System does not power down [solved]

[Cybergei]

Shutdown script question here.  Your script is checking for the value of /sys/class/gpio/gpio$SHUTDOWN/value which is set to 0, far as I can tell, always.

Then it is testing for a change to anything other than 0.  How does pushing your power button change it from 0 to anything?  Far as I can tell, the script issues a shutdown command if the value is != 0 but there is never a change.  Is there a line of code missing?

[Felix]

ATXRaspi will pull a pin UP when the shutdown condition is met (the button held for 3sec and BOOTOK signal from PI==HIGH). That will read HIGH (or 1) on the Pi side and the script will detect that and start a sudo poweroff.
The default pin assignments are:
SHUTDOWN=7
BOOT=8

To verify you have a good ATXRaspi please completely disconnect it from your Pi. You can test it by itself. Just turn it ON using the button as normally. At that point the output side will have 5V. Simulate a Pi boot OK by connecting the BOOTOK to the 5V output (with a jumper wire) - the LED next to BOOTOK will turn ON. Now hold the button for 3 seconds. That should assert the SHUTDOWN signal and turn the LED next to it, also the LED connected (if any)  to your button will start pulsing. Give it a few seconds and then disconnect the BOOTOK signal from 5V. At that point the BOOTOK LED turns off and the pulsing will be faster on the button LED. After another few seconds power will be cut off.

[Cybergei]

Great post on testing the ATX Raspi.  After connecting 5v to BOOTOK the green LED does come on.  While the Green LED is on, I have held the button for 3-4 seconds until the LED starts pulsing.  after around 2 seconds I disconnected the BOOTOK.  No other LED ever comes on and the Green LED goes off.  The poser LED does pulse for a few seconds and the unit does cut power after a few more seconds.  But no other LED other than that Green one ever comes on.

What ya think?

[Felix]

That would indicate the SHUTDOWN signal does not get turned HIGH. Sounds like a board with some missing/bad functionality.
They do all get tested with that same sequence I've described but the human factor, shipping, and user handling can still introduce an error or perhaps zapping the chip (very unlikely but not unheard of). To put this all aside, please contact me directly for getting a replacement out to you.

[Cybergei]

The error in the board is no problem.  What matters most is the support after the sale.  You have done a fantastic job.  I will buy another board as the damage certainly could have come from me during soldering or anything else.  I appreciate your focus with finding this issue with me.

[Felix]

Thanks, I guess I'm glad you can say that after it taking so long. It's partly my fault assuming the board was good the whole time, but it appears it is not, that's a bummer. Next time I will make sure to have the user test the board before chasing rabbits in wonderland. Maybe because I've seen this same type of issue so many times before where the script or wiring was not done as expected, or some OS access rights etc.

[Cybergei]

Nah, come on.  It is trouble shooting process.  You are faced with dudes that are beginners and have no idea what step one is.  It is hard to provide support via anything remote. You did absolutely fine, and that is what adds worth to your project.  I sincerely appreciate your attention and thought that you gifted to me.

It is kinda like when you are providing computer support.  You know that you have to ask if the plugged the computer in.  You do not want to ask, but if you don't, that will be the problem.  It is a process.  Best part is that now the entire process has been exposed for everyone to read about.  So it simply adds that much more to it.

Anyone says otherwise, they have not ever developed a product and put it on the market before.

[Felix]

You are spot on. I'm still learning how to best respond to user issues and be most efficient for both parties. The forum has really helped reduce my support time and having to repeat the same answer over and over. So thanks for your patience.
Unfortunately I'm not a big company to afford a direct line where someone can pick up the phone and walk through everything live. But I'm trying my best to help people and provide whatever support I can for my few products.

[Cybergei]

Thanks guy.  I will order soon.  I have a bit more work to do on my main unit but I think I will need at least 3.  I am building prototypes right now using embedded units so, a power on/off is necessary.

[Felix]

Ok just FYI in case you wanted to get some in the next week - I sold quite a few this past week and I'm down to a handful of the last batch right now. I have more PCBs on order but they will take another week or so to get here then some more time until I get to assemble and test them etc.

[Cybergei]

Thanks for letting me know. There is no hurry

[Cybergei]

UPDATE:  >Got it all working now.  I have hard wired my ATXRaspi to the P5 of my Pi using similar tools and connections to the original video on the LL website.  I am also using the ferrite cylinder along with a fuse inserted in to the board.  For the wiring diagram for the P5 header which I am using, most of the ones that I found were incorrect.  The correct one can be found at http://www.raspberrypi-spy.co.uk/2012/09/raspberry-pi-p5-header/  Also important is that I used  SHUTDOWN=28 and BOOT=29 in my shutdowncheck.sh script.  Make sure that you wire accordingly using the given diagram.  If anyone wants more details, feel free to ask.

[Felix]

Thanks for the update, much appreciated.