I’ve soldered up my kit and everything works great – all except for the E key which plays no sound unless I rest my finger on any of the resistors while tapping the E with my other hand. If I do that all the notes play normally including the E.
I’ve checked the values of the resistors and everything seems within tolerance. I’ve also reheated all the solder connections a second time just to be on the safe side, but the only way I can play an E is by placing a spare finger on any of the resistors at the front of the PCB.
Just put ours together (bought and built at the end of 2019, 10 months after Ian’s post) and had the same problem. A few keys do not work unless you also touch the battery’s metal case which implies that the code doesn’t auto-configure the key’s capacitive settings (hard coded?) or there’s an issue with the auto-configure (if not hard coded).
Probably would be resolved if the board is installed in a metal box and you held the metal box while using.
I noticed that the resistors connected to the different pads are different (high) values. Do they need to be tweaked to get this to work?
Mac, you are on the trail! First let me mention that it was an oversight that we failed to post a response to this thread: we talk with Ian a lot, mostly off line, and I overlooked this post. Mac, each key has a different capacitance influenced by its size, shape, and position on the PCB. The resistors were selected after some theoretical analysis and LOTS of empirical trial-and-error. Plus the whole system is influenced by its environment: temperature, humidity, and how dry your skin is! So we have found the following tend to stabilize the keyboard:
Touch one finger not to a resistor or the battery, but ANYWHERE on the PCB outside the 13 keys.
Experiment with a wall-wart power supply instead of battery, and earbuds or other external speakers instead of the (sketchy) on board speaker.
Lick your fingers before playing!
Adjust that E resistor down to a lower value.
Playing with the MIDI IN jack, controlling the Macchiato with an external control keyboard or sequencer, bypasses the capacitive touch keyboard all together and performs with no instability.
Regarding a metal case: Not such a good idea. The metal itself, if it contacts the PCB at any point, will affect the capacitance of the whole system. However, one user had great success adding an additional ground plane to the bottom-inside of the cabinet. I think he used a piece of heavy aluminum foil (copper would be MUCH better) and a little wire that he connected to ground at the center of the board, possibly using the center standoff screw. This gave him good stable performance. He found he had to isolate the foil from the five screws in the cabinet bottom; otherwise, when he held the synth in his hand, touching any screw would throw it off again!!
I am also having the same issue but with my C#/Db key. I can measure the resistor and it read accurate however it doesn’t work not matter where I touch the PCB or other keys and resistors. I do not have any 750k resistors around, but do have some 680k. However I am afraid of separating the contact to the lead in the board as I have already reflowed the solder then pulled the resistor out and re-installed. Will putting in a 680k affect the tone of the C#/Db key?
Derek, sorry for the late reply. No, the resistor value has no influence on the tome. The tone is generated digitally by the microcontroller at a frequency hard-coded into the software. You can lower the value of the resistor on your bum key by adding that second resistor in parallel. This will lower the resistance across the key. You can try this by just holding the leads in place across the pads. Tricky, but a quick check. If you want to increase the resistance, you will have to pull out that resistor again and replace it with a higher value. You can’t hurt anything with different resistor values, and the magic number may vary from unit to unit, so experiment.
I put foil under the board, as mentioned above, but that alone didn’t do the trick for mine.
I had to tweak a few R values differently; ended up with these:
R1 from 470k to 660k (330k + 330k series)
R5 from 560k to 820k
R8 from 560k to 820k
R11 from 750k got 2.4M in parallel (1.2M + 1.2M, so about 570k total).
waynemcl,
I’m glad you got it figured out. Capacitive touch systems can be pretty finicky sometimes. Usually altering the resistances is enough to get this system stable again.
-Brach