Order Tray | Contact Us | Home | SIG Lists

[aprssig] Suggestions for modem designs?

Jason Winningham jdw at eng.uah.edu
Fri Oct 13 14:01:08 UTC 2006

On Oct 13, 2006, at 7:45 AM, Rich Mulvey wrote:

> Can anyone suggest particualr implementations that work well?  They  
> don't have to be 614 based - they just seem to be the most common  
> ones I've run across.  I want something that handles both TX and  
> RX, as well.

Google for variations of "TNC ax.25 circuit" and you'll eventually  
dig out lots.

The tracker2 is open source; find it a source forge.  It uses a 2211  

The WhereAVR is a zero crossing detector; meant for tracking and  
light duty receiving.  http://www.knology.net/~gdion/whereavr.html

I've built a couple of these receiver/decoders; they tend to choke on  
long packets but otherwise work well.  It seems a shame to tie up  
several $ worth of parts on something that can't transmit, though:


There are plans in the March/April 2005 QEX for a "modemless"  TNC  
(zero crossing detector) based on a PIC.

There are various "software modems" out there that are nothing more  
than a hardware encoder/decoder tied to parallel or rs232 lines and  
the rest of the work is done in software.

You've got several choices for the modem:  The MX614 you've already  
found.  It does a decent job, requires its own crystal, and is  
relatively expensive ($10 range).  There may be other chips out there  
that do Bell 202 tones (which are what we use for 1200 baud packet),  
but I don't know about them.

There is a pair of chips, the 2206 and 2211, than respectively  
generate and decode tones.  The tones are set by external components,  
so you can decode 300 baud packet, 1200 baud packet, etc.  The  
devices are relatively cheap (compared to the 614, anyway).  The bad  
thing is, the parts count is a bit high.  Here's one example:


For output, there are schemes for digital to analog conversion based  
on either resistor-ladder networks (usually around 4 bits, I think)  
or PWM to generate the tones.  For decoding there are methods based  
on feeding the analog signal directly to a microcontroller's ADC,  
usually with a bit of filtering and/or amplification first.

A dedicated DSP device could be used, but I haven't seen an example  
of that.


More information about the aprssig mailing list