A question about limit switches, zeroing, and turning the machine off

Community Discussion đź“„ General
1

I have the LongMill MK.2, which includes limit switches and the spindle. All software is up to date.

When I cut out guitar fretboards, I tend to do them at the end of my work day, and finish off the frets the next morning. My issue is that when I turn the machine on the next day and pick up where I left off, the machine is not cutting in the correct spot. I’m not sure if I am missing something in my steps or in the software.

Let’s say I am shutting the machine down at fret number 3, and it only cut the slot halfway, so I need to go back to fret 3 and then finish the rest of the frets.

My process of shutting down and starting up the next day…

  • Stop the job.
  • Put spindle at the x,y,z coordinates(which is G54 workspace)
  • Disconnect from Gsender and shut off the power supply
  • Next morning, turn on Gsender and Longmill
  • Zero the x,y,z coordinates, and then home the machine
  • Load in Fusion file to start cutting at fret 3
  • Hit cut

The only thing is, the cut is off, and by a very small amount. It widens the slot slightly that it started cutting the day prior. I was under the impression that the limit switches allow me to save workspaces, but they are always off by a small fraction.

Nothing about the machine or part moved overnight, unless, of course, there are ghosts. I am stumped.

2

Just to make sure we are on the same page, homing sets the machine coordinates to zero at the home position and probing sets the work coordinates to zero. The machine coordinates are the source of truth and the work coordinates are just stored offsets.

There isn’t any need for the first part of your fifth bullet point “Zero the x,y,z coordinates”. X, Y, and Z work coordinate offsets persist through power off. The way your doing it now you have two places for small errors to add up. The accuracy of both the limit sensors and the probing process.

You need to keep the homing part because before you home the machine coordinates are zero wherever the machine is when powered up. Either that or just in case the machine gets bumped. I know there was a reason I kept the homing part.

The sensors, at least mine, tend to be accurate enough for me to continue work. Take a look at these cribbage boards I made awhile back. I used 4 pieces of of smooth 6mm rod and holes in my work/spoil boards to be able to take the stock on and off the mill. The tops with the inlay are multi-sided and the top needs to come off for the inlay, then put back on to plane. Then taken off to sand and put back on for the laser work. The holes form a rectangle with XY zero dead center and XY is set once for the whole job. Just need to zero Z for bit changes.

EDIT: I should add that I’m not familiar with how the start from line feature works. I think you’d be okay as long as you are using absolute and not relative moves in your G-code. My boards are a bunch of short jobs that only use one bit so I’ve never stopped in the middle of a carve on purpose. The few times I’ve stopped in the middle of a carve I just made new toolpaths for the parts that weren’t done because I was afraid of messing up my work. I guess I’m too lazy to do a couple of test carves and figure out how it works. Sigh…

3

Thanks for the reply.

Yes you are correct on the verbage, my
apologizes!

So are you saying I shouldnt zero the work coordinates on start up and trust the workspace that is already saved? But still home the machine on startup?

Does the machine need to maintain constant power at all times, like stay plugged in? Even when powered off? Because I do sometimes unplug it from time to time.

I have noticed when I turn my machine on it jolts. It has always done that even before I got the SLB. I always assumed that was normal. :man_shrugging:

I have never resumed the job with the start from line feature either, but I have done lots of jobs from editing the code in Fusion and just inputing the new code into Gsender.

Thanks!

1 Like
4

No apology necessary! I just wanted to make sure I didn’t confuse you too much with what I was gonna say.

Yes on the first question. I have the home on startup on in the firmware. My machine is locked when I connect gSender and homing is always the first thing I do.

No on the second question. Pretty sure the offsets are stored in the same type of memory as the firmware settings EEPROM.

I get that little jolt as well but I think I can avoid it if I connect gSender before powering the board. I push the E-Stop button to turn my machine off. Not sure if that’s how you power off a MK2 with SLB. I have the original LB and my E-Stop is the older model.

So I power up my pc, run and connect gSender then turn the E-Stop and home.

EDIT: By first and second question I meant the first and second question paragraphs not individual questions.

EDIT2: electrically erasable programmable read-only memory that’s EEPROM and it’s non-volatile which in the context of computer memory means it doesn’t have to stay powered.

5

Hi Houtwinkel,

Houtwinkel is undeniable Dutch, how did you end up with that name on your shirt?

Oh whoops there I start drifting right at the start. Silly me. Gosh lost in the woods, while walking in the desert. Lemme track back.

