#linuxcnc Logs
Sep 03 2025
#linuxcnc Calendar
12:30 AM Deejay: moin
03:05 AM t4nk_freenode is now known as t4nk_fn
03:57 AM lcnc-relay: <lucid.nonsense@> My revopoint trackit kickstarter is being sent today, please don't suck, please don't suck...
04:03 AM lcnc-relay: <lucid.nonsense@> Comes with a copy of Quicksurface, anyone used it?
04:18 AM lcnc-relay: <pink_vampire@> Good morning everyone
04:21 AM lcnc-relay: <pink_vampire@> I just came across this probe, it looks very nice and not too big, but the tip is M2.5
04:21 AM lcnc-relay: so standard m2 stylus will not work
04:21 AM lcnc-relay: what do you think?
04:21 AM lcnc-relay: https://www.amazon.com/PGFUN-Anti-roll-Waterproof-Coordinate-Compatible/dp/B0B3XMFDX5
04:26 AM jpa-: i have the probe, it is pretty nice but because of the length any runout in cheap spindles is amplified and results in inaccuracy
04:27 AM jpa-: to me it seems M2.5 probe heads are easy to find, so shouldn't be much of an issue
04:36 AM Tom_L: morning
04:37 AM -!- #linuxcnc mode set to +v by ChanServ
05:33 AM mrec: I wonder how stubborn a ballscrew service company can be to insist not using V blocks for testing the straightness of a ballscrew...
05:34 AM mrec: today was the last day to work with that company here I'll go for another one. They have a machine for bending a ballscrew but insist to use the disqualified pilot hole
05:35 AM mrec: also I wonder why this bending tool/press for ballscrews doesn't come with v-blocks, does anyone have experience with those machines?
05:36 AM mrec: the only valid part of the ballscrew are the bearing journals in my view
05:36 AM mrec: parts*
05:41 AM jpa-: if they are ground in one setup, the outside is likely as accurate as the grooves.. but yeah, measuring from the grooves would be best
05:44 AM mrec: so far the state of the ballscrew is unknown, when using the pilot holes the state is poor. But the service guys had no idea how to remove the coupler and we didn't do much better but managed to remove it (those two attempts might have deformed the ballscrew end, also there's burr and pits in one pilot hole caused from a puller (that's why I consider them to be unreliable)
05:45 AM mrec: I am not sure how much power was applied on the coupler end to remove everything, I just think it's not in a good state
05:45 AM mrec: so for them I'd just like to focus on everything between the bearing journals and that they just refused.
05:46 AM lcnc-relay: <pink_vampire@> if I go the extra mile to buy a probe, I want a ruby sphere on the tip. (ruby is kinda pink.. and you know me by now)
05:46 AM mrec: they just don't acknowledge that the pilot holes are unreliable, all they say but that's the standard way of getting it right
05:46 AM lcnc-relay: <pink_vampire@> Tom_L^
05:47 AM mrec: we measured a runout of 0.1mm
05:47 AM mrec: even 0.12mm but that was in the lathe which is absolutely unreliable
05:47 AM lcnc-relay: <pink_vampire@> have you pulled the screw?
05:48 AM mrec: no, it's a 40mm ballscrew around 1.4m I would say
05:49 AM jpa-: pink_vampire@: ruby probes are available in M2.5 too, https://www.ebay.com/itm/364684945948
05:50 AM lcnc-relay: <pink_vampire@> I have a a bunch but they are M2
05:50 AM lcnc-relay: <pink_vampire@> I must say, I am so tempting to but it.
05:52 AM lcnc-relay: <pink_vampire@> maybe I need to get the lathe connected. it is a strange 2 speed AC motor..
06:53 AM BorgPipe is now known as NetPipe
06:54 AM JT-Cave: I need to modify the case for my n100dc so I can put a pcie nic in there... it has a weird punch out for a connector of some sort
06:57 AM JT-Cave: https://pasteboard.co/gTnHgAGvIr9B.png
06:58 AM JT-Cave: but I should have a cooler master mini itx case around here somewhere
06:59 AM lcnc-relay: <pink_vampire@> I modified a case of dell P4 computer
07:00 AM lcnc-relay: <pink_vampire@> I think it will be fairly easy to 3d print something to hold the pcie
07:10 AM -!- #linuxcnc mode set to +v by ChanServ
07:19 AM JT-Cave: I "think" the pcie nic will fit in that knock out...
