|
From: Florian R. <fr...@fs...> - 2009-12-01 15:06:01
|
Hi, For the first time I have to mill a precise aluminum part and discovered that the x and y axes of our CNC mill are not properly aligned at an angel of 90° but show a deviation of about 10". Is it possible to correct this misalignment in software using EMC or Mach3 (both soft CNCs are used here)? Thanks Florian |
|
From: Jeff E. <je...@un...> - 2009-12-01 15:56:30
|
An emc user has contributed a kinematics module to correct for XY skew. http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ContributedComponents#millkins_trivial_kinematics_extended_by_XY_skew_correction I have not personally used this kinematics module. I think emc's main limitation when using kinematics to correct these kinds of machine imperfections is that jogs are not corrected--a "Y" jog will only turn the "Y" motor, not both motors together. Jeff |
|
From: Andy P. <an...@an...> - 2009-12-01 16:29:47
|
2009/12/1 Jeff Epler <je...@un...>: > I think emc's main limitation when using kinematics to correct these > kinds of machine imperfections is that jogs are not corrected--a "Y" jog > will only turn the "Y" motor, not both motors together. Do the coordinates of the non-jogged axes update correctly? If so it seems only a small problem. If not, a potentially huge one. -- atp |
|
From: Andy P. <an...@an...> - 2009-12-01 16:00:25
|
2009/12/1 Florian Rist <fr...@fs...>: > Is it possible to correct this misalignment in software using EMC or > Mach3 (both soft CNCs are used here)? Such a situation is specifically mentioned here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Kinematics Though I haven't found an existing kinematics model mentioned in the Wiki or manuals that obviously supports skewed axes. -- atp |
|
From: Rainer S. <lem...@gm...> - 2009-12-01 16:23:54
|
I'd shift it in the CD software and run the CAM again.... On Tue, Dec 1, 2009 at 9:50 AM, Florian Rist <fr...@fs...> wrote: > Hi, > For the first time I have to mill a precise aluminum part and discovered > that the x and y axes of our CNC mill are not properly aligned at an > angel of 90° but show a deviation of about 10". > > Is it possible to correct this misalignment in software using EMC or > Mach3 (both soft CNCs are used here)? > > Thanks > Florian > > > ------------------------------------------------------------------------------ > Join us December 9, 2009 for the Red Hat Virtual Experience, > a free event focused on virtualization and cloud computing. > Attend in-depth sessions from your desk. Your couch. Anywhere. > http://p.sf.net/sfu/redhat-sfdev2dev > _______________________________________________ > Emc-users mailing list > Emc...@li... > https://lists.sourceforge.net/lists/listinfo/emc-users > -- _________________________________ Rainer M. Schmidt Complex Consulting LLC Bo...@co... Voice (646)-657-8815 FAX (646)-435-9216 |
|
From: Stuart S. <st...@gm...> - 2009-12-01 16:38:40
|
Florian, EMC2 can do that. I corrected the perpendicularity of the linear axes and also the misalignment of the rotary axes A and B. The X, Y and Z axes are not mechanically orthogonal. The axis of rotation of the A axis is not exactly parallel to the X axis. The axis of rotation of the B axis is not exactly parallel to the Y axis. I have a kinematics file that does these corrections. Under this directory http://www.mpm1.com:8080/cinci/cinci_running you will find the files that configure my Cincinnati 5 axis hydrotel retrofit. The cincikins.c file has the orthogonal (what I call skew) correction. Look in the file for skew. The gear4.comp is for the 4 speed gearbox on the spindle. Try this: Grab your left earlobe between your left thumb and left index finger. Grab you right earlobe between your right thumb and right index finger. You now have the limitations of EMC2 between your hands. :) Have fun Stuart On Tue, Dec 1, 2009 at 8:50 AM, Florian Rist <fr...@fs...> wrote: > Hi, > For the first time I have to mill a precise aluminum part and discovered > that the x and y axes of our CNC mill are not properly aligned at an > angel of 90° but show a deviation of about 10". > > Is it possible to correct this misalignment in software using EMC or > Mach3 (both soft CNCs are used here)? > > Thanks > Florian > > > > ------------------------------------------------------------------------------ > Join us December 9, 2009 for the Red Hat Virtual Experience, > a free event focused on virtualization and cloud computing. > Attend in-depth sessions from your desk. Your couch. Anywhere. > http://p.sf.net/sfu/redhat-sfdev2dev > _______________________________________________ > Emc-users mailing list > Emc...