So, this has happened too many times lately. The Y-steppers loose sync with each other and then the gantry walks off the y-Rail. At first I thought it had to be a hose or cable getting hung up…NOPE. Then I thought perhaps a loose set screw… Nope, Nope. Tight and secured with Blue loctite. Cycling the board resolves the issue…HOWEVER the fact that is randomly happens has me scratching my head and reluctant to put any expensive materials on the bed.
Anyone else experience this phenomenon?
There have been a thread or two here where someone has mentioned the gantry walking itself off of the Y rail in ways consistent with your description. I can’t find the threads at first glance, so maybe we’ll identify those. I don’t recall any specific solutions as I’m not sure anyone understands why
Stating the obvious: The fact that it randomly happens means there’s something intermittent going on.
A note about how FluidNC controls the steppers- there is nothing closed loop about this.
With the two motors defined in the config for the Y axis, FluidNC duplicates the logical commanding of the stepper drivers so that each motor ‘should’ see identical commanding through each stepper driver. Nothing in the system can verify that it works- due to it being open loop.
This isn’t unusual- it is the same exact strategy employed by countless 3D printers and large numbers of various CNC machines.
The only obvious issue from system telemetry would be if there are any FluidNC error messages reported from the stepper drivers themselves.
Other things we should be looking very carefully for:
I suspect that it isn’t cycling the board per se that ‘fixes’ it. After you cycle, you repeat homing/squaring, which mechanically gets the two Y positions in sync as intended.
I believe the reason you’re seeing it walk out is your machine is losing steps on one side only. Once that starts, the error will accumulate until the machine is really wonky.
There’s a very small chance you have one of the known bad FluidNC software versions- so please confirm for us what version of FluidNC you have.
Other useful troubleshooting steps if you have the test equipment:
TMC2209s have internal thermal protections which will shut down motion if they overheat. If you have an IR pyrometer or similar, check the temps on your Y axis heat sinks as you operate. The key measurement isn’t just absolute temp (how hot are they?) , it’s whether they differ. If one is obviously getting hotter than the other then we have a suspect.
What PS do you have on your machine? What are its’ specs (voltage, current)?
Never move the machine at high speeds by hand when off, as this can also blow drivers. That said, if you unplug the Y axis stepper motor connections and move the machine through its extent of travel, does the gangry move freely through its’ range of motion on both sides? This is somewhat subjective but can be an indicator if you have binding or some weird alignment or mechanical issue.
Also consider the potential for spots along the rail that are on the edge of causing a bind. When moving through the section(s) at speed it may not bind, but when starting and stopping or going slow through the section(s) it may bind.
So… it would be good to have an idea of what sorts of travel operations are happening when steps are lost. It would also be good to know which Y side is losing steps (rail vs non-rail) and if it’s only losing steps in one direction or both.
Or it could all be electrical things like @MakerJim has pointed out.
For things like this I like to send the machine on laps at various speeds to see if it can trigger the problem, starting with simple table long squares, progressing to more complex random patterns. If you can get it to trigger on simple travel patterns it’s easier to point at potential causes faster.
An answer you might already know is which side and which direction(s) are losing steps.
Given the fact that it is intermittent/random, my first guess would be a wiring issue. Kinda how with some phone chargers you have “to hold it in just the right spot” after they get worn down, maybe one of the y stepper wires has a spot where the wire is connected most of the time, but if it is bent at just the right angle it will disconnect. Not sure the best way to check for this, maybe multimeter and try bending the wire around?, or potentially just replacing the y motor wire all together.
, and thank you all for your Replies. So a couple of things first. It’s only the non rail side that lags. And it’s not that it’s. Not moving at all, it’s just moving slower than the. Other side of the Y rail. In other words, one’s going at a faster speed than the other, which is kind of goofy. As far as the wiring, I don’t have any extension connectors. I soldered and heat shrunk all of the wires so that I have a nice clean. Install without the possibility of a jumper possibly being a weak connection.
Regarding the Suggestion that. The problem is only resolved after homing. Following a power cycle is actually incorrect. The minute I come off a power cycle I can then go and jog the machine and both motors move at exactly the same speed, albeit the gantry still slightly cocked. It is moving at the same speed, which is not what was happening when it. Dog tracked off of the wire. any additional suggestions? When I get to the high school, I will open the jackpot. Control box cover and check the connectors to make sure nothing’s remotely loose. Although I feel if it was I would have blown out that driver by now and had problems more frequently than what I’m experiencing.
Let’s see what is in your terminal window during this.
Pull a $SS when you power cycle it, and then several more from when it is misbehaving. Post those here for us to look at.
