Has anyone used the Adafruit MPR121 Touch Sensor Breakout with the Propeller?

Jim_Arlow

Hi All,

Has anyone got any code to drive the Adafruit MPR121 Touch Sensor Breakout? It is an I2C device.

Thanks!

Jim

Mike Green

It’s only a general I2C interface, but “Basic I2C Driver” in the OBEX would let you access the device’s registers. You can get the datasheet from AdaFruit. I don’t know if they also provide the sources for their Arduino library. You also might look at their CircuitPython library for ideas.

There are other I2C interfaces in the OBEX. This one is pretty straightforward to use.

JonnyMac

I don't have one of those devices but as Mike points out, it's a pretty simple interface with I2C. Adafruit does provide a simple Arduino library for the part and there was no problem porting that to Spin.

FAIR WARNING: the attached code compiles but I haven't tested it as I don't have the board.

JonnyMac

Whoops... while looking through the code I found a missing CR. Fixed in this version -- but I still haven't run the code.

Jim_Arlow

Mike and Jon - thanks for your help! Jon - that jm_mpr121.spin code looks like just what I need. I did look at the Adafruit library some time ago, and I can recognize bits of it even though I am new to Spin. I should have no problem getting it going. BTW - very, very nice code! Easy to follow.

JonnyMac

Thanks for the kind words, and I do hope the code helps.

Looking through it again I found a copy-and-paste gotcha. The rd_reg8() method should use i2c#NAK instead of i2c#ACK in the read line -- like this:

pub rd_reg8(reg)

'' Read 8-bit value from specified register

  i2c.start
  i2c.write(devid)
  i2c.write(reg)
  i2c.start
  i2c.write(devid | 1)
  result.byte[0] := i2c.read(i2c#NAK)
  i2c.stop

I will order one of those boards so that I can test the code. Once I've done that I will upload the final object/demo to ObEx. In the meantime, do let me know if you have any troubles or suggestions.

Jim_Arlow

Thanks Jon - that's fantastic! I think many people will find the board useful. I use them quite a lot with MIDI instruments, but I have only used them with the Arduino, and not yet with the Propeller. I am working on a MIDI Autoharp that has 12 carbon fibre strings, and a 24 button chord keyboard. My existing Arduino design uses an Uno to connected to the MPR121 to scan the strings and output MIDI, and a separately mounted ATMega328p to scan the keyboard. It is a clumsy design, so I want to move it to the Propeller where I can have separate cores handling the different tasks. One for the strings, one for the keyboard and one for the MIDI output. You never know, I might even add video output to it... There are so many more possibilities with the Propeller.

JonnyMac

That sounds very cool. I play guitar and recently bought and inexpensive keytar so that I had a standard MIDI output to send into a Propeller board (I use the EFX-TEK HC-8+, but then, I designed it!).

Since the MPR121 is I2C you could use a couple MCP23017s (also I2C) to monitor your buttons -- assuming, of course, that they are simple closures and not velocity sensitive. Depending on scan rate you might want to switch to a PASM I2C interface. If you decided to do that, let me know. I would have to update (well, create a secondary version of) my MPR121 and MCP23017 objects to use an external I2C object so that you're only consuming one cog and not creating issues with collisions.

I hope you'll post the project when you get it working.

Jim_Arlow

Keytars are great fun! The missing link between the guitar and the keyboard.

Yes, the buttons are just simple closures, so I could use MCP23017s. I'll consider it in the future. The thing is I am trying to keep latency down to a minimum so I want one core monitoring the strings and a different one monitoring the keyboard. The interface between the ATMega328p keyboard scanner and the Uno is 5 bit parallel and one interrupt, so that will be fast. All the chord keyboard does is output a binary number that corresponds to a specific chord.

In fact, I should do some investigation, because I know that latency can be a serious problem but I don't really know the extent of the problem, or where the latency is. It seems to be that if the Uno tries to do much else apart from scan the strings using the MPR121, and send some MIDI, everything gets too unresponsive. Throwing another ATMega328p at the problem is a lazy solution, but it will work.

Actually, I just got my chord keyboard PCBs this morning, so I'll be able to see how it works with the ATMega328p shortly.

EFX-TEK HC-8+ - I took a look at your site - you've got some good looking stuff there.