@li... > https://lists.sourceforge.net/lists/listinfo/emc-users > > -- The value of goods are expressed in money, while the value of money is expressed in goods. Money and goods are clearly not the same things, but are exactly opposite things. Goods are wealth which you have, while money is a claim on wealth which you do not have. Goods are an asset; Money is a debt. If goods are wealth; money is clearly not wealth, it is negative wealth, maybe even anti-wealth. – Quigley, Tragedy and Hope, pg. 44 (emphasis added) |
|
From: Kirk W. <kwa...@wa...> - 2009-12-01 17:39:56
|
On Tue, 2009-12-01 at 11:23 -0500, Rainer Schmidt wrote: > I'd shift it in the CD software and run the CAM again.... > At first, I agreed with you, but now I'm thinking, skew would be better fixed on the machine. You would correct the problem once and be done. With the skew correction in the CAM, the issue would need to be addressed with each part file. This also brings up whether it's best to handle mutili-part setups (which may need translation and/or rotation for each part to maximize material use) in CAM or the machine. I guess it all depends on whether you're a hobby shop or production (either of which are equally good). -- Kirk Wallace http://www.wallacecompany.com/machine_shop/ http://www.wallacecompany.com/E45/index.html California, USA |
|
From: BRIAN G. <gla...@gm...> - 2009-12-01 19:19:22
|
I am with Kirk. Months ago I played with the gantrykins to correct mechanical skew on my router gantry. After several days of fiddling, I did get it to work, but one comment from a email response suggesting I just dive in and correct the skew at its source stuck with me. I loosened the gantry up and restarted the alignment procedure on my machine. 30 minutes later, the skew was below my measurement ability. Squares returned. Go back and ensure you are certain you cannot eliminate it mechanically first. Only after failing to eliminate skew mechanically would I recommend using the software fix. $0.02 Brian |
|
From: Florian R. <fr...@fs...> - 2009-12-01 21:50:06
|
Hi Brian > Go back and ensure you are certain you cannot eliminate it mechanically > first. Of course that was the first think I tried, but it's completely miss aligned and would be quite a hassle to aligned it mechanically. I have a deviation of 0.3mm on every 100mm. By the way, it's not only the angel between x and y axis that is not 90.0° but also the one between the x/y plane and the z axis and main spindle motors axis´ is not parallel to the z movement, too. I want a new machine... :-( See you Florian |
|
From: Stephen W. P. <sp...@so...> - 2009-12-01 22:21:30
|
Florian Rist wrote: >Hi Brian > > >>Go back and ensure you are certain you cannot eliminate it mechanically >>first. >> >> >Of course that was the first think I tried, but it's completely miss >aligned and would be quite a hassle to aligned it mechanically. I have a >deviation of 0.3mm on every 100mm. > >By the way, it's not only the angel between x and y axis that is not >90.0° but also the one between the x/y plane and the z axis and main >spindle motors axis´ is not parallel to the z movement, too. I want a >new machine... :-( > > Um. Are you saying that the spindle is not perpendicular to the XY plane? If that's so, then no amount of software correction will allow you to drill a perpendicular hole. This is because software correction only corrects one point - the tool tip. The rest of the tool is expected to be directly above the tip (on a mill), which it isn't on your machine. So the tool tip will go straight down, by moving X/Y to compensate Z, and the top of the hole will be a slot. This is a very bad situation. - Steve |
|
From: Rob J. <ro...@my...> - 2009-12-02 08:13:12
|
