input plug problem (not multiplexing)

(6 posts) (2 voices)

Tags:

No tags yet.

stef007

Hi all,
this weekend I've been experimenting with the 3 input plugs I received over the week. I'm using the inputdemo sketch for testing. It looks like always the same input pin is routed to the jeenode analog in, no matter what channel is selected. Most of the time it looks like input channel 11 (or 10 on the chip, it's zero-based) is selected. When I connect I11 to the ground the logged line shows 16 times a value of around 17. If connect any other input to the ground I get all high values (because of pull-up).
Commenting out the select(i) function call gives the same behavior. Putting a call select(0) before the loop also gives the same result.
So it looks like the mini ATmega is either dead, not-working or not programmed? Oh yes, once I got the same behavior with I13, but for the rest always I11. I suppose that just depends on the default output values of the mini ATmega after boot.
Anyone any suggestions?

Posted 13/06/2010 20:53:57
jcw
I'm still trying to understand what level of clock variation there is in the ATtiny on the Input Plug.
Could you try something? In Ports.cpp, line 559, there are two timing constants:
```
        byte us = bitRead(data, 4 - i) ? 9 : 3;
```
Could you try varying the 9 and the 3 (microsecond) values?
Maybe also the "delayMicroseconds(4);" that is a few lines below, although I expect that to be much less critical.
I'll do some more test with the latest batch of Input Plugs here.
Posted 13/06/2010 22:02:15

stef007

Ok thanks for the tip. I tried several values. Multiplying all by 10 worked for me;-)


    byte us = bitRead(data, 4 - i) ? 90 : 30;

and


    delayMicroseconds(40);

No more time left today. Will test and see later if this keeps working for long.


				Posted 13/06/2010 22:28:33


	
		
			
								
					jcw

				
			
			
				
Whoa, that's a biiiiig difference! Sounds fishy - as if the fuses didn't get set properly, and the ATTiny's clock is still divided by 8! Thanks, I'll check here.



				Posted 13/06/2010 22:35:43
			
		
	
	
		
			
								
					jcw

				
			
			
				
Ok, confirmed. The fuses were wrong. I've implemented a s/w workaround. To be published in a few days on the daily weblog -  (link will work in a few days, added here for future readers).



				Posted 16/06/2010 17:41:08
			
		
	
	
		
			
								
					stef007

				
			
			
				
Ok, great:-) Will check it out in a few days.
Thanks



				Posted 16/06/2010 22:10:27