1

enter image description here

I have a question regarding PCB design and rubber keypad contacts. I did this circuit for prototyping a piano keybed but some contacts seem to be unstable: when pressed it switches from 0 to 1 continuously - this isn't a classical debounce issue but some kind of electrical instability. I know that it is best to use carbon ink, but it is way too expensive for a prototype.

My question is then: do you think my design of the pad could be the issue, or is this a pull-up resistor issue, or simply some dirt?

Thanks

Andrew Morton
  • 2,632
  • 1
  • 18
  • 27
Olivier
  • 41
  • 3
  • 1
    Have you tried shorting it with a wire? – HandyHowie Jun 19 '18 at 09:44
  • 3
    HASL surface finish? I would go for ENIG if carbon is out of the price range, likely to be a bit better then the typically unflat and prone to oxidation HASL. Also, your feature size looks very large, I would expect a lot more fingers under the button. – Dan Mills Jun 19 '18 at 10:40
  • 1
    Precisely what does this mean "switches from 0 to 1 continuously" – Andy aka Jun 19 '18 at 11:02
  • 1
    It's kind of hard to say when all we have is a photo of a part of a PCB. Is there a reason you haven't shared a schematic? – Bort Jun 19 '18 at 12:50
  • (a) "is this a pull-up resistor issue" We can't tell, with the info you've provided. As already requested, you must supply a schematic, preferably the whole schematic, but at least the signal paths from keyboard back to MCU (or whatever). (b) Do you have access to an oscilloscope? If so, what did you discover when you viewed the relevant signal(s)? (c) "some contacts seem to be unstable" If only some contacts are affected, then you have groups of "affected" and "unaffected" contacts. Compare them to find the difference(s). Try using "known good" conductive pads on "problem" PCB contacts. – SamGibson Jun 19 '18 at 15:49
  • 1
    This doesn't need a scope. Thoughtful debugging of the detection scheme, along with unpowered resistance measurements should be enough to figure it out. There are short run board vendors that offer ENIG basically by default, which would be better for this kind of thing. Have to say though, a piano with a rubber keypad is not going to be pleasant. – Chris Stratton Jun 19 '18 at 17:48
  • @ChrisStratton Nearly all modern synthesizers use silicone-dome contacts. (Do an image search for "fatar keyboard teardown" -- Fatar is the largest producer of synth keybeds today.) The lever and spring arrangement dominates the mechanical "feel" of the keys. – Russell Borogove Nov 18 '20 at 17:33
  • Given all but fairly cheap keyboard have velocity detection, how would that even work? – Chris Stratton Nov 18 '20 at 19:45

2 Answers2

1

I finally cleaned up the rubber contacts and the pcb with some alcohol and now it works as expected - so it looks like HASL finish is very sensitive to dust - I now understand why carbon ink is used. Thanks very much all for your comments.

Olivier
  • 41
  • 3
1

The velocity sensitivity of carbon pill based midi keyboard switches comes from having 2 pills which land at different times by virtue of the hinge geometry. The velocity is measured by the time interval between each contact point. Aftertouch is provided by a strain gauge - perhaps a piezo crystal. Piano keys are easy 'cos they're a lever. Drum pads / midi pads are harder and require the measurement of the impact. Piezo or similar again.

In answer to the OP, ENIG because it's flat and remains conductive; Higher density pcb feature - as tight as your pcb fab capabilities allow; Replacement conductive carbon pills are available from Aliexpress. Use a silicone adhesive.

nik cooper
  • 11
  • 2