Due won't start after power off-on, have to reset

I have a problem with Arduino Due. If I apply external power supply or just connect it to a computer with USB, it won't start at all. Then after I press the reset button or turn power off and on once again, it begins to work properly.

This issue doesn't appear if the board is put back to work only a short time period after its last turning off. But with a "cooling" time of about 5 min, the problem arises.

I have also tested with Uno. There seems to be no problem with it.

What should I do to fix this problem? Thanks.

I recall similar problems like this with early versions of the IDE v1.5 when it first came out. Make sure you are using the latest IDE v1.5.7.

hiduino:
I recall similar problems like this with early versions of the IDE v1.5 when it first came out. Make sure you are using the latest IDE v1.5.7.

Thanks for the reply. I was using 1.5.6, now I tried 1.5.7, it doesn't work either. I just uploaded the blink example, the on-board led didn't work after power off for a while. Reset button made it work again as I mentioned in the previous post. Any idea?

Let me ask you something. Which pins are you connecting your led to in the board? And which led are you initializing in your sketch?.

difuvi:
Let me ask you something. Which pins are you connecting your led to in the board? And which led are you initializing in your sketch?.

I was using pin13, and I used pin8 before. They all had the same issue.

I have the same problem. DUE doesn`t start at the first powering on or if it was switched off more than 5 se?

Hello, same problem here. First I thought I had a defective TFT, but also a H-bridge or any other components won't work. To get it work properly I have do disconnect and reconnect the power supply, or open the serial monitor in the arduino IDE. Also the reset-button on my board doesn't seem to work.
I am using IDE 1.5.7 and a Arduino DUE board from ebay/china (about 16€).

It would be great to find a solution because I need the DUE to start correctly in my project.
I hope someone knows the reason for this behaviour.

I have same problem. In my board and arduino-Due-schematic.pdf Vddout for core supply use 2x10 mkf capacitor - more than enough (in datasheet ATSAM3X8E for Vddout is offer 4,7 mkf capacitor).
For this reason may be violation of time in initialization flash and boot from SAM-BA .

I've attached a condenser 3000 mkf and board has been successfully starting with power outages in 5 minutes instead of 10 seconds. Now I'm thinking to delete one capacitor 10 mkf, I think this will solve the problem.

Did not help, the problem has not disappeared

Same problem here.
It seems to be common so a fix must be somewhere...
I hope.
Don

If taking out the cap doesn't solve the problem, it does not seem to be related to powering the voltage regulator (?).

It takes about 600µs until the 3.3V and 5V are fully on the 3.3V and 5V pins (> 30 sec off). It takes about 400µs when I reconnect the power supply again after disconnecting it (< 5 sec off).

Maybe the SAM3X tries to do some operations when the supply unit isn't fully stable, at around 500µs or so. Just a guess.

It's very annoying, I need the DUE to boot directly and correcty after it is switched on. It is used in an embedded project...

Its amazing! When I attached a condenser 3000 mkf to Vddcore board has been successfully starting with power outages in 5 minutes and more, more, more - every time !!!! But whith 220 mkf is works as before.

Where exactly do you connect the capacitor? Directly on pin 10 on the SAM3X? Or on the 3.3V pin? I will also try this.

Hi,
Some of my boards have the same problem.. Than noticed that, the boards with problem is R3 boards. But R3-E boards does not have that problem. Do anyone knows, what is the difference between R3 and R3E? thanks, reha

Hi, that's interesting. Where can the R3-E be found? At ebay I only found the normal R3 boards.

Hi Vile,
I purchased many (around 25) Due boards from several sources w/o knowing the "concept of E" (!) ...When I realised the problem, I noticed the difference.. Sadly, most of them R3 , not R3-E ;-((

So, I really dont know, where these boards are came from ;-(

Hi, any news or solution?

Hi,
In my case, I had been lucky. Because, my design has 2 uC's (one of them DUE, other is SANGUINO).. so the soln I found is, when Sanguino starts , it "activate" a relay, which controls the power to DUE card..

In short, Sanguino "activate the relay", so that DUE has initial power (certainly, code does not start) .. then "deactivate" the relay (ie. power of the DUE is cut ! ) , wait for 2 seconds and finally activate the relay ( so that DUE has the power again) .. In this case, code always starts (at least in my boards)

I know, it is not good design, but I could not found ant other solution. In fact another possible solution on "single uC environment" is dedicated WDT chip (will function like Sanguino) ..

but solutions are "unlogical" because R3-E version does not has that trouble but I could not found the difference btwn R3 & R3-E

reha

Hello there,

rehas solution works, but it is only a compromise. In my application, the DUE is the only controller and I don't have the time to wait a second or two until the code runs.
Is there any possibility to contact the developers as this does not seem to be an individual case?

Maybe 'wopo' can tell us where he got his DUE R3-E: Arduino DUE Booard Model DUE R3-E - Arduino Due - Arduino Forum

Solved! Just need to a capacitor which close to the reset switch. The capacitor will let the reset pin to force zero for a short period. However, if you are using many shield on this board, you also need to add a capacitor for each shield's reset switch.