it at the middle) is used.

I hope the method and approach turn out well. But if I ever encountered any difficulty, you all already know who I would run to 😅😅.

[Aduragbemi]

Just a bit of caution here. I've been led to believe poly can become sticky

Well, it did not take too long for this to occur. The varnish keeps the wood stuck to the poly foam. But the impetus (or impatience - I don't know which) to get the pedal working can't make me wait till I get a felt to use.

I yanked off the poly foam from the front piece, and as a more tolerable alternative, wrapped foam completely with a thick upholstery cloth and tacked them back to the front piece, without gluing them in permanently. That way I can remove them when felt is available.



Now I have all the pedal keys in place, with the springs underneath.

[Aduragbemi]

HAIRLESS MIDI

I am going to mention names and quotes in this comment. The intention is to recognize your contributions to the success of my project. I hope that is allowed on this forum.

I read Larason2 's reply to a post by Quemeshine; https://organforum.com/forums/forum/...232#post800232

In his comment, he mentioned an instructable by Amanda Ghassaei;
https://www.instructables.com/Send-a...-with-Arduino/

I learnt something BIG (at least to me) at Step 5 of the instructable - viz - eliminating the use of hardware MIDI interface, just by changing the Baud rate in the code to that which the Hairless MIDI will recognize.

I plan to use the Parallel Scanner code by Coenraads - https://sites.google.com/site/casava...ner?authuser=0 - but the code outputs MIDI via the traditional 5-pin MIDI socket, which means I will need a MIDI interface to connect the Arduino to the PC. But with what I learnt, connecting the Arduino's USB port directly to the PC is enough to transmit MIDI. I have tested it out and it worked.

What did I do? I replaced the line Serial.begin(31250) in the Parallel Code with Serial.begin(115200), the Baud Rate in the Hairless MIDI Preferences.

Once again, thanks to everyone who has contributed one way or the other to my project.


Any thoughts about the hairless midi? What are your experiences with it?

[Larason2]

Ah yes, I've known about Hairless Midi for some time. It's a program that converts serial messages into Midi, and many others have used it, or programs like it, to convert an Arduino's serial output to midi. I was under the impression that it wasn't available for Mac, but I see on their website that it is. Either it wasn't, or I must have given it a try in the past, but it didn't work (because of the baud rate problem). When I looked at it in the past, it didn't have the helpful website that it has now, and no mention of the baud rate then. So yes, this could be a useful way to go about it, and I may even try it again myself. Thanks for posting!

[Larason2]

One more comment though is that changing the baud rate may cause problems, so you have to be careful. With the code I posted, its function with the internal pull-ups depends on the code not running too fast. A faster serial connection may speed up the code a bit, which may cause problems. If you're not getting corrupted/spurious Midi signals or notes stuck on/off, or failing to read a note on/off, then you should be fine. If you do get those problems, then you will have to make some adjustments to the code. I don't know Coenraads code well enough to know if it might be a problem with his in more demanding circumstances.

[Coenraads]

Although I had heard the name Hairless Midi, I'll admit to being totally ignorant of its application. When it comes to code, I like to "roll my own." My Parallel Scanner code turns on the internal pull-ups during the initialization part of the program only and not in the Main Loop. As a result, I'd be surprised if the higher Baud rate creates a problem. I'm looking forward to hearing from Aduragbemi once he gets his pedal board up and running. In any case, it's possibly good news for Hairless Midi fans.

[Aduragbemi]

I'm still testing and so far I have no issues, even after wiring up the reed switches - see updates below.

[Aduragbemi]

MIDIFICATION DONE

The entire day (and night) was spent wiring up the pedalboard. I marked the positions of the reed switches on a piece of wood with the curve matching that of the pedals. Then I cut slots into those positions, mounted the switches into the slots and held them in place with hot glue.

