MIDI Pedal Board, the electronics
Continuing from part 1 of this project, I’ll run through what I did to get my MIDI Pedal Board to more than just a (nice?) shell. All that stuff which makes it tick!
Nowadays I just can’t be bothered to start etching PCBs and whatever to get the electronics of a project sorted out. And to steer well clear of all that, I chose to go with Roman Sowa’s excellent selection of MIDI converter boards. I was already a happy customer with his Midimplant board, and upon ordering that had noticed that Roman also had some boards on sale for midifying organs. A look through the catalog and there it was, the Pedscan, a board which can handle a keyboard up to 32 keys and one analog input. Just what the project needed! I also ended up ordering a Potamp board, more on that in the Volume pedal assembly section below.
Must admit that when I started tearing the organ apart, I didn’t at all check whether it’s actually possible to make the entire Pedal work electronically. A high trust for “surely someone out there has come up with a solution”, just what every project needs.. But it also could’ve ended up with the project getting binned. Yeah, I sure like to live dangerously :)
Volume pedal teardown & tests
Of the available pedals, getting the analog control (= volume pedal) to work seemed like the trickier one so I started with that. Without looking at all I had assumed that the pedal must have some sort of a potentiometer gear mechanic inside. So you can imagine my surprise once I got the pedal torn to bits and saw that it uses a 24VDC bulb and a photo resistor instead!
Ok so hooking this thing up is going to be a bit more work than just “solder some wires”..
First up I had to determine the ballpark resistance range of the photo resistor. Going from about 200R to 1Meg (light entirely cut off), well, that’s way beyond the recommended maximum range of Pedcan (50k tops). I then dug up some photo resistors from the parts piles, replacing the original one to see how the substitutes perform. At best I could hit a top resistance of about 300k, still way too much.. Definitely the plastic sheet needs some tweaking instead. But prior doing so, the 24V bulb needed to go. No point in adjusting the mechanical parts for resistance range if I’m going to swap out the bulb, and doing so very likely affect the photo resistors range in the process.
Now the bulb might’ve been perfectly fine for the organ with it’s internal 30VDC power supply, but the Pedscan only needs 5VDC and I’m certainly not going to include a 24V PSU just to power this one bulb! To make this part of the build “5V compatible” a LED it has to be, and I had just the replacement readily at hand: A similar jumbo LED to that which I used to mod a backlit power button on my Sega Megadrive 2. It turned out to be even mechanically about the same size as the 24V bulb, so it fit neatly inside the pedal mech :)
With the LED test setup stuffed inside the lamp assembly it was time to start tweaking the plastic sheet. Since I didn’t want to damage the black paint by scraping off some of it, the next best solution was to test moving the plastic sheet by drilling additional mounting holes to it. The sheet could be moved forward about 3mm as-is, and another 3mm more by trimming the long outer edge (the narrow green area on photo below).
With the 2nd pair of holes, I got the lowest high resistance down to about 6k8. This left some room for further tweaking required later on. The Pedscan manual mentions 20k as “optimal”, so with 6k8 there was good room to go higher too. Surely the resistance value would still shift a little once I get the LED hooked up to a proper PSU instead of the button cell battery, and installed properly. For the tests I just had the LED loosely placed inside the lamp assembly, not even properly pointed at the photo resistor.
During all the testing I managed to break off the feet off the original photo resistor. This forced me to use one of the substitute resistors I found for testing. Annoying but not a biggie; the substitutes low resistance value is about 100R higher than with the original. The latter would’ve of course fit the lamp assembly better. The substitute is much smaller, thus some light could possible leak inside the pedal assembly from the outside.. But it’s a very vague “if” anyway.
Volume pedal assembly
With the tests carried out and results pointing that the volume pedal could be modded to ‘something usable’, it was time to figure out how to hook it up to Pedscan. Roman was kind enough to help by giving pointers how to set things up. By adding a resistor between the top and middle pins of the Pedscan analog input, then connecting the photo resistor between middle and bottom pins I’d end up with a resistor divider similar to a potentiometer. There was one short-coming with this though..: Because of the fixed resistor at the top pin and the lowest resistance of the photo resistor, the pedal wouldn’t be able to transmit the full range of MIDI volume commands (CC7, 0x00-0x7f ). My plan all along was to use the volume pedal to control the master volume of my sound card, so a partial MIDI range just won’t do.
Now about that high trust towards “someone coming up with a solution”, coincidentally Roman also sells the Potamp board to solve issues exactly like this. This board allows you to tweak eg. a potentiometer to send the full MIDI range even only a partial mechanical range would be available. Thus perfect for my needs too. And coming in at 4.5eur a piece, I seriously can’t be bothered to start putting together some hack of my own. Pedscan & Potamp boards to the shopping cart please, and on with the build!