07:22 AM lcnc-relay: <pink_vampire@> is that a low profile card?
07:25 AM javabean_ is now known as JavaBean
07:27 AM JT-Shop: yes, but I would have to take the bracket off as there is no place for it
07:38 AM Tom_L: the punchout is for a D connector of unknown width
07:57 AM lcnc-relay: <pink_vampire@> caliper for the rescue
08:12 AM Tom_L: https://www.carbideprobes.com/ migit have a 2.5mm thread tip
08:31 AM {HD}_ is now known as {HD}
08:45 AM -!- #linuxcnc mode set to +v by ChanServ
08:49 AM lcnc-relay: <rs> mrec: what do you mean with "runout" in a ballscrew? is the machined end not coaxial with the "screw" part?
08:52 AM mrec: rs: yes, the ballscrew comes from a machine which is 25 years old my target is to improve it somewhat. Also the machine had a hard crash the first time I powered it on due to a faulty encoder.
08:52 AM mrec: it's not coaxial anymore, also I don't know the entire history of it the machine doesn't have any service documentation
08:53 AM lcnc-relay: <rs> how did you measure that?
08:53 AM mrec: the machine works and is usable, the only showstopper for me is the Y axis ballscrew, in the worst case I will change it
08:54 AM mrec: I just attached a dial gauge to the crosstable and monitored the deviation when moving it
08:55 AM mrec: however the best check in my eyes would be to put the bearing journal onto some V blocks and rotate the ballscrew and measure how much it's bent
08:55 AM lcnc-relay: <rs> hmm. don't understand. so you say the distance travelled per rotation is inconsistent on the order of 0,1mm ?
08:55 AM mrec: it's wobbly 0.1mm
08:55 AM lcnc-relay: <rs> that IMHO probably won't matter much
08:56 AM mrec: and the backlash is twice as much as it was before factory default (while the X axis is still the same as the factory default 25 years ago)
08:56 AM mrec: I tried to increase the preloading of the nut it didn't change the result
08:57 AM Tom_dev: are the balls worn
08:57 AM mrec: the backlash is 25.5 microns before it was 12.75
08:57 AM Tom_dev: before what?
08:58 AM mrec: balls are replaced, some of them were corroded a little bit
08:58 AM mrec: factory default was 12.75 microns
08:58 AM Tom_dev: what changed
08:58 AM mrec: that's the question.
08:58 AM mrec: the machine experienced crashes on the Y axis
08:58 AM Tom_dev: ok
08:59 AM mrec: I am not entirely sure if the backlash is 25.5 across the entire ballscrew I only tested a few sections
09:00 AM mrec: the wobble is up to 0.12mm at the bearing journal but all the measuring is wrong in my eyes.
09:00 AM mrec: without V blocks all the measuring is nonsense (I do not trust the pilot holes because there's burr from a puller, and the puller was used in a very hard way on it)
09:01 AM mrec: I believe the puller bent the shaft-end a little bit, so the only valid reference should be the bearing journal
09:02 AM mrec: and until now -- it was never measured like that. The bearing service company isn't doing it correctly in my eyes only using the pilot hole is definitely not a good way on that ballscrew
09:02 AM mrec: the ballscrew searvice company I mean
09:02 AM lcnc-relay: <rs> i don't see how a bend of 0,1mm over 1,4m can affect backlash
09:02 AM rdtsc-w: ditto
09:04 AM lcnc-relay: <rs> the screw will have some sagging due to gravity and may resonate like a guitar string (a bit) while rotating with high speed anyways
09:05 AM mrec: I have some moving issues on the Y axis
09:05 AM lcnc-relay: <rs> maybe the nut has damaged channel?
09:06 AM mrec: could be, when I rotate it manually it seems to be okay I feel some preloading on it too which should be okay.
09:07 AM mrec: (when I take the entire ballscrew out, or disengage the nut from the table
09:15 AM rdtsc-w: have seen more than a few "axis rebuild" videos recently, which ended up in total rework - hope scraping isn't in your future, but it may be if chasing microns
09:15 AM mrec: https://streamable.com/lc0mgg
09:15 AM mrec: I wouldn't have a problem to buy a new one but they told me it will take 3 months...
09:16 AM mrec: the entire machine still works fine the machine is useful for some parts definitely not for high accuracy but for some items.
