I don't know if it is specific 2.3.6 but I have seen irritationg phenomena:
(1) IDE changes USB ports after a while. E.g. you work with COM6 and sudently it doesn't work because another port was created and assigned e.g. COM7
(2) Ports die and no new are cereated. Now when I played with UNO R4 WiFI all ports except COM1 (communicationm) disapeared, and COM1 does not allow uploading. But if I use ESP32-C6 I get a COM8 (Serial USB)-port! Back to UNO R4 - only COM1 there.
(3) It does not help to reboot the PC. The ports do not come back for Uno R4.
Tried another instance of UnoR4WiFi and got COM10.
Plugged in the non working UnoR4 and no port again.
Tried another PC and the is no port assigned.
Is it bricked??? And how to recover since I cannot flash?
UPDATE for (3): it seems a I2C chain with Modulino Buttons and AdaFruit display killed the UnoR4! Modulione seems innocent here, after some disconecting/reconnecting.
It won't as its not related to the board. Its just the default COM port in Windows.
That can happen for a variety of reasons including a faulty USB cable, port connector or the port being already in use when an upload is attempted e.g. serial monitor opened in another session or a terminal program or script still connected. Sometimes you have to shut down the serial monitor.
Have you tried pressing and releasing the button marked "RESET" on the Arduino UNO R4 board quickly twice? The 'L' LED should start blinking and the board should then be in DFU mode. Then try uploading the blink sketch.
I don’t agree with your theory.
COM1 is a standard windows. But the other ports are created on demand when Arduino needs them eg when a new board is connected. Eg when you are plugging in and out three boards there will be three distinct ports allocated. I don’t know practically how the COM ports are created, if IDE asks for them or Windoze sees that someone needs them.
Any how when I had the case (3) no port was assigned at all to UnoR4 until I disconnected devices from the I2C.
(I probably connected a QWIIC connector upside down and shortened the bus that affected USB handling too. Then IDE did not see the board over USB).
If you have not experienced some of the other "features" of IDE 2.x.x... let me introduce you...
IDE booting without the toolbar (no chance of selecting anything)
IDE booting with the toolbar stacked in the upper-right corner
IDE stops saving code changes, stoos compiling and stoos uploading
IDE can not find your board or port after hours of use and no changes to hardware
I know I have missed some of the silly things I have bumped into... but nothing is "broken" ... just temporarily out of order. Restore order by using your troubleshooting skills and patience... and you will be back to code crunching in less timr than pouring your next cup of tea.
That's absolutely fine. You don't have to! Its only a suggestion.
Yes, correct. Windows will dynamically allocate the first available free port and create an association in the registry. But if that port is busy for some reason, it will then dynamically use the next available free port. I don't think that the problem is the driver since another UNO R4 apparently works fine but you were seeing no port at all. Cables can become intermittent though.
Are you now saying that removing the devices from I2C "fixed" the problem and the board is now identified and working as expected? Or are you still getting no COM ports at all?
If the latter, did you try pressing reset twice quickly? When you connect the UNO R4 board, in the IDE under Tools->Port: do you see a DFU port? In Windows device manager are you seeing anything identified as "Santiago DFU"?
Not necessarily. My laptop does not have COM ports if no Arduinos etc are connected. My previous desktop had COM1 although it did not have a physical RS232. And the desktop before that had both COM1 and COM2 (and no physical RS232).
Interesting to note. Just about every Windows PC I have used seemed to have a COM1 even if there was no actual external port present. Of course, some motherboards had COM port headers even when an external port was not actually implemented but I don't think that has been the case for a long time now. I had therefore always assumed that Win creates a "dummy" COM1 port for some reason.
I use Linux on my Laptop nowadays and TBH have never checked whether COM1 appears when its booted into W11. In fact I can't remember the last time I booted it into W11, but that's another matter entirely. Windows also seems to skip some of the lower port numbers when allocating COM ports for USB devices. On Linux it will always start at /dev/ttyUSB0 or /dev/ttyACM0 but on Windows its like COM5, COM6, COM7 etc. I know that Win does reserve ports for devices it has "seen" before, although I am not sure whether this behaviour is related to such reservations. Anyway, point taken and thank you for the information.
Exactly, when I disconnected QWIIC and the I2C chain I got a COM-port back. So I’m convinced I shorted power via I2C in some way, even if the QWIICs ”should” be entered in only one way. The QWIIC have very tiny guides and slightly asymmetrical pins, so maybe not fully fool proof?
The assignment of COM port numbers is very interesting. It jumps around unsystematically like COM6, 9, 8 , 11, 3. 5, …
But the most irritating is when in the middle of work. You start with COM6 and it works stable a couple of hours and studently you cannot flash or run the serial monitor because you now have COM5. It’s another Ph.D. to clear out…