Florian, I agree with Steve. If your Z-axis movement is not perpendicular to the XY plane you must fix this. The same for spindel alignment, that should also be perpendicular to the XY plane. Is this a home built machine? Any chance of changing it such that you can get rid of the misasalignment? That will certainly save you a lot of problems later on. Any chance of showing (larger) photos of your machine - maybe some of us have some good ideas. But bear in mind, it is only my third machine that will allow me to do the alignment in a very easy way without special tools (and I am currently building my second one ...) Rob Florian Rist wrote: > Of course that was the first think I tried, but it's completely miss > aligned and would be quite a hassle to aligned it mechanically. I have a > deviation of 0.3mm on every 100mm. > > By the way, it's not only the angel between x and y axis that is not > 90.0° but also the one between the x/y plane and the z axis and main > spindle motors axis´ is not parallel to the z movement, too. I want a > new machine... :-( > |
|
From: Florian R. <fr...@fs...> - 2009-12-01 21:50:23
|
Hi Jeff thank for you answer. > An emc user has contributed a kinematics module to correct for XY skew. > http://wiki.linuxcnc.org/cgi- bin/emcinfo.pl?ContributedComponents#millkins_trivial_kinematics_extended_ by_XY_skew_correction Great, that´s what I was looking for. I was easy to install and does the job very well. I think I´m going to extend it to correct a tilt of the machine table as well. > I think emc's main limitation when using kinematics to correct these > kinds of machine imperfections is that jogs are not corrected--a > "Y" jog will only turn the "Y" motor, not both motors together. In deed, that´s somewhat uncomfortable and misleading, but never the less the coordinate readout you while jogging is correct, so it´s not limitation. See you Flo |
|
From: Florian R. <fr...@fs...> - 2009-12-01 21:50:24
|
Hi Andy > > I think emc's main limitation when using kinematics to correct these > > kinds of machine imperfections is that jogs are not corrected--a "Y" > > jog will only turn the "Y" motor, not both motors together. > > Do the coordinates of the non-jogged axes update correctly? If so it > seems only a small problem. If not, a potentially huge one. The coordinates do update correctly, so it´s no a big deal. In fact if you hit the jog y button it makes sense that only the y aces moves. See you Flo -- Vienna University of Technology, 3D Design and Modeling E264/2, Dipl.-Ing. Florian Rist, Karlsplatz 13 E264/2, A-1040 Vienna, Austria, Phone +43-1-58801-26427, Fax +43-1-58801-26490, fr...@fs..., WWW http://e2642.kunst.tuwien.ac.at/ Ö: UID ATU37675002, DVR 0005886 |
|
From: Florian R. <fr...@fs...> - 2009-12-01 21:50:26
|
Hi Rainer > I'd shift it in the CD software and run the CAM again.... I thought about that. It´d be quick and easy but I not a very nice solution and one that would introduce an extra step and therefore an extra chance to make mistakes. The similar, but more robust solution would have been to correct the XY skew in the CAM postprocessor. Still adapting the kinematic module seams to make more sense to me. See you Flo -- Vienna University of Technology, 3D Design and Modeling E264/2, Dipl.-Ing. Florian Rist, Karlsplatz 13 E264/2, A-1040 Vienna, Austria, Phone +43-1-58801-26427, Fax +43-1-58801-26490, fr...@fs..., WWW http://e2642.kunst.tuwien.ac.at/ Ö: UID ATU37675002, DVR 0005886 |
|
From: Florian R. <fr...@fs...> - 2009-12-02 11:41:15
|
Hi Stephen > Um. Are you saying that the spindle is not perpendicular to > the XY plane? Unfortunately that´s what I discovered yesterday. > If that's so, then no amount of software correction will allow > you to drill a perpendicular hole. I know (could only be done on a 5axis machine). > This is a very bad situation. Indeed and I´m quit disappointed with the low precision this machine was assembled. See you Florian -- Vienna University of Technology, 3D Design and Modeling E264/2, Dipl.-Ing. Florian Rist, Karlsplatz 13 E264/2, A-1040 Vienna, Austria, Phone +43-1-58801-26427, Fax +43-1-58801-26490, fr...@fs..., WWW http://e2642.kunst.tuwien.ac.at/ Ö: UID ATU37675002, DVR 0005886 |
|
From: Florian R. <fr...@fs...> - 2009-12-02 11:57:19
|
Hi Rob > Is this a home built machine? No. > Any chance of changing it such that you can get rid > of the misasalignment? I haven´t tried it for the z-axis and the spindle motor, but I hoe so. The misalignment between x and y is to big (3mm at 1000mm length) that it was impossible to correct mechanically. It look like the whole machine table (2000x1000mmm) and the y guides are sheered to a parallelogram. What ever, the xy misalignment is corrected pretty well by the kinematics module. > Any chance of showing (larger) photos of your machine - maybe > some of us have some good ideas. Sure. I´ll take some later that day or tomorrow. It´s a simple 3+1 axis portal router type machine, working area 1700x950x240mm. I probably could correct all errors by adjusting the hardware, but his would most probably require disassembling most of the machine and take a lot of time. Maybe I´ll to that in January. See you Florian -- Vienna University of Technology, 3D Design and Modeling E264/2, Dipl.-Ing. Florian Rist, Karlsplatz 13 E264/2, A-1040 Vienna, Austria, Phone +43-1-58801-26427, Fax +43-1-58801-26490, fr...@fs..., WWW http://e2642.kunst.tuwien.ac.at/ Ö: UID ATU37675002, DVR 0005886 |
