Saturday, April 25, 2020

Audio Mixer as Opposed to Social Mixer

You'd think with Shelter in place I'd have a lot of time for electronics noodling around, and I do, but I am not sure my heart is fully in it. I must be more social than I thought.



I finished a 4x1 mixer I designed back in 2005, but mixers are so simple that I can't really say "designed".  For whatever reason, I replaced the 2005 era module in my rack with an new version, complete with new PCB and Front Panel Express expensive front panel. 

Whatever.

The original idea goes like this:


Nice schem eh? Um...no. Have I improved any in 15 years?  15 years ago everything was hand drawn and then built on perf. I look back at this work and marvel that it ever worked.  So maybe I've improved a bit, at least I don't duct tape things together nowadays, at least not often.

Anyway I replaced the hand built mixer module with the one you see on the left, the new one looks a bit more pro perhaps?


I dunno I kinda like the one on the right. 

Anyway with all this time I should be building tons of stuff but instead I find myself reading about politicians telling me to drink disinfectant. Pass the Lysol, Elmo! That's the good thing about electronics--unlike people, they work, or they don't, always in the same way, and above all they don't golf.



My psychiatrist girlfriend's college and high school kids are trapped at home, feeling sad and ripped off, and blame us old farts for all of shelter in place, their missed graduation, their boredom, and everything else. Maybe they are right about that. 

But we gave you iPhones!



My mom and dad talk to me via Zoom each week--they have been holed up at their place in the Central Valley for years, so what the hell does Shelter in Place mean to them anyway?  Their lifestyle hasn't changed much in the past few weeks compared to the last five years. But they are old and I worry about them.  



It's oddly quiet and peaceful outside, when I go out to buy groceries or whatever everyone is walking around with masks on, giving each other suspicious looks. I'm coughing because I breathed too many fumes and they look at me like I am contagious, but no, I explain: it's not those kinda fumes. 

 I am not so sure how bad it is out however, in my neighborhood I can hear birds sing better than before, there are less trains rumbling by, and less planes flying over my house all the time. As long as we all stay healthy, and not go crazy, perhaps it's a tiny bit OK.




I get some comfort from the universal language of the schematic, and the idea of things being mixed together kinda reminds me of what we, as humans, aren't doing right now. Wait, no, that's crap. This all appeals to my need for order maybe?  Banishing life's chaos? For sure: it makes me realize how much I love anything with right angles.

 There is something comforting about it, so maybe I'll get back into it today.

Anyway it beats ingesting UV light sources. Whatever you're up to, I hope you are surviving this very strange time with a degree of sanity.  Hang on kids, because the ride is just starting up.





Thursday, April 16, 2020

Turn On the Bubble Machine!

I have no idea what hit me upside the head: for my next audio DIY project in need of a numeric readout, no OLED rubbish, let's go retro. I always liked those little LED bubble displays growing up so how about those?


I looked around and found a bunch on Ebay that were quite expensive, but I managed to spouse up some New Old Stock National Semiconductor NSA1166 seven segment displays--6 bubbles for numbers + decimals--at Jameco, for less than $3 each--cheap.


Those showed up fast!

OK first thing I did was dig into the NSA1166 datasheet, which seems a bit, well, brief?--you can see it here

I knew zilch, and I mean zilch, about 7 segment LED displays, they had never crossed my path before. Wait, where is VCC and GND?  Turns out: doesn't work that way. You get a configuration like the one below (without the current limiting resistors). This is called, for obvious reasons, "Common Cathode"--in this case 8 LED anodes per bubble, but one shared cathode for all of 'em:


That means: current sourced to A1-A8 light up segments of the LEDs in a single bubble, which are laid out like this:


The idea is that an MPU (Arduino for instance--could also be a driver chip, transistors, electro-cups and string, or whatever) sends current to one set of LEDs, that is, one bubble, say LEDs "B" and "C" in the leftmost bubble, to make the number "1", then the MPU moves on to the next bubble--fast. The persistence of vision in your eye make each of the six bubbles look like they are always on even though they aren't. Marvey. Turn on the Bubble Machine!!

OK with that super basic knowledge under my belt, it was time to see if anyone had already written up an easy way to hook 7 segment displays up to Arduino type MPUs. Why reinvent the wheel? 

Answer: yes and yes and yes--this has been done many times before in Arduinoland....

But so far no perfect fit I could find, for me anyway.

The post here is a very cool implementation but appears to work with 3 different Arduinos, none of which I had.  