I am not familiar with stopping carves somewhere else than at the end of it. Selfinduced. My toolpaths are like Michaels, single jobs, I never stack single bit toolpaths let alone having a toolchanges in it. My workflow doesn’t seem to need them. I am prety lazy, though, so if it even helps me do less one second longer I’m in.

I don’t have lim switches so can’t be of much help there either.

So am I here just to stear jet another topic down the cliff. Jjjjup..
No wait! I do have jobs that require to keep the same xy0 at all times. Like Michael sometimes swapping on and off the machine bed. Sometimes even returning it on another location, using a small dimple to rezero. And here is the crux, that proces cannot ever result in a perfect return to xy0. You can get close, but not like micrometerperfect close. If I would stop mid carve, pick up my workpeace and replace it and rezero at my mark, I will have deviations.

The difference is that inbetween toolpaths, that deviation isn’t as critical. Who cares or sees a few micrometer difference between your pick up pockets and the cutout?

And after aaaaall those words, I come to what I think happens. When powering up my Longmill MK2, at the moment gsender pairs with the cnc, you hear a distinct kick in the steppers. I have never seen my machine differentiate after that, but if you ask me, and you did kindasortha, thát is where your tiny g@¿>⊙°m/t comes from.

And always will.

Now to test this hypothesis, you could see what happens if you only power down the screen and have the complete machine hold it’s position untill you come back to continue. If you manage to return without hickups after that, the hypothesis becomes theory and you can work out howto work around it. -either not stopping a job mid carve, or never power down mid carve. Maybe you can even test if leaving the controller powered up results in no or less kicking when powering up gsender.

Science it!

1 Like
6

Hi there!

That is a Dutch name, which is also my business name, translating to “The Woodshop”. Everyone in the luthier business uses their actual name, but my actual name is long and hard to pronounce anyway lol. So, I went with “the Woodshop” as it’s a great talking point for most people. Fun fact: In grammar school, my teacher gave me an F on a paper because they thought I spelled it wrong, and I didn’t capitalize the first letter. The first letter isn’t capitalized, and my father marched into the school the next day to berate said teacher. Fun times with my name most of my younger life lol.

Anyway! The fret slots are .023" wide, and any wider than that, it becomes a problem for the fret tang, and the fret won’t seat properly and stay in the slot. It widens the slot to about .025", so whatever is happening is off by my guess .002". In the guitar world, these small measurements are critical. It is just very odd for the machine not to remember the position it is in, even with the limit switches. Other things like pockets for other stuff aren’t as critical, and a ± of .002" is fine.

But I will try some of the different choices that are mentioned in this thread. I know these machines are very accurate, and to me, it must be a human issue. I also not everything is perfect, but I also need some things to be perfect. After all, that is why I got the machine!

I am going to try shutting down the machine with the estop first and power it back on with it engaged to see if that causes it to not jump. When I put my hand on it, I can feel it move, which you mentioned is probably the cause, and throwing everything off.

Thanks for the insight…I will be back with my findings.

1 Like
7

That’s exactly when it happens, on my MK1 with regular LB, but not if my E-Stop is pushed because the motors haven’t got power. So I connect and then turn my E-Stop to power the motors. Unless I forget and then I get a click or clunk.

A separate but kind of related issue I’ve had in the past is if you let your PC sleep. I had it set up so it would would never sleep when running a job but if I forgot to push the E-Stop after a job when the PC went to sleep gSender would lose connection, presumably, and the motors would make a quiet but constant humming noise. That also got me in the habit of always pushing the E-Stop when done, even after disabling any power saving stuff, just in case.

1 Like
8

Don’t forget that when you do your cuts you are exposing new wood to the air. I found that when doing deep inlays, if I cut the pocket and plug on one day (and confirmed that the parts fit together) and then come back the next day to actually do the glue up, the plug will no longer fit properly. That’s because of moisture in the air being absorbed into the newly exposed wood thereby causing it to expand and distort.

I’m not saying that this is for sure what your issue is but it’s something to consider. You could try doing your shutdown and restart steps all at the same time on the same day. If everything lines up then it’s a moisture moisture issue and you need to always finish your cuts on the same day. If they still don’t line up then it’s probably something to do with your Longmill.

Rob

1 Like
11

I highly doubt moisture would cause the fret slot to be oversized. I can observe that the cut is off by .002" in real time by watching the cutter not going into the exact slot it cut the day before. I’ve also stopped cutting on the same day and taken an extended lunch break, and had the same issue.

I can’t say I’ve had issues with inlays like you have mentioned because I also cut them and then install the inlays the next day or sometimes several days later. My inlays are always cut either .004"-.005" undersize to the pocket anyway.