[Laser] serial port PDM and PWM

[TWOSIG at aol.com]

Oooooooops!   After further study, I found that the standards for  UARTs, and 
therefore serial ports, is to shift out the bits in order from least  
significant to most significant.  I have corrected the bit patterns at the  end of 
this message.
 
I am pleased that my topic sparked a lot of discussion, but I am very  
surprised that it was about laser data links and not what I expected.  It  was even 
suggested that using a serial port to create analog signals modeled on  sound 
card HF modes was highly inefficient, to say the least.
 
I was hoping to communicate that there are ways to create analog  
transmission signals, be they modeled on sound card modes or not, on computers  that do 
not contain sound cards.  Equally significant is that  microprossesor often 
contain UART functionality that can be applied.  These  ideas could be used on a 
computers with sound cards, but I imagined very limited  appeal for serial 
port PDM ( spPDM ), mostly because my limited creativity would  have approached 
that situation with a hardware interface as needed.
 
On newer computers, particularly low end laptops, may not include the  serial 
port that fits what I was trying to describe.  Those have USB and /  or 
variants of the 1394 high speed serial bus.  So much is shared on those  that I 
presumed they would be burst mode and unsuitable for direct application  of 
spPDM, but an external interface box might be worth consideration.
 
Suppose that you built a circuit with a USB receiver that can feed data to  a 
cluster of as many as eight UARTs, each with its own FIFO buffer.  You  might 
need only one for the communication transmission channel, and that may  
benefit from much higher bit rates than I previously described (115,500 bits per  
second ).  The others could be used to drive servo motors to search for  
signals or track moving targets, much like the popular "go to"  telescopes.  Servo 
conrollers may need no more than 300 bits per second,  but need at least two 
axes, and might use differential contol paths.  
 
OK.  Maybe its a small niche.  There are already analog  controlers to do the 
same thing, particularly if you find the sound card  adequate for what you 
are doing.  Maybe I have created a solution for  problem that does not really 
exist.  I still had fun thinking about how to  make it work.
 
 
 
 
In thinking about spPDM, used 8N1 configuration.  It results in nine  
discrete states.  Perhaps there is a better mapping of the input data if  you use 
7N1, which only provides eight states, but that matches three data  bits.  This 
is coarse granularity, but should have some  usefulness, particularly if 
moderate speed is used.
 
I did notice that if you cascade, or "stack" the transmission of multiple  
bytes into one sample, it is better to use the 8N1 configuration.  Two  bytes 
give granularity of 17, three 25, four 33, eight 65.  There is one  more state 
than the number of bits transmitted.  
 
 
Correction of the state outputs on the original post should be shown  below.  
On my email the ASCII printable characters show, but they may not  on the 
post.  
 
PWM:

State 1        X'00'    ASCII  NUL
State 2        X'80'    ASCII  €
State 3        X'C0'    ASCII À
State  4        X'E0'    ASCII à
State 5     X'F0'    ASCII ð
State 6         X'F8'    ASCII ø
State 7        X'FC'   ASCII ü
State 8        X'FE'    ASCII  þ
State 9        X'FF'    ASCII  ÿ


PDM:

State 1        X'00'    ASCII  NUL
State 2        X'08'    ASCII  DLE
State 3        X'24'    ASCII $
State  4        X'49'    ASCII I
State 5     X'55'    ASCII U
State 6        X'B6'    ASCII ¶
State 7        X'DB'    ASCII Û
State  8        X'EF'    ASCII ï
State 9   X'FF'    ASCII ÿ


James
N5GUI