Or the one here, but it uses HP NOS bubbles which seem expensive and hard to find. 

Or...well....there are tons of other designs out there, but since I have lots of time on my hands, had never touched a seven segment display before, and needed a project to waste time on, I conjured up my best Nixon--I'll erase it myself!

Some design goals: 
  • My implementation must work with a wide range of Arduinos and other Arduino-like MPUs
  • It had to work with the NSA1166 since that's all I had in terms of 7-seg displays
  • The finished design had to be as modular (i.e., reusable in other projects) as possible.
  • It can't flicker, stutter, or looks obvius crap-like
  • It can't show the wrong numbers (duh)
  • A 3 digit number is shown from the far right bubble; and in that case, the left 3 digits must blanked--same idea for any "blank" needed. So no 000333. I want to just see 333.
  • The display had to be reasonably fast to respond to new values at input--10ms?
  • It has to take up as few pins as possible on the Arduino (since I might want to drive this from an ATTINY85 someday, or just have other things I want to use the Arduino for in addition to this display?).
  • OK to use external jellybean parts, but let's keep it reasonable.
OK, let's motorize this pursuit.....

I already had an UNO (easy to develop with) about 3 clono nanos (Ditto) and maybe 5 clono Pro Minis in my junk box. The latter meant finding, untangling, and then hooking up the tiny USB to Serial adapter for programming; 

Whatever, I developed with the trusty UNO. it has enough I/O pins for the bubble display but then I'd be out of pins for anything else fun--so let's use a shift register such as the venerable 74HC595 to free up pins; I used two 595 breakout boards from Sparkfun for my POC:


No Shift, Sherlock! There is a ton of information about using shift registers with Arduinos on Youtube (here and here for instance) as well as the forums etc. (post here was very useful for this project).

I had issues with the fact that pins in 595-land aren't always called the same thing from chip to document to breakout board to forum page...which got me feeling stupid and frustrated from time to time. I'll live.

Here is the "595 Cheat sheet" I came up:

HOW IT WORKS:
  • Bits of data--two bytes total in our case, come in the SERIAL in port.
  • Each HIGH on Serial Clock shifts everything over one to the right. The arduino "shiftOut()" function handles that and the timing. Yeh!
  • Each LATCH going high says "we are done with the bullshift, present everything to the world, ready to use".
  • You can put x number of 595s in series and it all works; I used 2x 595s for this, but you don't have to stop there; you can get as many digital I/O pins as you need by putting more registers in series, flowing your SER bits through all of them, then LATCHing all of it when you're ready to go. And through it all--even if you have 10 of the damn things lined up, that's 80 I/O pins, you'll never consume more than than 3 Arduino digital pins for the whole enchilada.  Obviously this is a very useful IC for what we do.

WHAT THE 595's PINS DO:

  • Q0-Q7 outputs. What we want to use to drive our bubble display's LEDs.  These pins can source or sink current and light up the LEDs (with current limiting resistors); I didn't find anything definitive about the current limits for the NS1166 in the NS datasheet--whatever; 5V with 330ohm-ish resistors did the trick, and 3V worked w 330ohm R's as well albeit a bit dimmer. No smoke!
  • DS-Serial data. Seen it as "SER" on some CMOS data sheets, and called "SER IN" on Sparkfun breakout. Your serial data--the 8 on/offs you want to blow into the shift register--goes in here.
  • OE--output enable. Active Low. Called /OE on the Sparkfun breakout board. We want our outputs enabled eh? Well Yes we do or else nothing happens. So, tie this to ground.
  • SH_CP is clock pin. Each time the clock goes low to high, the bits shift in the device one to the right. Called "CLOCK" on Sparkfun board.
  • MR--master reset. Active low. This clears everything in the register, and we don't want to clear everything ever for this application, so tie it to 5V. Called /RESET on the Sparkfun.
  • ST_CP "Latch" pin.  When that goes high, everything in the shift register is presented to the outside world. Called L_CLOCK on the Sparkfun device.
  • Q7' (Q7 prime) is SER_OUT on the Sparkfun. Use to daisy chain your 595s, so serial going into one IC can feed as many 595s in series as you need for your project. I've seen this called "QH" or "QH*" on some chips and data sheets.

SHUT UP AND BUILD THE THING ALREADY!

OK dammit!

I ended up bread boarding this:



Some bench photos:

I had to modify a cheap breadboard to accommodate the 595's. Sparkfun needed to make these breakout boards about 200mils wider!