After installing the LED inside the assembly, I hooked it up to a proper +5V feed via the usual current limiting resistor.. Well, in my case “usual” was a bunch of 220R resistors soldered in parallel as I didn’t have any smaller values at hand. Not a very neat looking solution, but you can’t see the resistor pack once the pedal is assembled :)
For the combined resistance value of top resistor + photo resistor I decided to target 10k. As said, the lowest high resistance I could reach during the tests was about 6k8, so I decided to use a 1k8 for the top resistor. I then started fine-tuning the high resistance upwards. This I did by adding to the black paint area using a black magic marker; add a thin line to enlarge the area, measure the high resistance, add more to the area, measure, and so forth. Once I got the resistance up to about 8k4, I decided it was ‘close enough’. I didn’t want to start experimenting with the area shapes, so just increased the area by following the existing border.
For the final version, I didn’t even need to use the new hole pairs I drilled for tests. The original position & increased black area on the plastic sheet were all good to hit the 8k4 value. Volume pedal sorted out!
On a side-note, I must admit I really like this light source & photo resistor approach mechanically. Having next to none moving parts it’s about as maintenance-free as it gets, just the pedal on it’s axle and that’s it! Having the 24V bulb replaced by a LED should also contribute well to ‘maintenance-free’.
I got the power (eventually)
As Roman advertises it the Pedscan could be easily run off batteries. I figured that since the pedal is going to sit next to the computer and be in use only when all my other gear is running, I could just as well add some external PSU. It’s not like the pedal is anything portable anyway, so if the need ever rises to have a pedal-to-go, I’m then probably better of getting something like the Behringer FCB1010 or similar. And on the other hand, having a solid +5V feed was a prerequisite to using Potamp.
So yeah, the thing needs to be powered up.. To boot with I decided to test leeching the power off a USB port on my computer (computer on = pedal on), but for some reason that didn’t power up combination of Pedscan + Potamp + volume pedal LED despite there should’ve been enough current. If I hooked up the Pedal using a separate PSU over the USB connection it powered up fine. Dunno, maybe something else was up with the Pedal but I didn’t bother investigating further and moved on.
Next, I dug up this AC/DC switch mode PSU brick that outputs +5VDC (@ max. 2A current) left over from some binned D-Link router. Way too overkill for the project, but whatever. With it the Pedscan powered up neatly, but now it was also outputting a constant stream of CC7 even despite I didn’t touch the volume pedal.. Crap! A closer look using an oscilloscope revealed that with electronics powered up, the PSU output contained switch-mode ripple of about ~350mVpp with frequency around ~14µs. I tried adding some smoothing capacitors to remove it, to no effect.. Ok then, on to what I knew would certainly work (cue the fanfare): a higher voltage external PSU and a internal +5V linear regulator.
Dug up another AC/DC brick outputting 8VDC, hooked it up to a LM7805 with the necessary bypass capacitors and there it was, a ripple-free power supply for the project! A higher AC/DC brick would work too, but (iirc) the LM series regulators need a minimum of +3V over the spec’ed output voltage to work. Since there was the one 8VDC PSU with no other use.. Yeah, fine by me. I didn’t bother soldering the parts to any prototyping board, so the regulator does look hilariously hacky.
LM series specs also state that for current draw in excess of 100mA you need to install a heatsink. My Pedal isn’t going to use as much, but I figured since the rear panel is all metal I might just as well use it as a heatsink by installing the LM7805 on it.. A similar approach as I used with my Midibox SID synthesizer.
Final thing on the list: hook up the keyboard pedals. I decided to spare its original harness cable just for this, but it needed to be modified to form a matrix (see wiring diagram below) and have diodes inserted before each pedal switch. Worth noting is that the Pedscan needs the lowest 10 keys of the matrix connected in order to adjust settings by hand. Had the organ keyboard had less keys I would’ve had to hook up additional buttons.
But yeah, a straightforward solder job with not much to it. I decided to solder diodes directly to the switches. And umm.. It was a nice day. I was listening to music while soldering.
Things come together
The final stretch for the build was mostly about drilling holes on the back panel to properly install connectors. I also hooked up a the Edit button which, once pressed, puts the Pedscan into settings input mode. The particle board was way too thick for the button mounting threads to appear on the inside, so I just made the installation hole a very tight fit. Maybe I’ll just give it some hot glue if the button decides to come off any time later.
The Pedscan is small enough not to have room for any mounting holes, so on my build it’s just all the wires that hold the board in place. It didn’t seem to move much when I tried shaking the Pedal.. Guess that’s “ok” then. Potamp is a single-sided board though, so it I could install by using a stack of double-sided tape.
Last few steps of the build, I decided to install the Hammond logo (mentioned in part 1) to the front and the original 123J3 serial number plate to the bottom. Got to include a bit of the history somewhere too ;)
Probably of little interest to anyone, but I’ll do a post about the setup too.. More of a series that way, perhaps?