Re: [avr-chat] Flow control with Butterfly

[Pertti Kellomäki]

18. maalis 2005 kello 14:56, Graham Davies wrote:
> If the Butterfly application does not use interrupt-driven input and 
> output
> from and to the UART, this is where I would start.

Yes, that's what I figured as well.

>  You say that you don't
> know how much CPU time is soaked up by the other things going on, so 
> you
> can't reason about whether or not the ATmega169 should be able to keep 
> up
> with the serial data using polled I/O.  The clock rate of the MCU on 
> the
> Butterfly is probably well below the fastest that the ATmega169 can 
> handle
> to conserve power so it may not be going as fast as you think.

If the source code is to be believed, the controller runs at 1MHz.
It seems that this should be plenty, since the data being logged is
generated by a person pressing buttons.

> Once your
> characters are received into a buffer by interrupts you can use 
> XON/XOFF
> flow control to hold off the PC when the buffer shows signs of filling 
> up
> (for example, XOFF when it filles to 75%) and XON when it empties to 
> 25%).

In my case the Butterfly responds to commands given by a PC,
and the commands are never longer than about 50 characters.
The PC waits for the Butterfly to respond, so it is enough if the
Butterfly can handle one bufferful at a time.

> The Butterfly has so little spare I/O that I would not spend a port 
> bit on
> hardware flow control.

In my particular case there is I/O left over, so I could do it,
but it does seem like overkill.

> **** Stop reading here if you object to advertising ****
>
> Working with the AVR Butterfly might be easier with an ECROS Technology
> Butterfly Carrier, which provides power, expansion space, a reset 
> switch, a
> DB9 and ISP/JTAG headers relocated to the rear.  See
> http://www.ecrostech.com.

I wish I had seen this when I started playing with the Butterfly :-(
--
Pertti