My apologies in advance for a long post. I’m going to try keep this to relevant details only in the hopes that someone can advise what I am doing wrong.
I’m yet to actually cut anything with the machine yet after about 5-6 hours of just trying to set it up. I’ll briefly outline below what I’m doing and where I’m having issues.
Connect to maslow directly (it really doesn’t like my wifi network for reason, tried 5ghz and 2.4 networks, disconnects all the time)
Retract all
Extend all
Bolt anchors into ground
Jog Z down to lowest height, run calibration.
Device says it’s calibrated. So far so good, no issues. I pack the machine away thinking I won’t have to complete calibration again.
Next session, set it up, extend belts, bolt into ground etc, but by the time I’ve done this (maybe 2 minutes?), the device has disconnected from the web interface. When I reconnect, it’s completely forgotten what it was doing, and the only thing it lets me do is extend all or retract all.
So I need to unbolt the pulleys, let it retract, extend them all and repeat. I was never actually able to get any further than this. I tried this numerous times, could never get further.
Then today I figured I would start from scratch. Let the machine redo it’s calibration etc. I went back to my computer every 30 seconds or so and clicked something in the web interface to prevent it from losing it’s connection. I was even able to install a router bit for the first time. I tried running some very basic gcode (just a circle) to see if it would move by itself. It moved for about 1 second and then I assume the device crashed and rebooted? It disconnected me from it’s wifi and was unresponsive for about 3 minutes maybe. It automatically downloaded a .dms file which just says “success” inside it.
I guess I’m just after some general help here if possible, but I also want to know if there’s any way to tell the machine “hey, your belts are already extended, find where you are and continue” or something along those lines.
Having to undo all the belts, retract them and extend them again every time the machine has a hiccup (which is about every 5 minutes) seems tedious. I’m sure I must just be doing something wrong?
Quick update, today I thought I was finally making some progress. Got it to “ready to cut” state. Generated a small square in Krabzcam, uploaded the file. Did everything with the router bit, got all that set up. Turned the router on, hit play. Nothing happened for about 10 - 15 seconds, I thought maybe it’s taking its time or something, and then I noticed it was doing something, it was unspooling one of its belt entirely. Only one mind you. I tried hitting the Estop, but it wasn’t responding. It continued to unspool until I pulled the power.
So now I’ve got one of the belts unspooled all the way to the point where I think it’s started going the other way on the spool. Not sure how I fix this now.
The spool which has unspooled, you may be able to ease the belt out without pulling the machine down, Don’t try applying to much force or you may damage the belt. The idea is to work the belt out slowly and carefully.
If it is too difficult you may have to pull that arm apart.
Thanks @ian_ab I managed to slowlyy ease it out and then back in eventually, so thats sorted. But I can confirm that regardless of what Gcode I try, it results in the same thing, zero movement from the maslow, then some tiny little vibrations from that top right motor, and then quite quickly it starts unspooling itself entirely.
If I pause the device at this point, it responds but then I can’t do anything else, zero commands seem to work. If I stop the device it seems to crash entirely and restart itself?
What may help if you can put a copy of the maslow.yaml file and a copy of the serial.log files up. Just do a Retract, Extend and Test, don’t try to jog at this stage. It maybe a magnet has come loose, but lets see if we can identify that before you have to pull the arm apart.
What happens if you press the test button on the Setup menu?
If the maslow starts unwinding belt, turn the power off. I recommend getting a push button power switch to make that easier. Incidentally, I am in Australia
@ian_ab I’ve done what you suggested and attached those files. I should mention that the device jogs without issue, all belts remain tight etc, nothing weird there at all.
Also, it now seems to be remembering it’s settings, so this time I didn’t need to calibrate it after turning on for the first time. Before now it was constantly going into “unknown state”, even if the browser disconnected for 30 seconds and I would have to start all over again. So it seems that issue is solved maybe?
Incase it’s helpful, here is the gcode file as well.. test.nc (2.3 KB)
Also, it now seems to be remembering it’s settings, so this time I didn’t need
to calibrate it after turning on for the first time. Before now it was
constantly going into “unknown state”, even if the browser disconnected for 30
seconds and I would have to start all over again. So it seems that issue is
solved maybe?
To be clear about what’s going on.
If the machine is in the state ‘ready to cut’ and paused, it will remember where
it is across a power off or reboot.
If it is moving, then it will not know where it is or where it left off if it
reboots. But even in this case you should not have to do a calibration again,
just the retract/extend to get the belts to a known state.
