Thanks so much for looking at all this and sharing your expertise.
No. I booted (powered on) the board on power supply.
Referring to the AP set of logs:
I powered-on, waited 30s for the SSID to show up on my MacBook sitting on the deck of the LR3.
Then I double checked that my phone also could not see the ESP32.
Then I connected by USB and used Terminal on the Mac to start FluidTerm.
The SSID shows within 10s.
I then sent $SS and captured everything.
Apologies for the double logs. - I captured the first set when I was testing to see if it would show up on my mesh network at boot up. It does not. And when I was testing it, I had moved a mesh node to the deck of the LR3. (Impractical, but I wanted no doubts about the signal strength.)
So -
I observe the same behavior when I use either AP mode or STA>AP mode.
I am only able to see evidence that it has booted after power-up if I connect by USB and start FluidTerm. Following that, the SSID is detectable on all nearby devices. It is stable (AP mode is the most extensively tested for stability, I am in the process of testing STA mode.)
I have also been doing something dumb, which is searching randomly on Google. I have run across a few posts where the talk about “rise time” or other characteristics on the 3.3V rail. I’m not an electrical engineer, but I’ve started to think than my Meanwell power supply may take too long to get the 3.3V rail to full power, thus creating a boot issue. (But then again, I’ve got no personal knowledge on which to base this.)
Ok, so I’m not an electrical engineer either (although I did play one in college once), but maybe power is another variable to consider.
I have not had any issues, but I’m using the 24V power supply from Ryan to power my board. Perhaps something to check on for the people having issues vs the ones who are not is how the power is supplied.
I wonder if at that point it is actually rebooting again, so we just aren’t seeing a problem log. Or that it just didn’t fully boot up anyway until you plugged in if it’s a power issue
Is it possible to have the steppers not power on until the ESP’s done booting? Would it be meaningful to just connect a big 12V battery in place of the normal power supply?
I’m not sure about the battery, if maybe that would be more stable, but maybe a good test would be to try with a 24V power supply if you have one handy to see if that makes it better without changing anything else
Edit: Although I see now that @dgkeith237 has responded that his is 24V
William I think we should have you open a new thread, chances of these being the exact same issue are slim. Lets separate them. Please tag me @vicious1 and start by stating your exact issue.
Yes - it’s grounded. I have to leave for my jobby job (I’m already late - I need to come up with an excuse…)
I do have a 12V laying around somewhere, but it’s a Meanwell also. I have the 12V supply that I purchased with the MPCNC kit last year sometime, but I can’t find that barrel jack (why is it always the dongle?)
I love the Meanwell’s so I’ve stuck with them on my projects. But we may solve this if I was using the same one you’re using. I’ll order one.
Cut the end off then you are left with a power and ground wire. I had one of those dongles go bad about a week ago powering my laser. Its now hard wired lol
Just a note on this. When installing my board and screwing down the terminals, the block does twist some. It is still connected, but there is likely enough slop or some space between the board and the block that will allow it to rotate slightly. So, it may come from the shop straight and then get rotated on install. This may not actually cause problems if it is only the wires moving and the solder joint on the backside is OK.
Yes, it is long pins. They can bend. It would take a whole lot to mess them up though. Be gentle the wires should be constrained after the board as well so they should have no way to come loose.
Tried 2 power supplies. One is the 12V supplied with my MPCNC kit in 2022. I also had a 24V 2A power supply for my laser. Neither of those resulted in broadcast of the SSID. While on the 24V, I did use Fluidterm to get the SSID, then power down and power back up but still no SSID detectable until after Fluidterm.
I did hear the motors energize with the 24V, and it’s a pretty wimpy supply. Tomorrow I can take it apart further and try again.
Just to follow up on this after doing a little research, the $WiFi/FastScan option will tell it how to handle this.
Setting Fast Scan to true makes it choose the first one, false makes it look for the best one.
FluidNC wiki is missing this information, so just wanted to put it out there in case someone else comes across it
Edit: It does have the ability to choose the sort method too, but FluidNC is not exposing it as a property, so it sorts by signal strength if $WiFi/FastScan is Off.
If you look at ESP-FAQ Wifi latest Documentation, I had problem 111. This could cause what you are seeing in STA>AP mode. My ESP32 was showing a completed connection with an IP address. My router was showing the ESP32 but reported “Failed To Assign IP” and would not connect. Most routers won’t show this, the ESP32 just won’t show on your network.
You may or may not be able to correct this the way I did.
I put the ESP32 on it own vLAN/guest network all by itself. If your router has a guest network setting, you can turn that on and connect with the ESP32 being the only client.
I set the network to 2.4ghz only.
I assigned static IP in the router, not the ESP32. (personal preference)
My router (with a mesh network) allows me to lock clients to a single Access Point. My other AP don’t broadcast this network. Doing this keeps the ESP32 from trying to connect to other APs just because the ESP32 moved a couple of feet. This setting is not available on all routers.
These steps basically put the ESP32 in STA mode on a router by itself.
As far as AP mode on the ESP32 not showing up on your devises, I’m stumped.
