Re: [Emc-users] G33.1 error then g-code file parsed
Using the last version 2.10.0~pre0 this problem is solved so time to pass this discussion to the trashbin. Should of course have checked the latest version before starting the discussion. Nicklas Karlsson ons 2023-06-14 klockan 12:27 -0700 skrev John Dammeyer: > Does your spindle have an encoder and the spindle at speed? > > https://forum.linuxcnc.org/38-general-linuxcnc-questions/43124-g33-1-rigid-tapping > > > -Original Message- > From: Nicklas SB Karlsson [mailto:n...@nksb.eu] > Sent: June 14, 2023 12:14 PM > To: emc-users@lists.sourceforge.net > Subject: [Emc-users] G33.1 error then g-code file parsed > > I put the lines below into a file: > > � M3 S100 > � G33.1 Z-30.474 K0.8 I3.000 > � M5 > � M2 > > Then I read into Linuxcnc I get error message: > > � parse_file interp_error > > Removing the line with G33.1 then no error message so it is something > with this row. Program do however execute as expected with G33.1 line > even though there is an error message so no real problem. Also > execute > without an error message if run manually in MDI mode. Use > origin/master > last commit Mon May 8 16:10:03 2023 +0200 > 404aa407f136ce91a3e6bf911c7bda54011a74e9 Anybody else had similar > problem? > > > Nicklas Karlsson > > > > ___ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > > > > ___ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Outside of machine limit bug? --> Partly fixed in latest version
Of course I should have updated to latest version before posting but at least have done this now. Update to the latest version 2.10.0~pre0 Date: Mon Jan 1 22:58:27 2024 +0100 the problem have been solved in z direction for different coordinate systems and partly for x,y direction. Use G10 L2 P0 moving G54 coordinate checking several times it seems machine and relative position is mixed up for (x,y) axis until file is reloaded. Re-Check button appear if G-code is outside machine limits then file is loaded. But load G-code and then changed the coordinate system which is usually done then part is touched no new check is done until machining is stopped because a soft limit is hit. Think problem is the same for tool length compensation. Nicklas Karlsson tis 2023-12-12 klockan 21:34 -0500 skrev gene heskett: > On 12/12/23 17:39, Nicklas SB Karlsson wrote: > > Have a quite large part compared to the machine there I put the > > zero > > point in the lower left corner. As machine limits for (x, y) > > movements > > are zero in the middle an offset is added then touch off part. > > Pretty > > sure I got it right and part is within machine limit but check fail > > complain axis is outside of machine limit. > > > > Then an offset is added either by G10 L2 or by touch off button > > machine coordinates and G-code coordinates does not agree > > anymore. So > > problem might be Linuxcnc check the G-code only but do not add the > > offset set during touch off or of simply make the calculations > > wrong. > > > > There also seems to be a problem with the minimum and maximum > > limits > > shown in the preview view. Then no offset is added machine and > > relative > > position is same and everything is OK. But then an offset is added > > by > > G10 L2 or touch off it seems selecting machine position in view > > menu > > show coordinates in G-code which must be wrong then an offset is > > added > > to the G-code. Selecting relative coordinates show other > > coordinates > > with same distance but they are neither the numbers from the G-code > > nor > > the absolute position if I maded the calculation correct. > > > > Anybody else had similar problem? > > > Those limits Nik, set in the .ini file, are relative to the g53 > coordinate map, or are supposed to be. That's so they don't move > physically when the mapping is changed by a touch off or whatever. > > The g53 map's zero position can be anyplace, but I usually set zero > near > the center of the envelope regardless of home switch locations which > are > generally just a few thou from mechanical end limits then offset back > to > near the center for the zero point, but that for me is just a > personal > pref as that allows me to use the home switch as a that end of travel > limit switch in the .hal file, with the other end just a software > limit > in the ini file. Saves me one or 3 pins on the end bob. > > It can be confusing too if the saved vars lists in > linuxcnc/configs/your-config/position.txt or linuxcnc.var file are > contaminated, so you might want to rm them, restart linuuxcnc and > rehome > to remove that potential src of the machines confusion. > > You also didn't note which version you're using. Andy will want to > know. > > I run master on all my wintel boxes and haven't noted a problem but > haven't really exercised either of them in about a month as I'm > fiddling > with final trimming of a vise screws production on a 3d printer farm > I'm > also re-building. Making the screw has now become a 2 machine process > as > the rough carving has been moved to the go704 from the 6040, taking > me > from 2 days to make one screw to 2 a day. I made an A axis for the > go704 which is about 300x faster at the rough rounding of a square > stick. An A axis that can spin 500 rpm with arcminute or better > accuracy. > > > > > Regards Nicklas Karlsson > Take care, stay warm and well. > Cheers, Gene Heskett. ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Capto