I added extra LEDs in parallel with the bubbles to better see what was going on with the 595s. 

Final POC was to have the pot make the display read from 0-1023 based on the pot's wiper position. Works! No flicker!

OK how about the code? Here is what I was using to get the pot to go--I wrote other variations to try out other things, comment if anyone wants to see those.

This wasn't too hard--it took a couple of evenings. I got stuck for maybe an hour on one strange thing--the bubbles wouldn't work reliably if I was using Serial.write to debug. Commenting the serial stuff out, the bubbles worked again. So I was introducing a bug by debugging? A pest, but yes.

A few >'s instead needed to be <' s but overall, not bad at all.


//////////******************************

long testnumber = 0;

//testnumber is what I want to show up on the NSA1166.
//you have to use long. int only goes to 32000 or something close!

//////////******************************


unsigned long startMillis;  //start millis timer
unsigned long currentMillis;
const unsigned long period = 10;

//each item in the array below ground a given pin of a 595, so the Bubble can display.

int mybubble[] = {124,188,220,236,244,248};


int c = 0; /* used to hold number of elements in the array */
long n;
int p = 0; // counter for filling array with elements.

int numberArray[6];

//this holds each value of number we are trying to display
// bubble zero is leftmost bubble!

long bubble5 = 0;
long bubble4 = 0;
long bubble3 = 0;
long bubble2 = 0;
long bubble1 = 0;
long bubble0 = 0;




//pins used to feed shift regs
//Pin connected to ST_CP of 74HC595
int latchPin = 8;
//Pin connected to SH_CP of 74HC595
int clockPin = 12;
////Pin connected to DS of 74HC595
int dataPin = 11;



//pot wiper connected to A0.  gnd AND 5V for the rest
int pot = A0;


void setup() {
              // put your setup code here, to run once:
              //set pins to output so you can control the shift register
              pinMode(latchPin, OUTPUT);
              pinMode(clockPin, OUTPUT);
              pinMode(dataPin, OUTPUT);

// debug using serial
//Serial.begin(9600);  
// BE CAREFUL serial stuff can step on Bubble display and make it mysteriously not work.

// initialize array

             numberArray[0] = 0;
             numberArray[1] = 0;
             numberArray[2] = 0;
             numberArray[3] = 0;
             numberArray[4] = 0;
             numberArray[5] = 0;
 }





            
void loop() 
{

//do a analog read every 10ms

currentMillis = millis();
   if (currentMillis - startMillis >= period)  //test whether the period has elapsed
    {
    testnumber = analogRead(pot);
    startMillis = currentMillis;
    }


c = 0;   


// turn that number into an array.  Note: ARRAY is backwards!


//first we need to determine how many digits the number in ques\ton has.  We need that to send //blanks out later
n = testnumber;

         while (n != 0)
     
           {
            n /= 10;
            c++;

           }
      /*      Serial.println("testnumber variable value is: ");
            Serial.println(testnumber);       
            Serial.println("length of array is: ");
            Serial.println(c);

*/
/////////////////

n = testnumber;
p = 5;
//now fill an array with each digit.

while ((p > -1) && (n != 0))
          
    {

          numberArray[p] = n % 10;
          n /= 10; 
 /*   
          Serial.print("----------");

          Serial.println("We are seeing output of array value");
          Serial.println(p);
          Serial.println(numberArray[p]);
          Serial.println("");
 */     
          
      p--;
    }
 
//wanted to time more CRAP then realized it wasn't needed, comment that!
//Leaving it here in case I remember what I was thinking.

////currentMillis = millis();
//if (currentMillis - startMillis >= period)  //test whether the period has elapsed

//{  

//startMillis = currentMillis;  //update current millis.


/////////////////////////////////////////////////

//YEH next part can be simplified and made into a single function.
//I am too lazy to do this. you do it!

//AND--I don't accommodate decimal points at all.
//Again I am too lazy.....

////////////////////////////////////////////

bubble5 = getbyte(numberArray[5]);        
 //write to shift reg. rightmost digit.