09:17 AM mrec: I don't want to let it go (also I got it for free..)
09:17 AM mrec: the jumping is solved in some areas as far as I have observed, the ballscrew was fully cleaned
09:18 AM mrec: but I think this is somewhat derived from the bending too
09:18 AM mrec: the resolution of this gauge is 0.001mm
09:23 AM lcnc-relay: <rs> hmm. i mis-filling a ballscrewnut such that some balls were in between two recirculation channels. that caused some very weird effects
09:24 AM lcnc-relay: <rs> but it was possible to turn without excessive resistance
09:24 AM lcnc-relay: <rs> +remember
09:25 AM mrec: I think the wobbly ballscrew doesn't make this effect better, my current goal is just to have it as straight as possible once that's achieved I will set the machine to 25.5 microns backlash compensation again.
09:25 AM mrec: so far we measured 0.1mm wobble on that 40mm ballscrew as mentioned
09:32 AM lcnc-relay: <rs> looks like an 80ies machine. surprising that the CRT is working
09:34 AM mrec: 2003, the software is from 1997
09:34 AM mrec: it's good enough to take data via ethernet
09:35 AM lcnc-relay: <rs> nice
09:35 AM mrec: I used paths generated from FreeCAD (and some special commands I patched in)
09:38 AM mrec: electronic edge finders are also supported, I wrote a program for it
09:38 AM lcnc-relay: <rs> my 80ies brother drill/tap even had tool breakage sensor
09:39 AM lcnc-relay: <rs> and a green CRT :)
09:40 AM Tom_L: wonder if it's the same brother that makes printers & sewing machines
09:40 AM mrec: it is
09:40 AM Tom_L: cool
09:41 AM mrec: however mine is a Mitsubishi I studied this machine so well I'd be sad to let it go now.
09:41 AM Tom_L: i kinda recall the beginnings
09:42 AM mrec: all the settings were gone, the backup was defect, the toolchanger program was missing, the ladder was "pseudo protected against curious operators" I recovered it from ground
09:42 AM Tom_L: that's all?
09:42 AM Tom_L: :)
09:43 AM mrec: endstops were all clogged, 2 encoders defect while one of them caused a crash unfortunately
09:43 AM mrec: and I bought the ethernet extension for it RS232 was too slow for trochoidal milling
09:46 AM mrec: I noticed the problem when milling round holes I had some flat spots
09:46 AM lcnc-relay: <rs> i replaced the control in my brother.
09:47 AM Tom_L: flats at the quadrants?
09:47 AM Tom_L: we had that issue on an old boss5
09:47 AM mrec: https://www.manford.com.tw/upload/product/215/manford_vh1300_03.jpg
09:47 AM mrec: basically that's the core of the machine
09:47 AM mrec: it's exactly the same I would say (since it's from the same manufacturer)
09:49 AM lcnc-relay: <rs> flat spots could be backlash
09:50 AM mrec: that's what I think. But I want to get that 0.1mm wobble out of the ballscrew first before proceeding
09:51 AM lcnc-relay: <rs> the thing that caused me to disassemble the screw where i made the mistake when reassembling the nut ultimately was some bolts not being torqued down properly
09:51 AM mrec: I took it out like 5 times already I'm getting tired of it
09:52 AM lcnc-relay: <rs> that caused the ballscrew (which is fixed in that machine, the nut is driven) to "slip" and the machine to lose position with time. looked somewhat like lost steps.
09:52 AM Tom_L: you're getting good at it
09:52 AM mrec: and I'm just losing faith in that service company, their attitude is good they don't even want to charge extra for the extra work now but it's difficult to discuss with them.
09:53 AM Tom_L: they probably have a certain way they test and don't want to deviate even though the pilots are compromised
09:53 AM mrec: I wouldn't mind to pay for the extra service but I'd expect that they dig into that problem a little bit more rather than letting me come up with ideas after thinking about what they are doing
09:55 AM lcnc-relay: <rs> can you measure the backlash with machine on and enabled and the servo keeping position while "rocking" the table back and forth?
09:55 AM mrec: just for fun I asked AI ... AI confirmed what I was thinking without giving it detailed information about which points I critizise from that service company (and usually AI is 50% wrong when I use it... eg. a discussion about PCI-Express end up with statements that the PCI-e card only has isolated memory access)
09:58 AM lcnc-relay: <rs> there is nothing intelligent in current "artificial intelligence"
09:58 AM mrec: statistics.