|
From: Andy P. <an...@an...> - 2009-12-02 12:50:17
|
2009/12/2 Florian Rist <fr...@fs...>: > I probably could correct all errors by adjusting the hardware, but his > would most probably require disassembling most of the machine and take a > lot of time. Maybe I´ll to that in January. I think that truing up the axes can be seen as part of the installation and commissioning procedure. Does the machine have a twin-drive X axis? It sounds like it is possible that the drives are out of synch, either due to misalignment of the home switches or possibly due to the homing procedure being wrongly specified in EMC. That would twist the Y axis relative to the table. If the drives are linked by a belt, then it might simply be a case of releasing some shaft clamps and running the axis up and down my hand, re-tightening and re-checking. Or you might need to remove the belt, move one leadscrew relative to the other, and reassemble. Z-axis alignment might also turn out to be relatively easy. A dial indicator mounted on the spindle such that it describes a circle on the machine table will give a very good measure of squareness, and will also indicate how your adjustments are working. It is not uncommon to find jacking screws built into the linear rail mounting arrangement, for example (but this might not be relevant to your machine) Given the machine type and the reported problem, I would be concerned that the axes are twisted and "fighting" each other. Is the machine securely bolted to the floor absolutely flat and level? Even traditional cast iron machines can be twisted by uneven floors, and it is more likely to be a problem with a fabricated steel/aluminium frame. A good spirit level should show up any twist (check it both ways on all 4 edges) -- atp |
|
From: Florian R. <fr...@fs...> - 2009-12-02 16:00:53
|
Hi Andy > I think that truing up the axes can be seen as part of the > installation and commissioning procedure. Well, it should be, but in this case the machine was delivered and used productively right away without any in-depth testing and up to now the errors were not noticed because the mill is usually used by students to mill simple parts for architectural model making. > Does the machine have a twin-drive X axis? No, it´s a single steeper motor on the y axis and it´s rotation is transmitted to the two y axis lead screw by a tooth belt. > It sounds like it is possible that the drives are out of synch, ... > If the drives are linked by a belt, then it might simply be a case of > releasing some shaft clamps and running the axis up and down my hand, > re-tightening and re-checking. I tried that, unfortunately I can only move one side by about 1 mm (about 3 mm would be required). > Z-axis alignment might also turn out to be relatively easy. A dial > indicator mounted on the spindle such that it describes a circle on > the machine table will give a very good measure of squareness, and > will also indicate how your adjustments are working. That´s how I measured the misalignment. I also have a coordinate measuring arm at hand; it might be useful when I deice to take apart larger parts of the machine to cure all misalignments. > It is not uncommon to find jacking screws built into the > linear rail mounting arrangement, for example (but this might > not be relevant to your machine) I don´t´ see any, unfortunately. > Given the machine type and the reported problem, I would be > concerned that the axes are twisted and "fighting" each other. It looks like the table is deformed from rectangular to a parallelogramlike shape. The table is just a bolted aluminum construction and it´s lacking a diagonal element to stabilize the rectangular shape. > Is the machine securely bolted to the floor absolutely flat > and level? Yes. > Even traditional cast iron machines can be twisted by uneven > floors, and it is more likely to be a problem with a fabricated > steel/aluminium frame. A good spirit level should show up > any twist (check it both ways on all 4 edges) I did when the machine was installed, even tough I didn´t have a high precision spirit level. The machine should be horizontal with a maximum deviation of 0.5 mm per meter. Good enough? See you Florian |