JonnyMac

The nice thing about the Propeller is that you can throw another processor at the Propeller without worry of processor-to-processor coms. Scanning 24 buttons could be handled with 10 pins (4 rows of 6 columns) -- you might even do it in Spin like this:

var

  long  kcog
  long  kstack[32]

  long  keys                                                    ' global keys mask


pri scan_keyboard | t, tkeys, scan

'' Scan 4x6 matrix keyboard every millisecond

  outa[ROW3..ROW0] := %1111                                     ' prep rows for high output

  t := cnt
  repeat
    tkeys := -1                                                 ' enable all keys
    repeat 4
      waitcnt(t += constant(MS_001 >> 2))                       ' wait 1/4ms
      dira[ROW3..ROW0] := %0001                                 ' scan row-by-row
      scan := ina[COL5..COL0]        
      dira[ROW3..ROW0] := %0010      
      scan |= ina[COL5..COL0] <<  6  
      dira[ROW3..ROW0] := %0100      
      scan |= ina[COL5..COL0] << 12  
      dira[ROW3..ROW0] := %1000      
      scan |= ina[COL5..COL0] << 18
      dira[ROW3..ROW0] := %0000  
      tkeys &= scan                                             ' debounce
    keys := tkeys                                               ' update global keys

You would launch this method into its own cog with:

  kcog := cognew(scan_keyboard, @kstack) + 1

Since this is in Spin and you have control of the design, it uses contigous groups for the rows and columns -- this helps keep the code efficient. The scan code is unrolled to help it be as fast as possible, too.

Fair warning: I've only been up for about an hour am just on one cup of coffee!

Again, this sounds like such a fun project -- I'm looking forward to seeing (and hearing) the end result.

Jim_Arlow

Yes - that code is pretty much what I have on the ATMega328p! In fact, what I originally wanted was a couple of 74922 keypad driver ICs, but these are now obsolete, so I programmed the ATMega chip to kind of emulate them. Shame - they were nice chips. I actually have 3 rows x 8 columns (not that it matters), after the Oscar Schmidt "Americana" Autoharp keyboard (http://www.oscarschmidt.com/products/autoharps/os11021ae.asp). There are 7 chords in each row, so that leaves me with 1 user programmable key per row. I might use 2 of these for octave up and down, and the 3rd to switch from Autoharp mode (chordal) to lyre mode (diatonic).

With regards to seeing the end result, I have an article in the next edition of Nuts & Volts on a MIDI Lyre, which is the basis of the MIDI Autoharp. I'm planning on 2 versions of the MIDI Autoharp. An Arduino version, that will be an add-on to the MIDI Lyre - this will be published in Nuts & Volts also, and a Propeller version. I can post the Propeller version here. With regards to how it sounds, the main thing design point is that it is a very responsive instrument. So you can stum the strings quite rapidly. This is why I am keeping the string scanning loop as tight as possible.

Publison

Jim,

This sounds like a pretty cool project! I'm a MIDI guy from the 80's. Look forward to N&V articles.

Publison

JonnyMac wrote: »

That sounds very cool. I play guitar and recently bought and inexpensive keytar so that I had a standard MIDI output to send into a Propeller board (I use the EFX-TEK HC-8+, but then, I designed it!).

Jon, Did you get one of the new Alexsis Wireless Keyboards?
http://www.alesis.com/products/view2/vortex-wireless-2
I put one on my cart to play with. I sold all of my Rolands and Casios a few years back. Still kept my Moog Liberation, but that's not MIDI.

JonnyMac

Jon, Did you get one of the new Alexsis Wireless Keyboards?

No, no, no. When I said inexpensive, I meant it.
-- https://www.amazon.com/gp/product/B003RS19XE/ref=oh_aui_detailpage_o04_s00?ie=UTF8&th=1
This was recommended by my friend Ryan Clarke (for a dirt-cheap MIDI controller). I wanted a standard MIDI out port which is really hard to find on budget instruments.
Ryan has a neat channel on Twitch that concerns itself with electronics and electronic music.
-- https://www.twitch.tv/1o57_curiouscodes/videos/all

Publison

JonnyMac wrote: »

Jon, Did you get one of the new Alexsis Wireless Keyboards?

No, no, no. When I said inexpensive, I meant it.
-- https://www.amazon.com/gp/product/B003RS19XE/ref=oh_aui_detailpage_o04_s00?ie=UTF8&th=1
This was recommended by my friend Ryan Clarke (for a dirt-cheap MIDI controller). I wanted a standard MIDI out port which is really hard to find on budget instruments.
Ryan has a neat channel on Twitch that concerns itself with electronics and electronic music.
-- https://www.twitch.tv/1o57_curiouscodes/videos/all

WOW! That is cheap for a MIDI!

JonnyMac

I really expected it to be Smile, but it feels quite substantial. That said, I haven't powered it up (on the road) so I will save final judgement for later.

Jeff Haas

Erco deal alert!

I got an ebay clone of a retired Sparkfun MPR121 touchpad - this has the chip conveniently set up on a board with a touch keypad:
https://www.sparkfun.com/products/retired/10250

And tried it out with Jon's code. It worked the first time!

Jim_Arlow

Jeff - fantastic! I haven't had a chance to try it out yet myself but look forward to doing so as soon as I have finished the Arduino version.

JonnyMac

And tried it out with Jon's code. It worked the first time!

I rule!