AT91SAM9260: How to output PCK0 clock on a GPIO pin

On Mon, 31 Aug 2009 16:44:39 -0400, Stephen Munnings
[off-list ref] wrote:

Pedro I. Sanchez wrote:

quoted

On Mon, 31 Aug 2009 15:58:50 -0400, Stephen Munnings
[off-list ref] wrote:
  

quoted

Sergey Matyukevich wrote:
    

quoted

Hi,
 
  
      

quoted

And how do I turn on the PMC clocks for the GPIO?
    
        

Assigning gpio pin PC6 to TIOB2 peripheral is only a part of the work.
Timer counters may operate in different modes. Take a look at
the datasheet for at91sam9260, timer counters are covered in chapter
34.
      

quoted

quoted

It looks like 'waveform mode' is what you need (see 34.5.10, 34.5.11).
Before you get desired output on PC6 pin you will have to turn on
waveform operating mode for TIOB2 and to configure its wave shapes.
 
Thanks,
Sergey
  
      

hmmmm....

PCK0 is *not*  a TC pin - it is a clock output from the Power

Management

quoted

quoted

Controller.
(Section 25.6)
It should be as easy as dividing down one of the primary clock sources 
and then enabling
the appropriate GPIO pin as the appropriate peripheral.

Ahah - there is the problem.....  PCK0 is not available on PC6 - it

must

quoted

quoted

be set as peripheral B
on pin PC1 or PB30 as Peripheral A - no other choices available...

PC6 is indeed TIOB2 (one of the counter timers)
    

OK, I see this thread unfolding, thank you!

This is what I have now to test pck0 on PC1:

at91_set_B_periph(AT91_PIN_PC1, 0);
pck0=clk_get(NULL, "pck0");
pll=clk_get(NULL, "plla");
clk_set_parent(pck0, pll);
clk_set_rate(pck0, 50000000);
clk_enable(pck0);

But I don't get a clean clock on PC1. In the end I will still need to

get

quoted

the TC clock on PC6 but I for the time being I want to be able to
generate
any clock out first!

Is there something else to do?

  

The code here will try to give you a 50Mhz signal.
That might not be feasible on a 9260.
This might be the source of not getting a "clean" clock on PC1.
Try adjusting it for something that your chip (9260) can reasonably 
generate.
Also, at 50Mhz, there can be all kinds of termination, ringing, and 
other problems.
To get "just any clean signal", try a frequency much lower.
Depending on which pll you use as parent, and what frequency it is 
already running at,
try for around 1Mhz to get a "nice clean signal".
The signal of PCK will try to be a "square wave".
You will find that the PCK signal programming is easier (but much more 
limited) than
programming the TC signals.
The TC signals are much more flexible in their range, etc., but take 
more work to set up.
This is a pretty common trade off.

Well, good news! I got the clean clock on PC1 (3.014 MHz), not at the rate
I requested (1 MHz), but very clean. I'll have to figure out the clock
divider thing now to get the right values.

Anyway, this is a big step for me. Thanks once more!

-- 
Pedro