|
From: Andy P. <an...@an...> - 2009-12-02 16:33:46
|
2009/12/2 Florian Rist <fr...@fs...>: >> If the drives are linked by a belt, then it might simply be a case of >> releasing some shaft clamps and running the axis up and down my hand, >> re-tightening and re-checking. > > I tried that, unfortunately I can only move one side by about 1 mm (about > 3 mm would be required). Sounds promising, though. Can you post some pictures somewhere we the wild speculation is a little more rooted in reality? If it is a bolted-up gantry then it might be possible to slacken some things off to give more freedom there (most probably the gantry slide bearings on the gantry, or the side columns to the upper (X?) axis. > I did when the machine was installed, even tough I didn´t have a high > precision spirit level. The machine should be horizontal with a maximum > deviation of 0.5 mm per meter. Good enough? It is more important that both long edges slope by the same amount so that the bed is flat. In principle I think you could work out the deviation from flatness by seeing if a pencil on a (massless, frictionless, inelastic) string tied to the ceiling above the table draws a perfect circle on the table. The smaller the included angle of the string, the more a small deviation from flat will distort the circle. Replacing the pencil with a dial indicator and marking the points where the reading is the same would be more precise. However, I don't know if this would work in practice, it is a thought-experiment based on geometrical considerations much more than it is anything I have ever heard of being tried. -- atp |
|
From: Jon E. <el...@pi...> - 2009-12-02 18:21:11
|
Andy Pugh wrote: > 2009/12/2 Florian Rist <fr...@fs...>: > > >>> If the drives are linked by a belt, then it might simply be a case of >>> releasing some shaft clamps and running the axis up and down my hand, >>> re-tightening and re-checking. >>> >> I tried that, unfortunately I can only move one side by about 1 mm (about >> 3 mm would be required). >> Right, the linear rails were aligned with this offset of the screws. So, you need to loosen where the sliders attach, align the leadscrews, then tighten the linear slides. I am assuming this machine is a bolted-together gantry router design. If this is a cast iron dovetail slideway machine, then something slipped when the ways were machined, and you are right, it is very hard to fix. |
|
From: Jon E. <el...@pi...> - 2009-12-02 18:13:41
|
Florian Rist wrote: > No, it´s a single steeper motor on the y axis and it´s rotation is > transmitted to the two y axis lead screw by a tooth belt. > Well, it may be possible to loosen one of the pulleys and turn the leadscrew until the gantry is perpendicular to the Y travel. It might be necessary to file a new flat for the pulley's setscrew to bear on to prevent slipping. For coarse adjustment, you might be able to loosen the belt and jump teeth until it is as close as you can get it, then make finer adjustments with the above procedure. It may then be necessary to loosen some of the structure around the rail sliders to allow them to adapt to the new alignment without binding. Jon |
|
From: <st...@gm...> - 2009-12-02 16:26:38
|
Tramming the spindle will tell you how perpendicular the spindle centerline is to that particular spot on the table surface. It will not tell you the perpendicularity of the spindle centerline (or the Z axis) to the XY plane. The actual level of the machine is almost irrelevent. If all the machine components are fixed at the same level values and the machine is straight and square you will be able to machine straight and square. Sent via BlackBerry from T-Mobile -----Original Message----- From: "Florian Rist" <fr...@fs...> Date: Wed, 02 Dec 2009 17:04:26 To: Andy Pugh<an...@an...>; Enhanced Machine Controller \EEMC\"<emc...@li...