I am using Arduinos since 2010; Uno, Nano, Mega, Due... I have developped several project, and I enjoy a lot. Arduino makes possible things that were not 20 years ago with 68 HC11 & Pic 16F84. Thanks a lot to the wonderfull team. I have also bought and started with the Atmel ICE for the Due : mandatory for this complex device. I use Arduinos directly by programming the registers for the Timers, the UART... Much more soft and powerfull than basic functions of the Arduino IDE.
I am engineer in electronic and low level code / firmwares.
I face today a very strange trouble on a Nano Clone from China.
My projects works fine, powered by an external 9 Volts battery.
After booting around 20 times, then the system does not boot anymore; nothing to do.
Then, I remove the 9 Volts batery, and power with USB 5 Volts. Does not boot.
I try to reflash the nano; flashing is OK but no boot after the reflash. I unplug the USB and plug again, press the reset button : now it works again. I unplug the USB and connect again the 9 volts, it works.
It works again for 10 to 20 Power on, then no boot anymore.
I repeated this procedure 3 times : always restart reflashing.
Seems to not be the known problem about the Schottky diode.
My program is very simple : just blinking the LED on D13 with tempo.
I did deep searches on the Web... I did not found such case (only the diode topic, and sometime the regulator, but nothing else). Seems to no be linked to the CH340; seems to be linked to power supply and/or Reset... Even when the Nano does not boot, the green LED (Power) is ON; I have the 5 volts on the Nano pin "5V".
Has someone faced the same trouble ? Has someone found the root cause ? A bad reset ? Could it be linked to a bad bootloader ?
Many thanks in advance.
PS : Below a picture of the board, and the test system with a relay that alternates Power On/Off.
TEK0010.jpg : 5 volts & Reset signal; both rising time is around 150 µs, both cases when the boot is OK and when the boot does not run;
TEK0011.jpg : rising time of the reset is around 2 ms when we press and release the Reset button.
I have compared with an official Nano : timings are the same; 5 volts & Reset are also synchronous, which seems strange to me... I was always told that the Reset must be delayed a little from the Power Supply....
I don't have much experience with the Chinese clones, but the problem sounds like a re-settable fuse to me. It may have nothing to do with your reset procedure, which may be coincidental to the fuse resetting. These boards usually run at 5 volts DC or some at 3.3 volts DC. When powered with 9 volts DC, your on-board voltage regulator may be shunting more voltage to heat than intended, allowing it to run a while, until the fuse blows, which needs time to cool and reconnect.
MBR_78:
OK; thanks for your answer.
I will check in the Atmel documentation is there is fuze somewhere.
There is no resettable fuse in the microcontroller. This is something I would put in the power circuit on board if I was designing it, however with the Chinese clone there is no schematic.
"Has someone faced the same trouble ? Has someone found the root cause ? A bad reset ? Could it be linked to a bad bootloader ?"
This may help.
Notice the USB boot EN on the CTS lead, 1K to ground on pin 13 atmega16u.
Page 7, CH340G, notice no 1K resistor on pin 13 DTR lead.
I tacked a 1K resistor onto lead 13 and Gnd. It now reboots on my smd Uno's and Nano's when I plug the USB cord into a 5V wall wart. Didn't test any further.
Thanks for your answer.
Yes, this could explain the problem.
I did some measurements on the scope (I attached a screen shot) of the Reset Signal; it is almost synchronous with the 5 volts, both on clone and official Nano board (as I said, it seems strange for me the Reset is not delayed....). In the mean time, I did not test several power on for the official Nano.
I will test both : several Power On for the Official Nano, and add a 1k resistor to Gnd on Pin 13.
I though also during the night.... May be a low quality XTAL that would not start.... ? (I faced similar problem 20 years ogo on a TMS370). At this time, I solved it adding 1 MOhms R in parallel to the capacitors.
So, I checked what you told me : this is the root cause ! ! !
With such design, the Reset is synchronous with the 5 volts power up : after 10 to 20 Power up, the Nano Clone is not running anymore : need to be reflashed, and it restarts.
Checking the schematic you gave me.... I thought a simple solution could be to add directly a 200 nF between Reset Pin and Gnd (quick and dirty solution...). This improves the things a lot : I could do around 150 Power up OK, then the problem appeared again. With this 200 nF, impossible to reboot from the PC / USB : I needed to unsold this capacitor to reflash.
In the next days, I will investigate a more proper solution, probably with a diode and a resistor in addition to the capacitor.
I wanted to do a video and a screen shot on the scope, but I was not fast enough ;(
So, as a conclusion, I think this is a critical issue on Nano Clones ! I guess the problem comes from this bull shit CH340G, that looks the Reset Signal in a wrong state...
If someone has an entry point to people who design and produce those Nano clones, don't hesitate to forward them the information; this issue must be natively fixed.
