[RTTY] Bit length errors

[Kok Chen]

On Feb 27, 2008, at 2:52 PM, Mike Mellinger WA0SXV wrote:

> To what extent does this same problem exists in the traditional  
> serial port?
> Is this just a problem of implementation of the particular interface?

Back when men were men and UARTs were UARTs (remember the National  
16550?), you can get quite close to the 45.45 baud value.

The baud rate of the 16550 is defined by 115200/factor.  So, with a  
factor of 2535, you can get 45.44 baud out of it.  You can get even  
more anally precise by changing the crystal that provides the 115200  
Hz time base.

I think some PCI cards still use UARTs, but real UARTs is scarce among  
USB serial port adaptors.  The only USB adapter that I have in the  
shack here with a UART in it is the Belkin F5U103.

But 45.45 has never been a standard in the computing industry -- it is  
not even in the Unix termios table, for example, which has 50, 75,  
110, 134, 150, etc baud.  So, even if a UART handles 45.45 baud,  
higher level software will not.  You might need a special driver.  The  
weird 45.45 baud number stems from the fact that a bit time in RTTY is  
22.0 milliseconds.

In today's world, you can probably homebrew some code for an EZUSB  
chip that "lies" to the system that it supports 300 baud, but instead  
puts out something close to 45.45 baud when 300 baud is selected.

But... why not wait until Doug K4DSP reveals his magic converter that  
takes AFSK from the sound card and regenerates FSK to key the rig?   
The baud rate can be anything you want as long as your software knows  
how to generate an AFSK signal at that baud rate.

Just tell your software to put out AFSK to a sound card, connect the  
regenerator to the sound card's audio output, and the regenerator  
provides both FSK keying signal and PTT for the rig.

I think that this kind of AFSK-to-FSK regenerator is the way to go if  
you want to use FSK.  There is no need for any USB serial port adapter  
whatsoever.

73
Chen, W7AY