I tried to install it on a spare and got a similar error
2 Likes
Another level in the IRAM vs flash is when the code uses virtual functions. Invoking a merhod on a class with subclasses makes use of function pointers to determine at run time what function gets invoked. If that function is not fixed in IRAM then it’s bad.
At least that’s my partial understanding.
Progress is being made. There have been a couple updates on the SteppingNoVTabs branch. My original panic issue is resolved. The next update was fun where Y and Z jogging always went in the positive direction and Y homed in the wrong direction. In the latest, that is fixed. In that version, opening the settings during a running job still crashed. But then I realized I had HTTP/BlockDuringMotion enabled. Once I disabled that, with a job running I could open settings without error, refresh the Flash file list, and download a file from Flash. However, currently refreshing the UI is crashing it.
3 Likes
Can you let me know if a suitably stable build appears over the weekend ?
I’ll put that on the JL1 laser to test with.
3 Likes
So do you think Mitch will go any further on your comment on Discord?
Trying to download a file from SD with a job running returned a 404 which makes sense. However, when I refreshed the WebUI, it crashed.
I think so. I should probably go repeat that UI crash with FluidTerm connected so I can provide additional detail.
3 Likes
Just remove the F5 key from your 
There’s been more updates to the SteppingNoVTabs branch. I’m getting a bit of testing fatigue. In a previous version, I was having erratic behavior. Currently, X, Y, and Z jogging and homing are going in the wrong directions and STA mode is usually falling back to AP mode. There were I2S enhancements also rolled into the branch which I think is the cause of some of this. I haven’t even tried the crash issues lately due to these other issues.
Here’s what it did a couple versions ago while jogging X+50. I about 
myself.
3 Likes
In my emails today:
5:57 pm Eastern Time:
@MitchBradley pushed 1 commit.
- 621e955 Fixed step count, timing and direction issues
7:55 pm Eastern Time:
@MitchBradley pushed 2 commits.
Yep, I’m on the latest version. The directions were wrong at one point, then fixed, then broken again.
2 Likes
Isn’t he working on totally different stuff right now, sort of just setting aside the random crash issues as well as the refresh issue? Or is that all in the hopes of fixing the random crashing as well as increasing performance and simplifying the code?
The focus at the moment seems to be I2S stuff. Butbhe made a fix intended to help with crash fixes but rolled in this other I2S stuff.
1 Like
Shoot, I really want to get back above 3.8…
1 Like
Had a bit of back and forth with Bart on this. They’re not seeing the wifi issues or direction issues.
From Bart:
We will have to get a few more testers to make sure the directions are the same with the same config with older releases and this branch.
If you could do that, it will help.
Can someone try this out from the SteppingNoVTabs branch and see if they are getting the same results as me? This is definitely easier if you have a spare ESP32 so you don’t have to muck with your working version.
1 Like
Have a lot going on but will try. To be clear, if I load this branch, the test is just to see if the machine moves in the right directions?
1 Like
Try jogging and homing in X, Y, and Z to see if it goes in the expected direction. Be prepared to stop it if it homes in the wrong direction. Also curious if you use STA mode if it falls back to AP mode.
1 Like
OK, will try to get that test done.
I do normally use STA so can try to check that too.
1 Like
I’m reading through this thread, but would this be something that everyone would see as good? I’m trying to come up with a modern UI for fluid that I can have a better workflow as I’m learning.
7 Likes
Wow!!! That is insane!
1 Like
I might be able to tomorrow afternoon
1 Like