Yellow is the boot pin
Blue is the power pin
Purple is the 24v PS
Clone MicroUSB
Clone USBC (it never gets high enough on the boot pin to boot)
Genuine - hmmmmmm that is interesting
Genuine 10ms instead of 1ms
Mike mentioned a boot delay…looks like that is built in?? That is a 30ms delay? If we add another 30ms so all the boards get it will that just make the genuine take 60ms? Adding a resistor and a cap seems like it is worth the $0.08 it might cost.
Maybe this is built in to the genuine dev boards, but not the knock-off?
Doubled down…
Could maybe add a jumper setting or something if you wanted that defaulted to knockoff mode, and if someone wanted to use genuine, they could change it if they want, just in case it ever caused a problem, although I can’t imagine and extra 50-70ms would matter.
Looking at it it boots so much nicer, everything is at full power when it comes on, the knock offs are booting while it is still powering up.
How would that delay add, if it was built into the jackpot and used with a genuine? I do have all the stuff to actually test it now. Does that time change with 3.3v and those values?
That is not double the time with double the delays thanks for checking that part. Still, who cares about 100ms? I don’t!!!
I assume there is a time that makes it flashmode vs run mode. I will dig around for that spec.
Cool!
I am going to try to wire that up after I grab some food.
I’m excited.
Never knew about gpio2, Boot Mode Selection - ESP32 - — esptool.py latest documentation
To be honest, I’m shocked at how fast it boots up most of the time. I can connect way sooner than I expect
How do we just edit Tsu, make it wait 30ms.
Quick search shows there are ways to change it, but quick search in FluidNC shows they are not as far as I can tell (Assuming what I found so far is correct…)
Must be one of those registry things I know nothing more about than the name.
Nothing.
Delaying the enable pin doesn’t seem to do anything the blue trace takes 30ms longer to get to 3.3v but it does not change the boot pin at all. It seems to get triggered almost teh same moment the enable pin does.
So then I tried the 3.3v in, same.
5V prevents it from booting.
I even tried the boot pin it just smooths out the initial curve but boots at the same moment it normally does.
Now that I say that…it was all on the genuine board. So that is good. Nothing changes on the real board. Maybe it only helps the other boards.
Curious what you get on the clone boards, particularly the USB-C one.
I will swap it but I do not think anything is going to change. The trigger seems to be super super low voltage. For some reason I hoped it was like 1.5V, but we can see from the first tests it obviously isn’t.
So the delay isn’t going to work. Next test is the pull up.
I tested this yesterday and here are the results.
Genuine with 10K pullup same time settings as the previous tests adjusted the voltage a bit.
Genuine, longer time
Genuine no pullup
Micro with 10k
Micro no pullup
USBC 10k
The clones eventually hit 3.3V with that really weird section (is that UART coms?). Why is the genuine not doing that?
Looking back I was in a hurry and I didn’t take notes for each test on the connection. I will recheck the genuine, but I am sure it connected great as did the others. So a pull up seems fine, and like a very cheap and easy fix for future boards. Unless you all see something I do not.
What else should I be testing before any changes are made.
I calibrated and set all the probes to 10k, might explain the extra “dirty” signals, but not fully. Previously I had them set to 1x.
Sorry I know I should not be changing things to make these results easier to read but I am still learning.
Wow, there’s much more common mode noise with the USB-C ESP32. Does that board just have a completely crappy 3.3V power supply?
Yes, I think that is UART communications.
10K what? Input impedance? You want more than that if possible. Much more. If the front end of your scope has only 10K impedance it will be an active part of the circut.
10X sorry.
Why not on the Genuine, then? Weird
I think this proves the USBC versions are not worth pursuing for the jackpot in any way.
