Falcon 2 Pro stuck in startup loop

garrettoberg · August 29, 2026, 7:10pm

Hello all. I have a Creality Falcon2 Pro that I got in a nonworking condition. My understanding is that this machine hasn’t functioned properly since a new controller board was installed (from Creality) to solve an issue with the air assist fault. That ended up being a bad airline connection. On startup the gantry moves to the upper right corner then to home in the lower left corner. immediately after homing it starts running in a loop, moving to the max XY upper right corner back to 0,0 home. I’m running the latest version of Lightburn and have tried refreshing the controller board and laser module with the current version of firmware from Creality. I’ve seen elsewhere where the G54 not a number or NAN code cause issues but so far I haven’t been able to clear it using the procedures.

garrettoberg · August 29, 2026, 7:20pm

Edit: I sent a basic file to engrave a rectangle from Lightburn and it executed it (although in the wrong location) and went right back to cycling from corner to corner.

MikeyH · August 29, 2026, 7:46pm

If in warranty, I would request a new controller board because the replacement appears bad.

Here are a few things to try first:

In the Console window, enter $RST**=*** to perform a full factory reset from the EEPROM.
If it stops all those extra moves when trying to Home the machine (front left), then enter $$ in the Console window.
Copy/paste that list from the Console window into a Reply so we can see if the parameters make sense.

garrettoberg · August 29, 2026, 8:27pm

Thank you for hopping in to help!

Not likely to get a warranty from Creality since I don’t have the proof of purchase as I’m the second owner. If needed I will buy a new board but I am thinking that it’s firmware related.

After trying #RST=* I got the message: error:1 G-code words consist of a letter and a value. Letter was not found.

Here are the $$ values:

MikeyH · August 29, 2026, 10:24pm

Too early to speculate.

This indicates it is communicating. However it also suggests a USB communications issue. Make sure GCode Clustering is turned off. If it is, then try to find a quality USB cable that is well shielded and has a ferrite (big knot) on it (translation: Not a $10 Amazon cable). Also be sure you are not more than 10’ (3m) for the USB cable length.

Cell phone camera images of a screen are difficult to view. For Console information, Windows copy/paste works best**. For screen captures use SHFT+WinKey+S to create an image.

Let us know the status if the USB thing.

** = Use the Machine Settings window to save a backup copy of these parameters. We may be suggesting some changes and need something to fall back on just in case.

parsec · August 29, 2026, 10:37pm

Warp8?

The correct command is $RST=*

MikeyH · August 29, 2026, 10:47pm

Darn dyslexic fingers

Curious, that should not have worked!

garrettoberg · August 29, 2026, 11:03pm

This worked, nice catch!
After the machine reset it didn’t solve the issue. I will work on posting the parameter list. There seems to be an issue with getting comms back in time as I get longer lists depending on when I send the $$ command. I didn’t note it before but I am only able to send commands while the gantry is moving and the response is delayed until the move is done. If I stop it by raising the lid or the stop button in Lightburn, no commands are recognized as it’s in a hold mode.

garrettoberg · August 29, 2026, 11:04pm

@mike-oz - I will see if I can find a better cable. Noted on the screenshots for clarity.

parsec · August 29, 2026, 11:18pm

The screenshot lacks a lot of settings before the OK 102 112 120 121 122 130 131 and 132.
Try typing them into Console Window to check/input/check values if they stick…

github.com/gnea/grbl

doc/markdown/settings.md

master

## Getting Started

First, connect to Grbl using the serial terminal of your choice.

Set the baud rate  to **115200** as 8-N-1 (8-bits, no parity, and 1-stop bit.)

Once connected
 you should get the Grbl-prompt, which looks like this:

```
Grbl 1.1d ['$' for help]
```

Type $ and press enter to have Grbl print a help message. You should not see any local echo of the $ and enter. Grbl should respond with:

```
[HLP:$$ $# $G $I $N $x=val $Nx=line $J=line $SLP $C $X $H ~ ! ? ctrl-x]
```

The ‘$’-commands are Grbl system commands used to tweak the settings, view or change Grbl's states and running modes, and start a homing cycle. The last four **non**-'$' commands are realtime control commands that can be sent at anytime, no matter what Grbl is doing. These either immediately change Grbl's running behavior or immediately print a report of the important realtime data like current position (aka DRO).

This file has been truncated. show original

garrettoberg · August 30, 2026, 7:25pm

Okay, after digging and more research and reading…

USB GCode clustering is off. USB cable is a standard data cable, and no indications that there is an issue sending/receiving commands in Lightburn Console, assuming the correct syntax is used . The issue occurs whether or not the PC is connected via USB port.

After entering $RST=*: (I deleted a bunch of the "ok"s since to make reading easier)

