You probably want to use an ordinary timer interrupt and sample at perhaps 8 times the baud rate though four times might possibly work (in hardware implementations 16x is common, but I've personally done it at 8x).
Most processors of modern clock speed should be able to fire an interrupt that fast, at least if your code in the ISR is efficient and you don't have much else interrupt-driven running.
You will probably want to run a state machine, advancing one state on each interrupt. Basically you look for the start bit transition and then read each data bit an appropriate number of oversampled cycles later. There are lots of implementations of this for FPGAs in verilog or VHDL floating around the net, translating back to assembly or C should not be difficult.
However, as you can likely reprogram your timers and interrupt modes quickly, there is another option for optimization: You could start out by having an edge sensitive interrupt on the input pin. When that fires, you then program a timer interrupt to generate delays of precisely a bit period (not oversampled), and simply read out the state on each interrupt. When the word is concluded, re-arm the edge interrupt.
But there is a catch with this optimized method: because you are only "paying attention" at the important times, you have to work harder to make sure that the initial transition is a start bit, and that you aren't catching some data word in mid-flight when your code starts up. Probably you want to do this by using your edge interrupt to make sure that the input has been in a quiescent state long enough (long enough since the last interrupt) - if it hasn't, discard what your get and wait for a clean character.
Some googling on "UART state machine" or the like would be informative. There's always the chance you will find reference code for your particular processor, particularly if it has no hardware UART or only one.
This doesn't answer your question, but might make the code a little easier for you to debug. The case statements are really long and may not be the best way to explain what you are doing with your outputs. I make no guarantees that the code is operational (I have not run it at all), but this should get you thinking about file size and readability.
Your singleminutes case statement has a truth table like this:
// | out
// in| 0 1 2 3 4
// ---------------
// 0 | 0 0 0 0 0
// 1 | 0 1 0 0 0
// 2 | 0 1 1 0 0
// 3 | 0 1 1 1 0
// 4 | 0 1 1 1 1
which might be better represented with output-centric code like this:
if (singleminutes >= 1)
PPEins = 1;
else
PPEins = 0;
if (singleminutes >= 2)
PPZwei = 1;
else
PPZwei = 0;
if (singleminutes >= 3)
PPDrei = 1;
else
PPDrei = 0;
if (singleminutes >= 4)
PPVier = 1;
else
PPVier = 0;
The nfminutes is a little more complicated, but here is the Truth Table:
// | MHUhr PMFuenf PMZehn PMViertel PMZwanzig PMVor PMNach PMHalb | |
// --|--------------------------------------------------------------|--------|-----
// 0 | 1 0 0 0 0 0 0 0 | 1000 0 | 000
// 1 | 0 1 0 0 0 0 1 0 | 0100 0 | 010
// 2 | 0 0 1 0 0 0 1 0 | 0010 0 | 010
// 3 | 0 0 0 1 0 0 1 0 | 0001 0 | 010
// 4 | 0 0 0 0 1 0 1 0 | 0000 1 | 010
// 5 | 0 1 0 0 0 1 0 1 | 0000 0 | 101
// 6 | 0 0 0 0 0 0 0 1 | 0000 0 | 001
// 7 | 0 1 0 0 0 0 1 1 | 0100 0 | 011
// 8 | 0 0 0 0 1 1 0 0 | 0000 1 | 100
// 9 | 0 0 0 1 0 1 0 0 | 0001 0 | 100
//10 | 0 0 1 0 0 1 0 0 | 0010 0 | 100
//11 | 0 1 0 0 0 1 0 0 | 0100 0 | 100
and again some output-centric code:
// MHUhr PMFuenf PMZehn PMViertel PMZwanzig
if( nfminutes == 0 )
MHUhr = 1;
else
MHUhr = 0;
if(( nfminutes == 1 ) || (nfminutes == 5) || (nfminutes == 7) || (nfminutes == 11))
PMFuenf = 1;
else
PMFuenf = 0;
if(( nfminutes == 2 ) || (nfminutes == 10) )
PMZehn = 1;
else
PMZehn = 0;
if(( nfminutes == 3 ) || (nfminutes == 9) )
PMViertel = 1;
else
PMViertel = 0;
if(( nfminutes == 4 ) || (nfminutes == 8) )
PMZwanzig = 1;
else
PMZwanzig = 0;
// PMVor PMNach PMHalb
if( ((nfminutes >= 1 ) && (nfminutes <= 4 )) || (nfminutes == 7))
PMNach = 1;
else
PMNach = 0;
if( (nfminutes >= 5) && (nfminutes <= 7 )
PMHalb = 1;
else
PMHalb = 0;
if(nfminutes >=8)
PMVor = 1;
else
PMVor = 0;
The code above might do well with some #defines too
#define UHR 0
#define PHUENF_NACH 1
#define ZEHN_NACH 2
...
if(nfminutes == UHR)
Again for hours. Truth Table:
| 12 1 2 3 4 5 6 7 8 9 10 11
//----|------------------------------------
// 0 | 1 0 0 0 0 0 0 0 0 0 0 0
// 1 | 0 1 0 0 0 0 0 0 0 0 0 0
// 2 | 0 0 1 0 0 0 0 0 0 0 0 0
// 3 | 0 0 0 1 0 0 0 0 0 0 0 0
// 4 | 0 0 0 0 1 0 0 0 0 0 0 0
// 5 | 0 0 0 0 0 1 0 0 0 0 0 0
// 6 | 0 0 0 0 0 0 1 0 0 0 0 0
// 7 | 0 0 0 0 0 0 0 1 0 0 0 0
// 8 | 0 0 0 0 0 0 0 0 1 0 0 0
// 9 | 0 0 0 0 0 0 0 0 0 1 0 0
// 10 | 0 0 0 0 0 0 0 0 0 0 1 0
// 11 | 0 0 0 0 0 0 0 0 0 0 0 1
and code. Slightly different structure with all outputs being cleared, then only the correct output turned on.
// one-hot, clear all will not cause a glitch
PHZwoelf = 0;
PHEins = 0;
PHZwei = 0;
PHDrei = 0;
PHVier = 0;
PHFuenf = 0;
PHSechs = 0;
PHSieben = 0;
PHAcht = 0;
PHNeun = 0;
PHZehn = 0;
PHElf = 0;
if( hours == 0 )
PHZwoelf = 1;
if( hours == 1 )
PHEins = 1;
if( hours == 2 )
PHZwei = 1;
if( hours == 3 )
PHDrei = 1;
if( hours == 4 )
PHVier = 1;
if( hours == 5 )
PHFuenf = 1;
if( hours == 6 )
PHSechs = 1;
if( hours == 7 )
PHSieben = 1;
if( hours == 8 )
PHAcht = 1;
if( hours == 9 )
PHNeun = 1;
if( hours == 10 )
PHZehn = 1;
if( hours == 11 )
PHElf = 1;
All this also allows you to do your input calculations together before your case statements.
// update single minutes
int singleminutes = (int) (unbcd(tm.min)%5); // 1, 2, 3, 4
// update 5 minutes
int nfminutes = (int) (unbcd(tm.min)/5); // Fuenf Nach, Zehn Nach, ...
// update hours
int hours = (int) (unbcd(tm.hour)%12); // 12, 1, 2, 3, 4...
if(nfminutes>=5) hours++; // 7:25 = Fuenf Vor Halb Acht (8)
Best Answer
Ok, I solved it (kind of a rookie mistake) - I had to adjust the configuration of the UART interface