> Subject: Re: [Emc-users] How to Correct Angular Axes Misaligment in Software? Hi Andy > I think that truing up the axes can be seen as part of the > installation and commissioning procedure. Well, it should be, but in this case the machine was delivered and used productively right away without any in-depth testing and up to now the errors were not noticed because the mill is usually used by students to mill simple parts for architectural model making. > Does the machine have a twin-drive X axis? No, it´s a single steeper motor on the y axis and it´s rotation is transmitted to the two y axis lead screw by a tooth belt. > It sounds like it is possible that the drives are out of synch, ... > If the drives are linked by a belt, then it might simply be a case of > releasing some shaft clamps and running the axis up and down my hand, > re-tightening and re-checking. I tried that, unfortunately I can only move one side by about 1 mm (about 3 mm would be required). > Z-axis alignment might also turn out to be relatively easy. A dial > indicator mounted on the spindle such that it describes a circle on > the machine table will give a very good measure of squareness, and > will also indicate how your adjustments are working. That´s how I measured the misalignment. I also have a coordinate measuring arm at hand; it might be useful when I deice to take apart larger parts of the machine to cure all misalignments. > It is not uncommon to find jacking screws built into the > linear rail mounting arrangement, for example (but this might > not be relevant to your machine) I don´t´ see any, unfortunately. > Given the machine type and the reported problem, I would be > concerned that the axes are twisted and "fighting" each other. It looks like the table is deformed from rectangular to a parallelogramlike shape. The table is just a bolted aluminum construction and it´s lacking a diagonal element to stabilize the rectangular shape. > Is the machine securely bolted to the floor absolutely flat > and level? Yes. > Even traditional cast iron machines can be twisted by uneven > floors, and it is more likely to be a problem with a fabricated > steel/aluminium frame. A good spirit level should show up > any twist (check it both ways on all 4 edges) I did when the machine was installed, even tough I didn´t have a high precision spirit level. The machine should be horizontal with a maximum deviation of 0.5 mm per meter. Good enough? See you Florian ------------------------------------------------------------------------------ Join us December 9, 2009 for the Red Hat Virtual Experience, a free event focused on virtualization and cloud computing. Attend in-depth sessions from your desk. Your couch. Anywhere. http://p.sf.net/sfu/redhat-sfdev2dev _______________________________________________ Emc-users mailing list Emc...@li... https://lists.sourceforge.net/lists/listinfo/emc-users |
|
From: Andy P. <an...@an...> - 2009-12-02 17:05:35
|
2009/12/2 <st...@gm...>: > > Tramming the spindle will tell you how perpendicular the spindle centerline is to that particular spot on the table surface. It will not tell you the perpendicularity of the spindle centerline (or the Z axis) to the XY plane. Good point, I am used to machines where that can be assumed. Not the case with a fabricated gantry router. Distance from fully-retracted spindle to table at various points will tell you how the table aligns with the true X/Y plane. There might be a work sequence somewhere on the interweb which runs through the optimum sequence for aligning such a machine. I have never even seen one in-person. (I will be putting some of all this in practice this weekend, we are moving my dad's machine tool collection from the temporary to permanent workshop. http://www.lathes.co.uk/smartbrown/page6.html http://www.lathes.co.uk/colchester/page14.html http://www.zhongyang-engrg.com/photos/se8089.jpg http://www.scottmachinery.co.nz/dbimage.php?meth=norm&img=BORE-1-0003__1 Likely to be a long hard weekend. Even the saddle of the borer is the sort of mass you can't even move by hand.) -- atp |
|
From: Jon E. <el...@pi...> - 2009-12-02 18:17:58
|
st...@gm... wrote: > Tramming the spindle will tell you how perpendicular the spindle centerline is to that particular spot on the table surface. It will not tell you the perpendicularity of the spindle centerline (or the Z axis) to the XY plane. > The actual level of the machine is almost irrelevent. If all the machine components are fixed at the same level values and the machine is straight and square you will be able to machine straight and square Right, I have a procedure for this. I mill a circular track in a piece of scrap. Then I move the spindle to the center of that circle, and sweep it with a dial indicator. Unfortunately, my machine has wear on all the ways, and so I get a horse saddle shape, with an upward curve in one axis and a downward curve in the other. But, sweeping it with the indicator, I can split the difference and get the spindle as close to perpendicular to the "average" of the X-Y plane. Jon |