09:59 AM mrec: probability.
09:59 AM lcnc-relay: <rs> stochastic parrot
10:01 AM lcnc-relay: <rs> i wonder if Emily M Bender is an ancestor of Bender Bending Rodríguez
10:03 AM -!- #linuxcnc mode set to +v by ChanServ
10:07 AM lcnc-relay: <_randymuc_@> 🤣
10:19 AM lcnc-relay: <roguish000@> good morning............. is the link to Matrix dead ???
10:21 AM lcnc-relay: <TurBoss> hello, good question
10:21 AM lcnc-relay: <roguish000@> I keep getting "Connection to the server has been lost"
10:23 AM lcnc-relay: <roguish000@> ok, well, i'm going to restart my modem and router........... bbl
10:23 AM turboss: hello from irc
10:24 AM lcnc-relay: <TurBoss> hello from matrix
10:24 AM lcnc-relay: <turboss@> hello from discord
10:28 AM lcnc-relay: <roguish000@> ok, back (on discord)
10:30 AM lcnc-relay: <TurBoss> 🙂
10:32 AM lcnc-relay: <roguish000@> still error message. just signed out of Element, not signing back in
10:38 AM lcnc-relay: <roguish000@> trying to log back in. get error "Cannot reach homeserver"
10:38 AM lcnc-relay: <TurBoss> what homeserver are you using? matrix.org?
10:39 AM lcnc-relay: <TurBoss> https://status.element.io/
10:40 AM lcnc-relay: <roguish000@> yes
10:40 AM lcnc-relay: <TurBoss> ok matrix.org down
10:43 AM lcnc-relay: <roguish000@> ok, just reboot computer too
10:44 AM lcnc-relay: <roguish000@> TurBoss. I think you are right. same error again
10:44 AM lcnc-relay: <roguish000@> on discord now
10:44 AM rdtsc-w: Neo has been unplugged from the Matrix
10:47 AM lcnc-relay: <roguish000@> so, what or who is Neo ???
10:47 AM lcnc-relay: <TurBoss> the choosen one
10:47 AM lcnc-relay: <TurBoss> 😄
10:52 AM lcnc-relay: <roguish000@> dang. Elemnt in browser is just hung trying to get to matrix
10:54 AM lcnc-relay: <roguish000@> web search says it down, after a RAID meltdown
10:55 AM lcnc-relay: <roguish000@> lots of down reports
10:55 AM lcnc-relay: <roguish000@> oh well. it will eventually come back up
10:57 AM rdtsc-w: https://www.imdb.com/title/tt0133093 :)
12:37 PM -!- #linuxcnc mode set to +v by ChanServ
01:08 PM lcnc-relay: <sysrover@> Good day, trying to add in gmoccapy button to release brake on lathe spindle like in axis. make a button its easy but still cant figure out how to make in hal this connection
01:49 PM JT-Shop: no one here uses gomappy
01:49 PM JT-Shop: ask on the forum
01:51 PM lcnc-relay: <Mister Wallace> So I managed to find a guy with nearly the exact same Shizuoka ST-N that successfully converted it to run LinuxCNC with the existing drivers & steppers
01:52 PM lcnc-relay: <Mister Wallace> His mill is close by enough that he can even share the config files and let me take a look at the wiring
01:53 PM JT-Shop: nice
01:53 PM lcnc-relay: <Mister Wallace> Sounds like I’ll need a helluva lot of pins going to each driver board, like 16 each and there’s 3 drivers on mine
01:53 PM JT-Shop: are you sure they are all used for something?
01:54 PM lcnc-relay: <Mister Wallace> He said _most_ of them are used, the control signals on these are really odd
01:54 PM JT-Shop: interesting
01:54 PM lcnc-relay: <Mister Wallace> Details on his conversion: https://nickparker.co/project/cnc-mill/2-controllers
01:54 PM JT-Shop: at least you found someone that has been there and done that
01:55 PM lcnc-relay: <Mister Wallace> A guy willing to help me at that 😅
01:55 PM lcnc-relay: <Mister Wallace> He said I might need a 7i75 breakout board for all the pins
01:56 PM lcnc-relay: <Mister Wallace> I think I might be able to do it with the 7i76EU
01:57 PM JT-Shop: on his page he used a 6i25 but didn't say what daughter board he used
01:58 PM lcnc-relay: <Mister Wallace> 7i75 is what he told me via email
02:00 PM JT-Shop: interesting, I have a 7i75 somewhere
02:02 PM JT-Shop: the linuxcnc stepgen can do 15 different step types
02:10 PM lcnc-relay: <roguish> whoo hoo............ Matrix is back up..........
