[aprssig] Alinco DR-135T - EJ41U internal modem

[John Hansen]

Scott Miller wrote:

>
> But blaming it on the buffer doesn't make sense either.  KISS frames 
> get passed to the serial port with no flow control (according to the 
> spec anyway), and the only overhead is some framing and escape 
> characters. Converse mode expands callsigns, adds carriage returns, 
> and so forth - it has MORE data to send to the serial port than in 
> KISS mode!
>
It depends on just how poorly designed it is.  Suppose that it never 
envisioned getting two packets in a row.   With continuous reception, 
you have to have enough buffer for two frames, because you have to hold 
on to the first one until you can check the CRC and decide what to do 
with it.  However, while you are checking the CRC and sending it out to 
the terminal, you've got to be receiving the next frame.  So regardless 
of the terminal rate, you need room for two frames.  If you designed it 
only to receive one frame at a time (that is, you could process the 
first frame before you had to worry about getting the next frame) and if 
it was a circular buffer that didn't have any protection against the 
second frame overwriting the first, really bad things could happen.  But 
it would have to be an incredibly poor design.

> Or is the buffer problem only when sending?  


This is more likely the case.  You can't send until the channel is clear 
so you have to have some storage capacity. 

John W2FS