Then I soldered one terminal of the switches to a "ground bus" made of solid copper wire. The other terminals were soldered to individual 24AWG appliance wires, which terminate at a socket in groups of 9 wires per socket. I used 8 out of them in 3 sockets, but I connected the 9th wire in the 4th socket to the ground bus.

To connect to the Arduino, I used a M/F DuPont cable.The male pins connected to the appliance wire sockets, and the female end connected to header pins, which are connected to the respective ports on Arduino.

[Aduragbemi]

...and the pedal speaks!

[Coenraads]

Congratulations! Well done.

[jbird604]

Excellent work! So happy that it works.

[Aduragbemi]

Wires and Arduino stacked behind the pedals


I attached metal brackets to the end of the keys and stuck the magnets to them



Pedal now in position under the console desk, illuminated by a USB-powered LED lamp.

[Philip Powell]

Congratulations!

[Aduragbemi]

I'm looking forward to hearing from Aduragbemi once he gets his pedal board up and running.

I have played the pedals together with the manuals. So far so good, no problem at all. Working perfectly.

Hairless MIDI connects with GrandOrgue via another app, LoopMIDI.

[Coenraads]

Wow! This looks fantastic. I've never had reed switches work perfectly the first time. You are good.
I'm surprised to see you activate the reed switches end on. Could you tell me exactly which reed switches these are? I'm also a bit surprised that they don't seem to mind the hot glue given that they are so fragile.

That learning curve you mentioned in your first post wasn't so steep after all. Thanks for letting us share in your success.

[Aduragbemi]

I count myself fortunate to get this working at the first run! I borrowed heavily from the experiences of what many others did, both their successes and failures. I owe this success to all of you!

I'm surprised to see you activate the reed switches end on.

I got the idea of the end-on orientation (I called it pole to pole approach earlier) from the reed-switch/magnet interaction document. This approach has a narrower window of activation compared to side-to-side approach, and hysteresis is smaller as well. Another advantage is less interaction with adjacent switches, because there is greater separation between them.

A disadvantage of this orientation is the possibility of "push-through", when the magnet overshoots the activation window and the note goes off. I tried to eliminate this by bending the switch terminal along the axis of movement of the magnet. While most people cut the terminal short because it adds to the magnetic imprint, to me that's additional 12mm window that the magnet can still activate.

That's a sketch of what I'm describing.

[Coenraads]

That is exceedingly clever and I like your reasoning.

[Aduragbemi]

Could you tell me exactly which reed switches these are? I'm also a bit surprised that they don't seem to mind the hot glue given that they are so fragile.

I got the reed switches from jumia.com.ng , a local online store that imports from China after one makes an order. The product comes in batch of 50 pieces, at the cost of about 36 USD equivalent of Naira, including shipping.


I particularly ordered them because they are described as plastic-covered, after I saw gtc 's post on the type that he used.

I think the plastic covering are for serious protection! Given what I subjected those switches to, I am amazed that not a single one failed, particularly those that I used for various testings on the breadboard.

I got the hot glue idea from Ben Carter; https://virtualpipeorgans.wordpress....ls-and-pedals/ (I also got the idea of hairless midi from him, before I learnt how to modify Coenraads parallel code for hairless midi use). I thought, if the glass reeds could stand the hot glue, these ones should be able to.

[Aduragbemi]

I got the neodymium magnets from stuffslane.com, another local online store (not on pre-order). They are 5mm diameter and 3mm thick, in batches of 100 pieces at the cost of about 30 USD equivalent of Naira, including shipping

At first, because of the small size, I thought I would need two or three pieces to activate a single switch, but I quickly discovered that only one magnet per switch does the job perfectly.

The metal bracket mount for the magnet is also an idea from gtc 's setup. I think the metal acts as a shield around the magnet, but does not affect the side that faces the switch, so helping to reduce interference with adjacent switches.

The magnetic attraction holds them in place on the brackets, but the point of speech can be adjusted by just pushing the magnets with a finger. When everything is finally set, I will hold them in place with a reversible glue.