Scott Dattalo wrote:

> No, I said 4 cycles while reading from RAM. The ROM part your thinking about is
> that the loop is unrolled such that indirect addressing can be avoided.
> Consequently, the code looks rather large. Also, I notice that the loop you show
> above has no means of exiting. If you can live with this, then it's possible to
> read the values from RAM and update the samples once every 3 cycles. (It's easy
> enough to break out of the loop if you have the interrupt-on-change interrupt
> enabled. When the interrupt is processed, you exit with a goto instead of a
> return. Take a second look at what I (and also Nikolai) posted the other day.
>
> Scott

Hi Scott,

I never though about a frequency generator on the PIC's clock - Cool
idea.

I can vary the frequency by using a scaler value 7 - 255. It's a bit
rough, but I'm not building something for Hewlett Packard :-)

How do you read from RAM in 3 cycles? I thought 4 minimum to keep data
evenly spaced.

Loop
  movf RAM1,W
  movwf PORTB
  goto $ + 1
  movf RAM2,W
  movwf PORTB
  goto Loop

I actually implemented an IRQ on USART receive so that I can stop the
wave when the PC sends a char.


--
Best regards

Tony

http://www.picnpoke.com
mailto:sales@picnpoke.com