    
    // take the latchPin low so 
    // the LEDs don't change while you're sending in bits:
    digitalWrite(latchPin, LOW);
    

    // next line puts numeric data onto shift register
    shiftOut(dataPin, clockPin, LSBFIRST, bubble5);
   
    
    // then pick which bubble to impact
    
     shiftOut(dataPin, clockPin, LSBFIRST, mybubble[5]);
    //take the latch pin high so the LEDs will light up:
    digitalWrite(latchPin, HIGH);



//write to shift reg digit. 2nd from right 

if (c < 2)

   {
    bubble4 = 0;
   }
   else
   {
    bubble4 = getbyte(numberArray[4]);        
   }

    // take the latchPin low so 
    // the LEDs don't change while you're sending in bits:
    digitalWrite(latchPin, LOW);
    

    // next line puts numeric data onto shift register
    shiftOut(dataPin, clockPin, LSBFIRST, bubble4);
   
    
    // then pick which bubble to impact
    
     shiftOut(dataPin, clockPin, LSBFIRST, mybubble[4]);
    //take the latch pin high so the LEDs will light up:
    digitalWrite(latchPin, HIGH);

    
 //write to shift reg digit. 3rd from right 

if (c < 3)
  {
  bubble3 = 0;  
  }
  else
  {
  bubble3 = getbyte(numberArray[3]);        
  }
    

    // take the latchPin low so 
    // the LEDs don't change while you're sending in bits:
    digitalWrite(latchPin, LOW);
    

    // next line puts numeric data onto shift register
    shiftOut(dataPin, clockPin, LSBFIRST, bubble3);
   
    
    // then pick which bubble to impact
    
     shiftOut(dataPin, clockPin, LSBFIRST, mybubble[3]);
    //take the latch pin high so the LEDs will light up:
    digitalWrite(latchPin, HIGH);



  //write to shift reg digit. 4th from right 


if (c < 4)
  {
    bubble2 = 0;
  }
  else
  {
  bubble2 = getbyte(numberArray[2]);        
  }  
    // take the latchPin low so 
    // the LEDs don't change while you're sending in bits:
    digitalWrite(latchPin, LOW);
    

    // next line puts numeric data onto shift register
    shiftOut(dataPin, clockPin, LSBFIRST, bubble2);
   
    
    // then pick which bubble to impact
    
     shiftOut(dataPin, clockPin, LSBFIRST, mybubble[2]);
    //take the latch pin high so the LEDs will light up:
    digitalWrite(latchPin, HIGH);       



  //write to shift reg digit. 5th from right 


if (c < 5)
  {
    bubble1 = 0;
   }
else
{
  bubble1 = getbyte(numberArray[1]);        
}   
    // take the latchPin low so 
    // the LEDs don't change while you're sending in bits:
    digitalWrite(latchPin, LOW);
    

    // next line puts numeric data onto shift register
    shiftOut(dataPin, clockPin, LSBFIRST, bubble1);
   
    
    // then pick which bubble to impact
    
     shiftOut(dataPin, clockPin, LSBFIRST, mybubble[1]);
    //take the latch pin high so the LEDs will light up:
    digitalWrite(latchPin, HIGH);       


  //write to shift reg digit. leftmost


if (c < 6)
  {
  bubble0 = 0;
  }
  else
  {
  bubble0 = getbyte(numberArray[0]);        
  } 
    // take the latchPin low so 
    // the LEDs don't change while you're sending in bits:
    digitalWrite(latchPin, LOW);
    

    // next line puts numeric data onto shift register
    shiftOut(dataPin, clockPin, LSBFIRST, bubble0);
   
    
    // then pick which bubble to impact
    
     shiftOut(dataPin, clockPin, LSBFIRST, mybubble[0]);
    //take the latch pin high so the LEDs will light up:
    digitalWrite(latchPin, HIGH);       


  
   
} // end main loop

//shiftOUT does not seem to accommodate bxxxxxxx typle format?
//whatever, I am using decimal values for this.

byte getbyte(long x)
{
  int a;
  
 if (x == 0)
   {
   a = 252; //11111100
   return a;
   }
  
if (x == 1)
   {
  a =  96; //011000000
  return a;
   }

if (x == 2)
   {
  a =  218; //11011010
  return a;
   }
if (x == 3)
   {
  a =  242; //11110010
  return a;
   }

if (x == 4)
   {
  a =  102; //01100110
  return a;
   }

if (x == 5)
   {
  a =  182; //10110110
  return a;
   }

if (x == 6)
   {
  a = 190 ; //10111110
  return a;
   }

if (x == 7)
   {
  a =  224; //11100000
  return a;
   }

if (x == 8)
   {
  a = 254 ; //11111110
  return a;
   }

if (x == 9)
   {
  a = 230 ; //11100110
  return a;
   }


}

///////////////////////////////////////////////

OK so this all works!  What is next?

