I am just starting to use the Edison with the development board ("Intel Edison Kit for Arduino Antenna"). I got the "Blink" test program to work with the USB connected to the PC. Now I would like to be able to run the board by itself (without the USB connections). I'm using a 11V DC, 800mA adapter. I hear that the board can draw more than that, but only when communicating through WI-FI.
The problem is that as soon as I disconnect the USB, the program stops (the DS2 LED stops blinking). When I press reset, DS2 turns on momentarily, then dims off and only DS1 remains on.I tried several combinations of pressing the PWR button, disconnecting the supply and re-connecting, soft-disconnecting the USB "nicely" from the taskbar before physically disconnecting them. Sometimes I get different results like DS2 blinking rapidly (not in accordance to the Blink program), but I never got it to run the program correctly without the USB connected. If I reconnect the USB and re-upload, it works fine, until I disconnect the USB.
This should be very basic, so it is frustrating...
Am I missing something simple?
Do I need to use a beefier adapter?
I left it connected to the adapter for 5min, and the Edison SOC got burning hot! Also the DS1 LED was dimmed. I turned it off right away. I reconnected the USB cables, uploaded the Blink program and now it's been running for 5min and does not heat up. It is warm, but not burning hot. The only explanation I can find is that since the adapter cannot supply enough curent, the voltage drops (I will measure it to see if it drops), and the SOC goes into a weird, fast looping state that makes it draw a lot of current and overheat. Is that possible?
Thanks for your help in advance!
That's a known issue that has been reported before. I recommend you to follow the instructions in this thread: https://communities.intel.com/thread/77945 https://communities.intel.com/thread/77945. The thread describes a workaround to make the Arduino sketches run automatically after a power off.
Regarding to the overheating issue, have you tried using another DC wall adapter? The issue might be caused by the DC wall adapter, or by the internal voltage regulator. Please try using another DC wall adapter and let us know if the issue is still happening.
Not yet. Tomorrow I'm planning to fix the issue with OS not booting up first, then I will look at the power supply issue.
Thanks for following up. I've been wanting to fix this sooner, but I also realized that one of the two micro-USB cables I was using was just a "power" USB cable, not a "data" cable, so I could not access the board through the serial interface. Now I found a second good cable, so I will do that tomorrow.
I loaded the older Yocto image and now it powers up with a 9V battery and USB cables disconnected (yeah!). I mounted the thumbdrive and copied the files, but I did not have time to re-flash it with the latest image and copy the files. But the fact that it boots with the battery confirms that the issue is the image, like reported by many other people.
I tried the external adapters again. With the small one (200mA), it cannot boot. The LEDs keep blinking, like if it keeps looping half way through the booting sequence.
With the bigger adapter (800mA), it is still not booting and overheating. I have not checked the supply current when running with the 9V battery, but the battery gets hot, so I guess it is drawing quite a bit of current. I'm still thinking that the 800mA is not enough and the CPU goes into a fast loop. I will bring it home and power it with an adjustable power supply set to 9V and I will monitor the current.
Looks like I will need to order a 9V/2A adapter.
I would suggest you to try using another power supply. In my case, I use a power supply rated at 12V/1.25A and so far I haven't had issues with it.
200mA is too low, however, 800mA should be enough to boot the Edison. Anyhow, try to use another power supply. It might be that there is a hardware issue in the voltage regulator circuit. It would be great if you can measure the output voltage of the power supply and check if it changes once it is plugged into the barrel jack. If there is a short circuit in the board, the output voltage should decrease and the current draw should increase. That would explain why the 800mA are not enough to power the board, and why the battery gets hot.
I ordered a 11V/2A adapter, so that should take care of it.
Yesterday I measured the supply current while powering from an adjustable 1.5A power supply at 9V. It started at ~800mA for a few seconds, then dropped to 500mA. So that may explain why my 800mA adapter is just barely enough. We know that these cheap adapter are usually not well regulated when pushed to their limits.
But now I have a new problem: The blink program does not work anymore... It is very strange: The Arduino development tool says "transfer complete", but the LED never blinks. The LED is working ok since it does turn on when I reset.
I was able to re-flashed the image with the older Yocto (18-15), I can connect serially to it and send/receive commands. But the blink programs does not work...
