Multiplexing is the way to go. All the individual segments are tied together and then the digit drive determines which digit is illuminated. For 4 digits it only takes 11 lines of drive.

Check out the MAX6954.

Charlieplexing works fine with 7-segment displays but I don't think it would be a good solution for Erik's needs.

Dedicated display driver ICs would reduce I/O pin requirements but can be a bit pricey. Multiplexing is a viable solution. Perhaps a modular design? How about three seperate 4-digit modules running off a 5-pin interface with PWM brightness control? Just connect the 'SER' line on each module to a unique line (RB2, RB3, or RB4) and load the shift registers on the three modules in parallel (a bit banged 3-channel SPI bus of sorts) by retasking the column driver lines for use as 'SER' and 'CLK' lines during a short display blanking interval at the beginning of each column update interrupt cycle.

Anybody want to see software for this design?

Attached untested BoostC demo for 16F88 below. Most of the interrupt driver code just sets up the special 8 byte 'srdata' array which is used for a 3-channel SPI update of the three 74HC595 shift registers in parallel in just 24 instruction cycles at the beginning of the next interrupt.

Originally Posted by Mike, K8LH

I don't think you know as much about Charlieplexing as you would like to think. First you say "it won't work" and now you're saying you can make it work by breaking the Charlieplexing signal rules.

May I suggest you do a little more research before offering more faulty advice?

Kind regards, Mike

Only semi-faulty... you can do inverse parallel charlieplexing with a common anode and common cathode 7-segment display, which will give you 2 displays w/the same number of pins as 1 display w/o charlieplexing, but if you know of a way to do tri-state charlieplexing with 7-segment displays, I'm all ears/eyes (referring to LEDs 5 and 6 in this diagram):



Plus, perhaps I did some research after the 1st post and before the 2nd...

kpatz,

Wiring common cathode and common anode displays back-to-back with one line driving the two 'common' lines is not Charlieplexing.

if you know of a way to do tri-state charlieplexing with 7-segment displays, I'm all ears/eyes

So you're willing to be enlightened if someone brings you the info' but not interested enough to find it yourself? LOL, just kidding...

There are plenty of examples for Charlieplexing 7-segment displays. The basic Charlieplexing signal rules still apply so we use active high NPN column drivers for common anode displays or active low PNP column drivers for common cathode displays. In the case of the common anode 7-segment matrix below, only one pin will ever be driven high to turn on one of the active high column drivers. The other pins are either driven low to sink current and light up associated LEDs in that column or tri-stated (high-z) to leave LEDs turned off.

Charlieplexed matrix wiring is peculiar in that we use a "float" pin to drive LEDs that occupy a position in the matrix where the column and row would use the same pin (illegal col/row combinations RB0/RB0, RB1/RB1, etc.).



The driver for a Charlieplexed 7-segment matrix isn't much more complex then the driver for a traditional multiplexed matrix. We use a ring counter variable (only 1 bit high) as the port output pattern and to detect when we need to set the "float" bit then we turn on or turn off individual outputs by manipulating the TRIS register.

Code:

void interrupt()
{ pir1.TMR2IF = 0;              // clear tmr2 interrupt flag
  data = display[column];       // new column segment data
  trisb = 0xFF;                 // blank the display
  portb = colsel;               // new output pattern (only 1 bit hi)
  if(data & colsel)             // if 'float' pin required
    data.7 = 1;                 // set the 'float' pin (RB7)
  data |= colsel;               // pick up column select bit
  trisb = ~data;                // display new column (digit)
  if(colsel.7)                  // if last column
  { colnbr = 0;                 // reset column number
    colsel = 1;                 // reset column select bit
  }
  else                          // else
  { colnbr++;                   // increment column number
    colsel <<= 1;               // shift column select bit
  }
}