If you keep having problems, then one thing that it may be useful to do is to
hook up to it with a USB cable and connect with a terminal program (putty can do
this amoung others)
If something is happening that is causing your machine to reboot, it won’t show
up in the log that you access via the browser (that log can only show startup
messages and what happens while the browser is connected)
I think we’ve been assuming this is a wifi connectivity problem, but it could
also be a hard crash that’s causing a reboot.
I think I’ve resolved the wifi issue by locking the Maslow to a single access point. I suspect maybe it doesn’t like something about my network, not sure. I wasn’t thinking the wifi was causing this but was simply mentioning it as an additional, possibly relevant piece of info.
I’m assuming I can SSH into the machine? Are there guidelines for this and any info I should be getting?
Just on a side note, is it normal to have these sorts of problems with a brand new machine?
I think I¢ve resolved the wifi issue by locking the Maslow to a single access
point. I suspect maybe it doesn¢t like something about my network, not sure. I
wasn¢t thinking the wifi was causing this but was simply mentioning it as an
additional, possibly relevant piece of info.
Ok, that behavior is your system not accepting that you really want to use this
network, even though it doesn’t have Internet access.
can you change logging from info to debug (fluidnc tab under flash settings) and
then send the log that you get when you connect? I think I left in the URL
reporting so that would tell us what your system is trying to access (so that we
can tell the maslow to lie to you that you really do have access)
I¢m assuming I can SSH into the machine? Are there guidelines for this and any info I should be getting?
there isn’t ssh, but there is telnet. but if you are really loosing your network
connection, that won’t help (other than confirming that it really is a network
problem)
Just on a side note, is it normal to have these sorts of problems with a brand new machine?
@bar I tried that gcode, exactly the same behaviour. I clicked pause the moment it started unspooling that same belt and then the device becomes completely unresponsive. At the end of the log file attached you’ll see I tried to retract the belts but had to reboot the machine to do this.
Oh, forgot to mention, jogging works completely as expected, well, I’m not sure if what I’m seeing is normal but it seems to move at a reasonable speed in whatever direction I click and all the belts remain tight.
Hi Alex,
I tried running your file, very slow, but it ran fine and completed, so it’s not the code. Didn’t run the router. The maslow.yaml file looks good, would suggest you change your grid sizes and rerun Find Anchor Locations (previously Configuration). FluidNC, Config Items, change the three values Set each then Save and run Find Anchor Positions.
I’d recommend deleting the maslow.yaml file entirely to wipe the settings are starting fresh. I’m not totally sure what setting is acting up, but it seems like the quickest way to find it is to start over.
When the belt starts unspooling like that, if you give it a tug on the extended part does it stop? I think that what we’re seeing there is the belt trying to extend, but not having any force on it to pull it out. Giving it a little pulling force to work against I think should stop it, but that won’t fix the issue of moving so slowly.
The reason it was slow was the value F400 for movement in the file.
I compared the yaml file to mine and the only differences were the Calibration Grid entries and the kinematics: which was expected. I also had an entry under MaslowKienematics: of fixedZ: false which wasn’t in Alex’s yaml.
I compared the yaml file to mine and the only differences were the Calibration
Grid entries and the kinematics: which was expected. I also had an entry under
MaslowKienematics: of fixedZ: false which wasn¢t in Alex¢s yaml.
check for whitespace or unicode characters. fancier editors make it easy for
such things to sneak in.
if something is corrupted in the maslow.yaml file, it could get itself into
grief.
but this is a case where we need something to stop if the belt is being given
power to move out with no encoder movement for more than 5 seconds.
When I started looking into that, copilot reported that there was one odd case
(which doesn’t sound like the situation here) but that there was a mechanism in
place that should catch that. I believe you did merge the fix for that corner
case in 1.13
This is why I implemented the watchdog to catch cases where it was not
computing, but was running the motor out.
There were so many things fixed in 1.13 that I think he needs to move up to that
and see if it solves the problem.
Still have yet to get it to work. I’ve tried all sorts of things, started completely from scratch. Current issues include jogging but belts constantly going very slack, console log constantly giving “error:3” and “error:152”. Device regularly just stops responding all together, requiring a reboot, then is still unresponsive after a reboot. Also regularly giving the message “panic” which I assume is a kernel panic?
I don’t want to diminish the likely many thousands of hours that have gone into this, but I really don’t think it’s fair to sell something that requires this much work to get going. I was completely happy to spend a “fair amount of fiddling to work well” - As stated on the shop page. But I feel the key word in that sentence is “work”. At the very least I expected it to work. From what I can tell, I’ve got a very expensive paper weight that I can move about on the floor.
Again, I really don’t want to be negative and I especially don’t want to be mean, just stating facts