It would be cool to use some pins on the Arduino (Something w/ a smaller form factor than an uno) to make this into an I2C slave device. Then, I could drive the NSA1166 with I2C and never have to think very hard about adding bubble displays to my projects again. Update: I2C control of the NSA1166 works. Part II, where I get I2C working with 32 bit datagrams--needed for a 6 digit display, is here, part III, where I get it all working with I2C, is here.

Also, would be nice to get decimals going, but I don't have an immediate need for that, so for now, wah-wah.

That's it, no smoke, no fumes, bubble up the covid and leave it behind!

Sunday, April 12, 2020

Using PCB material for a Front Panel: I said it wouldn't work, and I have no idea why I said this. It works.

Before all this bizarre and tragic COVID hibernation stuff began, I was at a local geeky synth meetup discussing (of course) DIY front panels, going through my usual diatribe about using PCBWAY "alubase" material to provide my inexpensive metalwork. I was pretty damn proud of myself. But then! one of the much smarter members told me that's stupid--and expensive--why not just use PCB material for this? Much cheaper! Much Easier! Mo Better!

I said, well you can't because--because.....because??



Turns out there is no because!! Of course, PCB material works great for front panels.

Yeh, and any PCB fab shop can make these. You already have the board layout right?  Just throw another gerber in there in the same run for the front panel. Viola.

So who came up with this DIY breakthrough? No idea, but the Meetup tech I was talking to was turned onto using PCB material for front panels because LMNC uses it in his VCO kit (oh no!!).

So it must work, right?

All Hail this dude?


Let's motorize this pursuit! Using Eagle CAD, I created a front panel design for a NOISE! board from Reverselandfill that, after many PCB level repairs to fix some of traces I savagely busted, worked again. It desperately needed a front panel--Orig. post about this module is here.

Creating the new raw front panel was easy--I measured the board, created a quickee PCB with just the dimensions and drills (and nothing else), created a gerber, and off it went to JLCPCB along with some other "real" PCB layouts.

You can try adding Gerberized silkscreens for the legends, but for this project I didn't do that. I used Mr. Label instead (blog post about Mr. Label is here--1001 uses?)  Yeh, I get all graphics and fonts etc. I can dream up and are easily printed out on my super cheap laser printer.


Mr Label ready to apply

Say What? JLCPBC didn't process the order right away (it was "Under review" for at least a couple of days) They figured that I forgot to put traces and a soldermask in the gerber and didn't want me to go wah-wah. But I emailed them again: this was for a front panel! And after they pretty quickly gave me the thumbs up. So I paid 'em and waited.

The PCB for "front panel only" showed up in Sunny N. California about 6 days later. After a serious amount of Howard Hughes ready Covid scrub downs I was ready to work.

So what: a few things about this--first, the PCB material is more brittle than aluminum, and if you crack the panel you probably have to start over, so best you can, make sure you get the layout right or close to right the first time.

Second, you might need a large metal file (I did anyway) to file away excess material, for me, it appeared that the panel shipped back to me was maybe 1mm-2mm off on one side. But it was pretty easy to file the PCB stuff down to size.

Third, color issues? since I was working with a black background in Illustrator, I used black PCB material. This turned out to be a good thing, because even though the Mr. Label didn't quite fit  (I put his jock strap on crooked?) the colors matched and everything looked good. I also found that black sharpee matches the flat black color of JLCPBCs black panels pretty well. I have used that in the past to cover up minor mistakes.

Fourth, you can move around the holes if you want! If you have a preexisting PCB front panel that's not quite right try this: put some tape on the back of the hole you want moved.  Then: fill the hole with JBweld. JBW sticks to PCB material like...glue? Next: after the jbweld cures, file things do so the hole is filled but flush. Drill out new holes if needed--you can drill small holes and then go at them with a hand held T handled reamer, a great tool for enlarging holes in PCB material. Finally: Paint or cover.

Right: original PCB.  Left: new holes for pots.

Here's an early attempt w/ PCB and Mr Label:



While I was at it, I created a better Mr. Label for one of the modules from my CEM VCO project, to replace the one that burned up rel good:


Look Mum, No Covid?  Damn I hope so. I will be using a lot more PCB material for front panels in the near future. If there's a reason to get back to aluminum I'll do that, but for now, PCB works and I'm happily taking in the fumes. You should do the same, but not the Corona Kind, rather, the good kind. Stay safe!

JTAG to SWD Converter

Readers: If you'd like to build the project featured in today's post, please go to PCBWAY's Community pages--gerber file, KiCAD ...