02:11 PM xxcoder: thanks neo
02:58 PM lcnc-relay: <rs> FYI matrix.org had major failure attacking the postgres db supported by general error runing a RAID. they restored ~55tb database from backup and replaying transactions from today. story was on hn.
02:58 PM lcnc-relay: <rs> ruining
03:11 PM andypugh: Does anyone here remember how to interact with the buildbots on the dev list?
03:15 PM lcnc-relay: <rs> say /msg linuxcnc-build help
03:18 PM andypugh: msg doesn’t seem to work, just opens a new window that I can type into, but which doesn’t accept the enter key
03:18 PM rmu: seems the buildbot is too chatty for the poor irc server
03:43 PM lcnc-relay: <skunkworks8841@> aww - it is only 470 lbs of lead
03:44 PM lcnc-relay: <skunkworks8841@> I really have a hard time estimating the weight of things over about 150 lobs
03:44 PM lcnc-relay: <skunkworks8841@> lbs
03:45 PM lcnc-relay: <skunkworks8841@> https://photos.app.goo.gl/G9wgHuzD16jUv7uK6
03:47 PM jasen: hi Pete
03:47 PM pcw-at-work: skunkworks8841@: Gallons x #88 ?
03:49 PM lcnc-relay: <skunkworks8841@> do - I did a quck estimation from the size..
03:50 PM lcnc-relay: <skunkworks8841@> it was obviously wrong
03:50 PM lcnc-relay: <skunkworks8841@> should have modeled it in cad.. lol
03:50 PM JT-Shop: I've done that before
03:50 PM lcnc-relay: <skunkworks8841@> me too..
03:50 PM rdtsc-w: wonder if scad can calculate volume...
03:51 PM lcnc-relay: <skunkworks8841@> but now I have a cheap 4400lb hanging scale for other uses..
03:51 PM lcnc-relay: <skunkworks8841@> I checked it with my wieght - seems accurate up to that
03:53 PM JT-Shop: I finally got the n100dc setup with a 7i96s for testing
03:54 PM JT-Shop: now I need to sort out where all the wood working equipment needs to be placed
05:12 PM mrec: is there any update on the S-curve implementation? I wonder how it would perform with a robot arm
05:25 PM lcnc-relay: <skunkworks8841@> I check and I haven't seen an update in months
06:09 PM andypugh: mrec: I don’t think he was going about it in a useful way. He gave up quite early on making it work in LinuxCNC, as far as I can tell.
06:13 PM lcnc-relay: <skunkworks8841@> I think he was going to make his 'own thing'
06:14 PM lcnc-relay: <skunkworks8841@> s curve is hard when you actually have to worry about synced motion, Feed overrides and such...
06:15 PM lcnc-relay: <skunkworks8841@> I 'think' tormach just worked on smoothing motion at the jerk points..
06:15 PM lcnc-relay: <skunkworks8841@> lol - if it was easy - mach would already have it.
06:18 PM lcnc-relay: <skunkworks8841@> (art had a version of mach years ago that had jerk control - it never went anywere)
06:29 PM pcw-at-work: The jerk limited stuff on the forum did not obey per joint constraints so a bit picky with real hardware
06:33 PM pcw-at-work: (that is it only had velocity/accel/jerk constraints for the resultant 3d motion vector)
08:48 PM Tom_L: golfball size hail here
08:57 PM rdtsc: hope for no damage
09:38 PM lfluffywof is now known as fluffywolf
10:08 PM xxcoder: ugh smokey hellscope season sucks. even if minor one this year
10:11 PM fluffywolf: so far.
10:16 PM fluffywolf: I should be doing useful things, but I'm grumpy and ate way too much for dinner, so instead I'm laughing at bowsette memes.
11:50 PM fluffywolf: grrr, non-standard threads... it's a #8 but super fine. might be finer than 8-36.
11:50 PM fluffywolf: not going to bother finding my thread gauges, because whatever it is, the hardware store isn't going to have it anyway. lol