[En-Nut-Discussion] Back in the game

Thu Feb 15 22:21:52 CET 2007

Hi!

Very strange indeed! 
My reasoning would be: if 2 ethernuts show the same ("dead") behaviour and
am reasonably certain this is not the case, what remains to be suspect is
the SW i.e. the flashing process.
You could try to read the flash memory to see if it really contains the
binary to think should be there. 
Maybe even using another flashing program to upload / download and compare
results...
I running out of ideas...

regards
ernst 

-----Ursprüngliche Nachricht-----
Von: en-nut-discussion-bounces at egnite.de
[mailto:en-nut-discussion-bounces at egnite.de] Im Auftrag von Rich Wellner
Gesendet: Dienstag, 13. Februar 2007 04:43
An: Ethernut User Chat (English)
Betreff: Re: [En-Nut-Discussion] Back in the game

The meter shows .38V on pin 41 running the stock led01 example.

I didn't hook up the LED without a resistor, and I've tried it on two
different nuts, so I'm pretty sure that the port isn't bad.  That said, I
went ahead and reprogrammed for DDRB and get a similar .36V reading on pin
49.

I tried reprogramming to hold the pin high or low per your suggestion, but
showed a .37V in both cases.

My power supply is rated at 9V 1000ma and seems stable.

This just keeps getting more mysterious to me.  I keep going back to the
fact that I can load the httpd demo, but not see any dhcp requests on the
wire and that I don't see anything on the serial port either (having now
convinced myself that I don't need a null modem).  It seems as if the
programmer is happy, but then nothing ever runs and I'm not sure what else
to check.

rw2

Ernst Stippl wrote:
> Hi!
>
> Not knowing how much of an "HW-guy" you are:
> Did you look at http://www.ethernut.de/en/documents/led-hardware.html ?
>
> It shows the importance of the LED's polarity vs. Vcc and the neeed 
> for a resistor.
>
> When you tried to measure the voltage levels, what did the meter show? 
> 0V or
> +5V or ??
>
> I would also modify the program to just issue the
>
> sbi(DDRE, 2);
> and go into and infinite loop afterwards
>
> then modify it to just issue
>  
> cbi(DDRE, 2);
> and go into and infinite loop afterwards
>
> This way, you could measure the voltage in both situations without the 
> time constraints (of the program toggling the io port bit).
> An you could try the LED (plus resistor) in both cases (of course this 
> does only make sense if the measured voltage is more than 2V-3V, i guess.
>
> You could also retry this on another pin after adapting the program,  
> just in case the port at pin 41 got destroyed by using the LED without 
> the resistor ( i do not know how likely this is).
>
> regards
> Ernst
>
> P.S: are you sure the Vcc voltage (i.e. the power supply) you use 
> supplies enough current and is stable enough?
>
> -----Ursprüngliche Nachricht-----
> Von: en-nut-discussion-bounces at egnite.de
> [mailto:en-nut-discussion-bounces at egnite.de] Im Auftrag von Rich 
> Wellner
> Gesendet: Montag, 12. Februar 2007 04:06
> An: Ethernut User Chat (English)
> Betreff: Re: [En-Nut-Discussion] Back in the game
>
> Thanks for the response.
>
> I hooked the LED to VCC and pin 41.  Nadda.
>
> So I hooked a meter to VCC and pin 41 (and every other pin in PE in 
> case I was reading the manual wrong) looking for a change in voltage, 
> but nothing happened.
>
> Since I wrote my other note, I've also gotten out another nut and 
> tried the same process with no luck.  So it appears I'm doing 
> something wrong rather than having a bad board.
>
> rw2
>
> Ernst Stippl wrote:
>   
>> Hi!
>>
>> To which pins on the expansion socket did you connect the LEDs? Are 
>> you sure they match the port (B,D,E) and the respective bit you use 
>> in
>>     
> your program?
>   
>> regards
>> ernst
>>
>> -----Ursprüngliche Nachricht-----
>> Von: en-nut-discussion-bounces at egnite.de
>> [mailto:en-nut-discussion-bounces at egnite.de] Im Auftrag von Rich 
>> Wellner
>> Gesendet: Sonntag, 11. Februar 2007 21:09
>> An: Ethernut User Chat (English)
>> Betreff: Re: [En-Nut-Discussion] Back in the game
>>
>> Ok, I seem to be burning things using uisp.  But nothing ever comes 
>> to life on the nut.
>>
>> Specifically I've tried
>> http://www.ethernut.de/en/documents/led-blink.html and the httpd 
>> sample that came with the dev kit.  Both go through "make burn"
>> apparently successfully, but neither seems to do anything (i.e. no 
>> flashing LED and monitoring the dhcp server shows that no request is 
>> being made for an IP nor does the server respond if I hardcode an IP 
>> and remove the #define to request an IP via dhcp).
>>
>> For example:
>> /dev/ethernut/ethernut-4.2.1/app/httpd$ make burn uisp -dprog=stk500 
>> -dserial=/dev/ttyS0 -dspeed=115200 -dpart=atmega128 --erase --upload 
>> if=httpserv.hex Firmware Version: 1.14 Atmel AVR ATmega128 is found.
>> Firmware Version: 1.14
>> Uploading: flash
>>
>> I tried hooking up to the serial port, but no luck.  I suspect I need 
>> a null modem, but don't have one at the moment.  In any case, given 
>> that the LED program isn't causing an LED flash, I suspect that the 
>> serial line wouldn't tell me anything either.
>>
>> So, I'm looking for advice on how to continue debugging at this point.  
>> Thanks in advance to any who respond.
>>
>> rw2
>>
>> Rich Wellner wrote:
>>   
>>     
>>> After a long absence, I've some time to get back in touch with my 
>>> Rev F boards.  So I went to the site to get a development stack
installed.
>>> I thought I'd share my results.
>>>
>>> This page: http://www.ethernut.de/en/documents/debiansage.html
>>>    has specifically calls out ethernut-3.9.9 when the current 
>>> version is
>>> 4.2.1
>>>
>>> The main doc link: http://www.ethernut.de/pdf/enswm24e.pdf
>>>
>>>    Page 14: In summary it says "download then unpack then configure, 
>>> but before you configure"  The preconfiguration requirements should 
>>> be listed before the instruction to "run ./configure"
>>>
>>>    Page 14: It specified to run ./configure, but never instructs to 
>>> run make && make install.  Unless these commands are run, the next 
>>> step "Configuring Nut" won't work because nutconf hasn't been built yet.
>>>
>>>     Page 16: It specifies to enter the correct top of source 
>>> directory as /opt/ethernut/nut.  As above the nut directory doesn't
>>>       
> exist.
>   
>>>    Page 17: It would be helpful if the location of the sample apps 
>>> were given.
>>>
>>> After this I can build samples.  Tomorrow perhaps I will try to burn 
>>> something.
>>>
>>> Actually, on that point.  How do I determine which burner I have?  I 
>>> have a grey dongle with a yellow LED on it, but no other markings.
>>> Is there a document somewhere with the various gadgets egnite has 
>>> sold over the years specified?
>>>       

_______________________________________________
http://lists.egnite.de/mailman/listinfo/en-nut-discussion

--
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.441 / Virus Database: 268.17.35/680 - Release Date: 10.02.2007
21:15