|
From: Gene H. <gen...@gm...> - 2009-12-02 17:00:06
|
On Wednesday 02 December 2009, st...@gm... wrote: > Tramming the spindle will tell you how perpendicular the spindle > centerline is to that particular spot on the table surface. It will not > tell you the perpendicularity of the spindle centerline (or the Z axis) to > the XY plane. The actual level of the machine is almost irrelevent. If all > the machine components are fixed at the same level values and the machine > is straight and square you will be able to machine straight and square. > Sent via BlackBerry from T-Mobile > >-----Original Message----- >From: "Florian Rist" <fr...@fs...> >Date: Wed, 02 Dec 2009 17:04:26 >To: Andy Pugh<an...@an...>; Enhanced Machine Controller > \EEMC\"<emc...@li...> Subject: Re: [Emc-users] How to > Correct Angular Axes Misaligment in Software? > >Hi Andy > >> I think that truing up the axes can be seen as part of the >> installation and commissioning procedure. > >Well, it should be, but in this case the machine was delivered and used >productively right away without any in-depth testing and up to now the >errors were not noticed because the mill is usually used by students to >mill simple parts for architectural model making. > >> Does the machine have a twin-drive X axis? > >No, it´s a single steeper motor on the y axis and it´s rotation is >transmitted to the two y axis lead screw by a tooth belt. > >> It sounds like it is possible that the drives are out of synch, ... >> >> If the drives are linked by a belt, then it might simply be a case of >> releasing some shaft clamps and running the axis up and down my hand, >> re-tightening and re-checking. > >I tried that, unfortunately I can only move one side by about 1 mm (about >3 mm would be required). Sounds like its belt removal time so you can slip it a tooth or 4 on just one shaft. Or would that much actually cause a bind in the carriage? >> Z-axis alignment might also turn out to be relatively easy. A dial >> indicator mounted on the spindle such that it describes a circle on >> the machine table will give a very good measure of squareness, and >> will also indicate how your adjustments are working. > >That´s how I measured the misalignment. I also have a coordinate >measuring arm at hand; it might be useful when I deice to take apart >larger parts of the machine to cure all misalignments. > >> It is not uncommon to find jacking screws built into the >> linear rail mounting arrangement, for example (but this might >> not be relevant to your machine) > >I don´t´ see any, unfortunately. > >> Given the machine type and the reported problem, I would be >> concerned that the axes are twisted and "fighting" each other. > >It looks like the table is deformed from rectangular to a >parallelogramlike shape. The table is just a bolted aluminum construction >and it´s lacking a diagonal element to stabilize the rectangular shape. Good Grief. OTOH, TSC et all sells all-thread & pump rod couplings in case the all thread isn't long enough. How about loosening the assembly bolts and pulling it into square with a long 'pony' clamp on a suitable length of pipe, then re-tighten the assembly bolts? Or even make an x brace from 3/4" square steel tubing. Pull it square with the pony clamp, then clamp the x brace in place under the table, drill and bolt in permanently. Super glue before you drill to prevent slippage. And is there room under the frame to install the x bracing all-thread permanently? No use doing the corrections every week. :( And since its students, I think once the x bracing was verified square, I'd super glue the all-threads adjusting nuts if its left in place. In any group of 'students' there will always be one that thinks its funny to subtly screw something up. [...] -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) The NRA is offering FREE Associate memberships to anyone who wants them. <https://www.nrahq.org/nrabonus/accept-membership.asp> senility, n.: The state of mind of elderly persons with whom one happens to disagree. |