On 1/4/24 10:08, andy pugh wrote: On Thu, 4 Jan 2024 at 13:39, gene heskett wrote: Humm, that may call for a higher resolution spindle encoder than I have on my Sheldon, I have a resolver on the spindle of my lathe, which the Mesa 7i49 converts into a 14-bit number, so should have enough resolution. (linearity of the resolver is not something that I have ever tested) The linearity may not be the most important characteristic, although in a lathe where the chuck is relatively massive it shouldn't matter that much for steadily turning use, but in for instance a 3d printer, all those conversions take time, which has bad effects on the nyquist stability of the feedback loop regulating that speed. in the lathe case it might mean a following error because by the time that position is known, the spindle has turned further. This is not a problem with a quadrature encoder as the position and direction is known fresh with every edge that goes by. No time lag. About 3 years ago one of the Chinese motor makers came out with the ideal sized nema 17 motor, using a hall effect as the feedback generator, while the hall effect can work at several times the optical accuracy its always working with stale data, so it jerks to catchup. Everyone who put the 42C motors on their printers, took them off the next day as their prints looked like they were shingled. That motor is a good motor, for something that needs a programmed position and stopped. Great for a smaller tool changer. The same motor with an optical encoder moves so smooth I had to grab the printhead to verify klippers "stepper_buzz" motor tester was indeed moving the motor 1mm back and forth. Originally intro-ed at about $35 to $40 USD, now they can't give them away at $9.95 But I don't quite understand why you are worrying about whether you can make Capto shanks on your Sheldon? Take care, stay warm and well. Cheers, Gene Heskett. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Capto
On 1/4/24 10:08, andy pugh wrote: On Thu, 4 Jan 2024 at 13:39, gene heskett wrote: Humm, that may call for a higher resolution spindle encoder than I have on my Sheldon, I have a resolver on the spindle of my lathe, which the Mesa 7i49 converts into a 14-bit number, so should have enough resolution. (linearity of the resolver is not something that I have ever tested) But I don't quite understand why you are worrying about whether you can make Capto shanks on your Sheldon? Like some other things I've done Andy, just to prove I can do it. That's why my Sheldon is being run by a pi. And run very well. I'm not a working shop, just a hobbiest (with poor math skills) in search of the knowledge. Take care, stay warm and well. Cheers, Gene Heskett. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Capto
We had gotten some 5000 line - 2 count encoders on ebay a while back. It certainly helped.. Higher would be nicer.,.(for bigger diameter testing) sam On Thu, Jan 4, 2024 at 9:08 AM andy pugh wrote: > On Thu, 4 Jan 2024 at 13:39, gene heskett wrote: > > > Humm, that may call for a higher resolution spindle encoder than I have > > on my Sheldon, > > I have a resolver on the spindle of my lathe, which the Mesa 7i49 > converts into a 14-bit number, so should have enough resolution. > (linearity of the resolver is not something that I have ever tested) > > But I don't quite understand why you are worrying about whether you > can make Capto shanks on your Sheldon? > > -- > atp > "A motorcycle is a bicycle with a pandemonium attachment and is > designed for the especial use of mechanical geniuses, daredevils and > lunatics." > — George Fitch, Atlanta Constitution Newspaper, 1912 > > > ___ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Capto
On Thu, 4 Jan 2024 at 13:39, gene heskett wrote: > Humm, that may call for a higher resolution spindle encoder than I have > on my Sheldon, I have a resolver on the spindle of my lathe, which the Mesa 7i49 converts into a 14-bit number, so should have enough resolution. (linearity of the resolver is not something that I have ever tested) But I don't quite understand why you are worrying about whether you can make Capto shanks on your Sheldon? -- atp "A motorcycle is a bicycle with a pandemonium attachment and is designed for the especial use of mechanical geniuses, daredevils and lunatics." — George Fitch, Atlanta Constitution Newspaper, 1912 ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Capto
On 1/3/24 18:54, andy pugh wrote: On Wed, 3 Jan 2024 at 22:01, andy pugh wrote: This means that the large and small arcs are tangent to each other, which requires that their centres lie on a perpendicular line, and that in turn means that the arc centres of small and large arcs must be coincident. Close, but no cigar, it seems. There is an equation in an online document (https://archiwum-bazakonkurencyjnosci.funduszeeuropejskie.gov.pl/file/download/587886) that isn't actually correct: x'= Dm/2×cosØ −2×e×cos(2Ø)+e×cos(4Ø) y'= Dm/2×sinØ -2×e×sin(2Ø)+e×sin(4Ø) However, after getting hold of a pirate copy of the standard, it turns out that it was only slightly wrong. This is what it should be. ( -+++ not -+-+) x'= Dm/2×cosØ −2×e×cos(2Ø)+e×cos(4Ø) y'= Dm/2×sinØ +2×e×sin(2Ø)+e×sin(4Ø) Plotted out this is very close to the shape that I had but differs by about 30 microns at the worst spots. The good news is that this is very easy to convert to either XY offsets for boring, or a radius for computed offset on a lathe as it is already in terms of spindle angle Ø. Humm, that may call for a higher resolution spindle encoder than I have on my Sheldon, I'm watching the 60 tooth gear on the spindle si my ultimate accuracy given perfct gears is 1.5 degrees, and that is noisy as hell because of the wear from 75 years of back gear over engagement which has rounded the taller teeth quite noticeably. I would not swear that any one edge is any closer than two degrees because of this wear. The vfd makes a good filter. I'd bet a $20 you could see the teeth as a pattern in an otherwise good grind. Letting my mind out to play w/o a chaperone, a 1000 line Omron encoder for $20.00 shipped on fleabay, driven by printed 1/1 timing sprockets with the driving gear pressed onto the smaller gear that is now unused on the back of the spindle would give a .36 degree accuracy. I just looked to see whats in the way, nothing as thats all been removed. Plenty of room for a 6mm belt sprocket, as actually there's 2 gears, one on each side of the rear bearing. And the outside one is totally unused now. This might be a good excuse to put a higher resolution encoder on my Sheldon. And teach it yet another new trick. The wheels are spinning, but need lubed, at 89 they are getting rusty. :o)> Thanks Andy. -- atp "A motorcycle is a bicycle with a pandemonium attachment and is designed for the especial use of mechanical geniuses, daredevils and lunatics." — George Fitch, Atlanta Constitution Newspaper, 1912 ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users Cheers, Gene Heskett. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