As far as I know I never did anything that could damage it physically, so I'm at a loss...
It did overheat with the issue of power adapter, but since I can communicate with it and flash it, I don't understand why it refuses to run the program.
I'm not sure why you are having issues with the Arduino sketch. Please try the following and let me know the results:
I did all that. I reloaded the latest Yocto image and tried the blink example again. Same thing. It says "Transfer complete", but the LED is not blinking. I checked everything like the board selection, serial port setting, etc. I also removed the Base Shield and all sensors/actuators. I removed, inspected and re-connected the Intel Edison on the development board. I tried various combinations of reseting, PWR on/off. Nothing works.
I also tried the "Grove_Indoor_Environment_Demo" sketch described in the small manual (which by the way is not clearly written, full of typos and bad syntax...). Again, it says Transfer Complete, but it does not work. The serial monitor shows information from the sensors, but the values are all wrong. For example the temperature is always stuck at -50, which I guess is the extreme of the low range of the sensor. The LCD is blank and nothing appears when I turn the encoder. It really seems like the Edison is alive and can load programs, but the whole I/Os are dead.
I'm running out of ideas of what to check. This has been a very frustrating experience so far. I spent more time than should have been necessary hunting very basic documentation that should have been provided, either with the kit or centralized in a single website. Now I may have a defective board. This takes the fun out of it.
Do you have another idea of what I should check? If not I will start to see if I can get a replacement.
Thanks for your help,
Could you please share the link of the small manual you are referring to? I'd like to test the Grove_Indoor_Environment_Demo sketch, or any other example just to check what results I get from them.
Meanwhile, just to check if the issue is happening only with the Arduino sketches or not, please try to run a Python script. The following is the Blink example for Python: https://github.com/intel-iot-devkit/mraa/blob/master/examples/python/blink-io8.py
In order to run it, follow these steps:
Anyhow, please post the link of the small manual so I can check it and try to test it.
I did what you suggested. I modified the python script to use pin 13 instead and saved it as "blink.py". I copied it to the Edison flash drive. Then I found how to mount the flash drive and I copied it to /dev/root/python/blink.py.
When I do "python blink.py", the command seems to run with no issues, but nothing happens on the board. The LED does not turn on and does not blink.
When I did "CTRL-C" to exit the python program, it reported something about being interrupted and was at a certain line in the program. So it was running it.
Again, this is like if the processor runs, but the IOs are dead.
By the way, I also noticed that whenever I connect the second USB cable to get the serial link, it disrupts the WIFI connection on my laptop. I know this sounds strange, but it always happens. Soon after I connect the USB, my WIFI disconnects. I can fix it with the troubleshooter, then it disconnects again. Very annoying. Maybe there is a conflict with the drivers and ports?
Regarding the small manual, this is what is provided with the "Grove Indoor Environment Kit for Intel Edison". I searched their website but could not find an electronic version. Their website is not very good either. I could not even find the kit I bought listed in their products. Here is an example of a sentence that is so cryptic that it causes more confusion than it provides help:
"Step 2: A actuator can only be controlled by a sensor. If A sensor wants to control a actuator(has be controlled by B sensor), B sensor should be set sleep."
Here is a picture of the manual.
Here is the link to the Grove_Indoor_Environment_Demo:
https://github.com/seeed-studio/grove_indoor_environment_demo GitHub - Seeed-Studio/Grove_Indoor_Environment_Demo: Demo for Intel Edison
In the manual they provide additional information regarding the libraries. Please let me know if you need that information or if the webpage is self-explanatory for you.
Thanks for the detailed information.
If you have flashed the Edison but the sketches don't work, and also the Python script seems to be running but there is no activity on the GPIO ports, I'd say that the board might be damaged.
I recommend to contact the place of purchase and let them know about this issue. If you got it recently, the board might be under warranty. On the other hand, if you purchased the board a long time ago, and the place of purchase cannot offer you a replacement, please submit a ticket using the following form. The Warranty Team will receive the ticket and contact you back to provide further assistance: http://www.intel.com/support/mailform/maker/emailsupport.htm
I will give it one last try with an external LED instead of the on-board LED. I don't have much hopes, but this is just to make sure I cover everything.