Can you swap the two TMC2209s between the Y positions on the Jackpot after recreating the issue to pull the logs, and see if the problem stays with the rail or follows the driver?
$SS will also show us what version you are running. I do believe there was one firmware version that did this, we skipped it but maybe you flashed it on there?
Ryan,
I haven’t flashed the firmware. It is whatever it was when you shipped me the board.
How/where do I pull the $SS from?
I only use the Web UI to load files to the SD card (via wifi). Then I use my pendant to home the machine, load and run the Gcode files…
The terminal in the WebUI.
Any config changes (like to make a pendant work) can make issues like this. Something as simple as an extra space at the end of a line can cause issues.
When you added the changes for the pendant, you could have easily changed the firmware. It can be done by a file on the SD card, OTA, or wired.
Also if you are suing the pendant, I think it is important that that is up to date as well, that was “experimental” for a very long time and it might still be. I do not use the pendant so I do not know. It does add a huge variable though.
I have the old config file. I could swap it and unplug the pendent and test from the laptop.
That would be an easy test. Worth trying?
You don’t event need to unplug the pendant. Without being defined fluidNC will ignore it.
Ok. So I was about to head to the HS to load the original config.yaml and run the LR4…Then I started thinking, a dangerous thing I know! But I reread all the possible items mentioned in the comments and thought, while testing is good, there is no rhyme or reason to when or if it will happen. Point in case, Wednesday I was at the HS flattening the spoilboard and the LR4 ran flawless for 1/3 of the job, better than 1hour 15mins before it STB. It then did it again 2 more times within a shorter interval. then last 1/4 of the job ran without issue. There have been runs, in xps foam that completed without issues, and some that walked off the rail.
What’s the chance that it’s static electricity confusing the board? and if that’s a possibility, why just the y-min stepper.
who actually grounds their dust hose and how have you done it?
OR am I barking up the wrong tree and should just load the old config and do another very small flatting pass (3 hours for my full sheet table)
Thoughts, insights, suggestions all appreciated…
I had to ground my dust hose for some weird issues a while back, but I cant remember if that was when I was still running the SKR board or after switching to the jackpot.
If you are running the hose that Ryan links in the docs then it has a wire running around the outside of it. You can hook a wire to the end of that and run it to the ground on the cord you plug into the outlet.
Also double/triple check your wiring. That could very well be a loose connection causing the intermittent issue. Have seen that many times.
If your hose is not grounded, you must do that. It can cause random issues, or full on break stuff. If you do not have a hose with the wire built in just suck some bailing wire down yours from the dust shoe and ground the other end.
Just to point it out, there are a number of things that would help us help you in terms of collecting diag data… which you still haven’t done.
Very possilbe.
Where is that cable routed with respect to the vacuum hose? All you have to do is reset a stepper driver and it then has the wrong config in it. (FluidNC only goes in and configures the TMC2209 settings on boot or on command.)
If it were the static thing, there are some steps we could take to try and further isolate/verify that. Unfortunately, that would need the outputs for the diag steps mentioned above.
I appreciate your insight and offer to look over reports to help me sort out my random issue.
I will get the requested diagnostics the next time I’m in the school running the machine. As I mentioned, aside from loading gcode to the Jackppot via the FluidNC wifi, I dont use the webui at all. I have been solely using the pendant to Home, Probe, Load and run my gcode files. so I don’t have the info you are requesting. If I did you would have had it as soon as you asked to see it 
I simply don’t have it…yet.
In the meantime I’m just trying to eliminate any other possibilities before needing to run a job…
I did order the hose Ryan had the Amazon link to. I don’t recall the coil being metal/wire. I’ll check the end next time im there. if it is then thats a simple fix, if not I’ll use Ryan’s bailing wire suggestion.
Don’t get me wrong, I just assumed the grounding recommendation was more a of a rare occurrence CYA statement. Like ground yourself before handling PC Boards and Computer components…Shit, I’ve built hundreds of video editing workstations over the past 3 decades and never used a grounding strap or F’d a chip or board due to SE.
BUT…since Im having a weird issue and the grounding is something i’m missing, that could easily be a random culprit, I’ll fix it before I fire the LR4 back up…
All my control wires are in wire-looms and run in the hose valley with the hose. BUt, oddly enough, the y-min, z-min (left side, problem side) do not, they go right into the jackpot control box that sits closest to the left edge of the gantry…so i’ts like a 6-8" wire loom …
And, It is always the y-min that falls behind/slows down. Very Odd
That hose can build up a ton of static, and certain materials and weather make it far worse. I use one of those dust buckets as well, that little booger can build up enough that I am scared to touch it. I always tap it with my fist before grabbing it.