DuetWifi won't boot without USB connected.
Flashed my DuetWifi with the latest and greatest, listed below. Tried uploading them all at the same time via the web interface settings page upload button. Received message asking if I wanted to flash / update the firmware, said yes and web interface went into it's progress bar state, but the progress bar never moved for hours. Reset the duetwifi by powering down and powering back up. Flashing red light to between micro USB and micro SD card slot, normal lights for voltages eg; 5v and 3.3v, light up when ATX PS is swtiched on. But the DW doesn't boot. Plug in USB, red light on opposite side of micro USB comes on and DW boots.
At this point I flashed one firmware at a time, confirmed they all had the expected progress bars and reboots afterwards, but behavior still exists where I have to have the USB powered before the DW will boot. Voltages check out, no wiring was changed, no shorts that I'm aware of. SD card checks out in other systems.
Any thoughts on what I borked up?
Did more voltage checking and jumper verification. I've got things configured as follows.
DW: Ext 5v - connected to ATX constant 5V line, purple wire.
DW: ATX PS On connected to ATX PS On, green wire.
DW: VIN and GND, connected to ATX 12V Switched, 20a capable line.
DW: 5V jumper set to 5V EXT
DW: V_Fan jumper set to VIN
DW: 5V voltage at Ext 5v is 5.001
DW: 5V voltage at 5V jumper is 4.63 when not connected to USB, 4.7 when connected
DX5: 5V AUX jumper set to 5V EXT
DX5: 5V Ext and GND is NOT connected
DX5: VIN and GND connected to ATX 12V Switched, 20a capable line. Not the same set of wires from ATX harness as DW VIN though, could be separate line.
DX5: V_FAN jumpers set to VIN
DX5: E_STOP jumper set to 3.3v
DX5: 5V voltage at 5V pin in GPIO row is 4.63 when USB is not connected, 4.7 when it is.
DX5: 5V voltage at 5V AUX 5V INT pin is 4.63 when USB is not connected, 4.7 when it is.
Behavior is intermittent now. Sometimes boots and sometimes does not when USB is not connected. Always boots when USB is.
Can see effect of cpu / wifi running when it boots as 5v voltages drop to 4.5.
Looks like the problem is that your 5VSB power from the ATX power supply drops too much when the processor starts up. Do you have a PanelDue connected? The backlight of a TFT panel draws a lot of current.
The voltages I measured above are without the panel connected. I've since connected it to see if it the panel still works, it does, but of course it drops the voltage further.
Pardon my probably ignorance, but is the .4V voltage drop from the the DW 5V EXT pin to the various 5V pins primarily due to the processor, etc….? I was a bit shocked that it dropped that much from the 5V input pin to the rest of the board.
Where can I measure the voltage going to the processor?
I've had suspicions about this PS for a while now, definitely willing to have it replaced, but would first like to verify.
There is a Schottky diode between the USB 5V power input and the internal 5V rail, also between the external 5V power connector and the internal 5V rail. That's why you are losing 0.4V. It doesn't matter in normal use.
UPDATE - RESOLVED:
Short Version: 3rd party ribbon cable used for Duex5 was causing problems, suspect voltage drop.
After another couple of weeks, the issue spread to attempting to boot with USB connected as well. Connected the DuetWifi to a known good USB3 port on a windows system to update the firmware. Did so successfully, then after hitting the reset button, followed firmware instructions to a T, the DuetWifi went into a reboot cycle. It was cycling about 1 - 1.5 times per second.
After some more troubleshooting, I identified that the longer ribbon cable I used to connect my Duex5 was causing the reboot cycle in some fashion. Exact root cause is unknown at this time, I suspect drop in voltage due to cable length / gauge. Replaced cable with one that came with the Duex5 and DuetWifi immediately booted and does so evertime now. Plus I no longer need USB connected to boot and serial over USB is working again.
So the longer ribbon cable was causing nearly every issue I was having. Not sure if it's just a cheap cable, got two in a pack, both exhibit the same behavior. Regardless I am back online and letting folks know in case they run into the same issue.
Dave, not sure if you have a warning on the Duex5 page about this, but you may want to add one. Use good ribbon cables…