$RST=*
<Run|MPos:19.100,19.812,0.000|Bf:126,65485|FS:3000,10|WCO:0.000,0.000,0.000>
[MSG:Pgm End]
ok
ok
<Run|MPos:3.438,3.562,0.000|Bf:127,65510|FS:0,0>
ok
ok
[MSG:Restoring defaults]
[OK]system start-up CV50-Pro-MASTER-Release V1.0.1 20240325-1042
[OK] ILmp_app_init OK
[OK] gas_init OK
usb connect succeed
<Idle|MPos:0.000,0.000,0.000|Bf:127,65535|FS:0,0|WCO:0.000,0.000Idle|MPos:0.000,0.000,0.000|Bf:127,65534|FS:0,0|Ov:0,0,0|A:S>
[MSG:Using cpu_map…esp32-s3]
Grbl 1.1f [‘$’ for help]
ok
ok
[OK] ILmp Link connect ok
<Run|MPos:56.400,58.513,0.000|Bf:126,65478|FS:3000,10|Ov:100,100,100|A:S>
inquire_info->power:22W ,CV50-LASER-22W-Release V3.0.5 20231129-1001
<Run|MPos:383.038,397.400,0.000|Bf:126,65465|FS:3000,10>
[MSG:Pgm End]
ok
ok
Homing
[MSG:Pgm End]
ok
ok
[G54:-nan,1.579,2.085]
G54 Offset values not 0
[G55:-nan,1.579,2.085]
[G56:-nan,1.579,2.085]
[G57:-nan,1.579,2.085]
[G58:-nan,1.579,2.085]
[G59:-nan,1.579,2.085]
[G28:-nan,1.579,2.085]
[G30:-nan,1.579,2.085]
[G92:0.000,0.000,0.000]
[TLO:0.000]
[PRB:0.000,0.000,0.000:0]
ok
ok
<Home|MPos:-0.363,-0.363,0.000|Bf:127,65535|FS:5657,0|WCO:0.000,0.000,0.000>
ok
ok
<Run|MPos:378.150,392.325,0.000|Bf:126,65470|FS:3000,10>

Then the $$ command:
$0=10
$1=250
$2=0
$3=4
$4=0
$5=0
$6=0
$10=3
$11=0.010
$12=0.002
$13=0
$20=0
$21=1
$22=1
$23=3
$24=2000.000
$25=4000.000
$26=20
$27=5.000
$30=1000.000
$31=0.000
$32=1
$100=80.000
$101=80.000
$110=50000.000
$111=50000.000
$120=5000.000
$121=5000.000
$130=400.000
$131=400.000
$150=36
$151=100
$152=30
$153=0
$154=1
$155=1
$156=0
$157=1
$158=1
$160=1
$161=13.00
ok
ok

What I am trying to understand is where the Run to position commands are stored and how to view them. They seem to be on an infinite loop and everything I read about GCode and grbl says they don’t operate that way. Homing works fine. I read the linked GitHub document last night and tried looking at $N0 and $N1 and both are empty. I am familiar with using an Arduino based controller for 3D printing. Since those config files are open source it’s relatively easy to view and edit code. This has me stumped.

garrettoberg · August 30, 2026, 7:25pm

Thanks for this, it’s a good resource document!

MikeyH · August 30, 2026, 10:33pm

G0 Xnnn.n Ynnn.n is rapid traverse to the X,Y coordinates.
G1 Xnnn.n Ynnn.n Fnnn Snnnn is the slow feed to the X,Y coordinates at Ffeedrate and Spower%.

GCode commands are not “stored” in Lightburn. They are created as the design is examined. These are sent to the controller memory, where they are execute in the order received. The controller should not do anything it is not told to do.

Part #1: In the Console window there is a Show All switch. Turn it on, then power up the laser. The Console window may show everything sent to the laser. If you get that improper module travel during Homing, copy all of that window (right click) and paste it in a Reply.

Part #2: Next, load the attached program and use the Save GCode to save it to a text file (.txt). Then run the program to see where the laser actually goes. Make sure Absolute Coords is selected in the Laser window. I deliberately kept it simple (no curves) so a huge program is not created. Let us know if it behaves the same as the design suggests. Laser power is set to 0.5% so it will not burn anything.

Garret Test 01.lbrn2 (3.6 KB)

If we keep chipping away, we might figure out what is happening with your system.

garrettoberg · August 30, 2026, 11:42pm

“If we keep chipping away, we might figure out what is happening with your system. ”

Really appreciate all the help. Here’s hoping!

Part #1, Full list of startup:

The run to max X,Y and 0,0 just keep looping after this. I did turn off the Home on startup selection, so it’s just the unwanted moves.

Part #2 to follow…

garrettoberg · August 30, 2026, 11:46pm

Part 2:

Laser appeared to trace several rectangular shapes near the lower left corner, then a few larger ones in the middle of the bed. Once done it was back to possessed diagonals. At least I know I can get the thing to work with a file while I troubleshoot haha.

Console Output:

JohnJohn · August 30, 2026, 11:56pm

The laser is being commanded to fire with the S5 in the line above.

S5 is on the bottom of the scale representing very low laser output.

$30=1000 so the control board is expecting a number from 0-1000 for how much power is expected.

S Value max is the next setting to look at.

MikeyH · August 31, 2026, 12:07am

W O W ! ! ! That thing is all over the place. Even if it is not moving like that, it thinks it is. Those bolded numbers look like a countdown from a starting 65,535 starting number.

Out of the yellow idea… (sorry, no blue) Is it possible for you to ground the frame of your machine to the ground wire in an outlet or something else?

If that does not solve it, then my opinion is to return it and order another machine.

garrettoberg · August 31, 2026, 12:18am

The low power was intentional in the Gcode file that @MikeyH sent me. Not trying to cut just test movement for now. We are trying to figure out why my laser will accept commands, but once they are done it reverts to perpetually cycling from 0,0 to 400,400. Any thoughts?

garrettoberg · August 31, 2026, 12:22am

From Google:

In GRBL real-time status reports, Bf stands Buffer State, showing how much memory is currently available on the controller.

It appears inside angle brackets in status strings like <Idle|MPos:0.0,0.0,0.0|Bf:15,128>.

I can try the ground wire later. Unfortunately I bought the machine second hand, knowing it was a bit of a project so I’m determined to fix it

parsec · August 31, 2026, 1:20am

You may check for a newer firmware…
Or try reflashing the same firmware or even an older version.
Also pay attention that some lasers have one firmware for the controller and one for the laser head