Have you tried making sure everything is loose? For your issue you would want to loosen everything and only tighten the long 5" bolt. Tightening bolts moves the tubes away from them.
What exact dimensions are you tubes? Your gantry is so far off I it seems like you either build the wrong parts size or your printer needs to be checked.
This is going to sound crazy, but are you sure that the square is square?
To test squareness of the square, place an edge against a straight edge of the table and draw a line. Next, flip the square (like a book) so that the same edge is on the same straight edge you used above, but on the other side of the line you drew. Now try to line the square with the line you drew, do they fit perfectly over one another? If so your square is square. If not, your square is off and you’re fighting the same battle I was last week.
Good luck
here’s an article with a GIF at the very top that shows what I’m talking about:
https://www.familyhandyman.com/tools/check-if-a-square-is-square/
Thanks you for your quick response ! My Z axis is not twisted and seem to be parallel : only 0.1mm between top width and bottom width. I don’t know if it’s significant
Thanks you. I will read this post but I want correct the mechanic if possible
It’s basic but it’s important indeed ! I tested the squareness of the square and it’s ok
Yes, bolts A and C are very loose (nylon just engaged). But if I put Z axis, I can’t move (or with difficult) these bolts by hand. Dimensions tube are 25mm (real dimensions between 24.93 and 25mm), I’ve print “25mm OD” files. What do you mean by “gantry is so far off” ?
I’ve printed a 40mm test cube. Real dimensions of this one : Z=40mm, X=40.15mm, Y=39.95mm
Can you give me important dimensions of XY and XYZ piece to verify these 4 printed pieces ?
Check the diagonals (edge to edge) (corner to corner)to see if your printer is square, the cheap printers have been known to be very far off.
The error on your angle is huge, and you are saying everything is loose and you snugged the 5" bolt something is wrong. I could not make mine that far off if I tried. Unfortunately this is not easy to figure out of you made the parts yourself. I have to assume the prints are bad if you are sure the rails are right.
I’ve had the same issue as you, with the same direction on the tilt. I reprinted parts many times, switched the parts placing, and so on. My pipes are fairly round and straight as well. I’ve now finished the mpcnc, its obviously out of square still - but the drawings are surprisingly square!! I don’t understand anything. People are saying that it’s good to use it for a while and let it "settle ". I hope that my machine is square enough, and at one point I’ll add end stops to improve it further.
yes this!!!
Listen mate. I was so worried about so many details in the prints. So many small inconsistencies and mysteries. In the end I gave up - and just assembled the whole animal. I don’t know if I ever will figure out if my prints were good enough, the models not correct, or if atmospheric conditions influenced. When I look back - the main experience I’ve earned is: get it dirty and try it out. Learn to use it, and build it over again, if things seem too bad. Good luck!!
I make sure the plate surface is not of any importance, the only place I even use it is on the locks and there I recommend glue to take care of any surface issues. For those parts you are actually more worried about them being vertical but that is hard to actually test. Super easy to test your printer by printing a very very large cube and checking all the diagonals.
Turbinbjorn is right just get it dirty, you have the parts use them. if you need other parts later, revisit them. Maybe I will have new ones by then.
Hello,
I don’t want to make a new troubleshooting point, because I have the same problem.
I’m German, but I’ll try to describe the facts in english, sorry for the errors in my text.
My MPCNC has the overall dimensions 800x930mm, I’m supporting the middle of all tubes, they are 25mm type. In the past I disconnected the supporting, but no change (I want to say, they create not the problem).
The dimensions in the diagonals are 101,6cm and 101,7cm (measured at the upper housing corner, near to the tubes), so only 1mm difference that means lower than 0,1%, for my opinion is this really good.
The squaring-error is approxomately 5mm between Y0 and Y1, and 10mm between X0 and X1 (hope you understand what I mean, X=930mm, Y=800mm). This is an other world when I see the difference in the diagonals.
I had the hope, that the belts will help to reduce these errors, but no change.
When the middle assembly is free, I made some checks with the integrated X- and Y-tubes over the complete length, and the squaring looks good.
When I integrate all parts then I get the squaring problem, it’s not necessary to fix the screws on the rollers (I made in the past a lot of things, and today I started again to find the error). In the past I 'd checked that I build the roller in the correct direction, so now the X-axis has on one side the standard and on the other side the mirror-roller, same on the Y-axis. Maybe there are some more aspects I could check here, but I found nothing in the docs.
The tension Bolts A are assembled but no force, only the bearing wheel has contact.
Today I disassembled the Z-axis, opened all screws from the X- and the Y-axis at the rollers, opened the 130mm screw and took a look if the big XY-burly-parts are to small (because of the 3D-printer), I found in a german forum this point. I had no idea to measure, so my idea was, when these parts are to small and I’ll open all screws, the lower tube must go up a little bit. But there was no tolerance in the X- or Y-tube. Nevertheless I integrated a washer between the two XY-burly-parts, at first thin one 0,4mm plastic, then 1mm metal and at the end 1,2mm metal. I saw with the last washer, that this was too much because the upper tube pressed up in the roller. For my opinion, it was not the right way, so now all washers are removed.
I’m not sure in which direction I can force the squaring with the 130mm screw, for my opinion only in the wrong direction.
In the moment I have no idea what to do. Normally I’m ready with the mechanic and the cable management (no force on the Z-axis with a dragchain), and in the next step I have to connect the cables. I prepared the double-endstop-solution (mechanic, switch, cable), but I don’t like to force the mechanic over this way. My plan was to use only 3 stepperdrivers TB6600, and the two X- and Y-motors are in a row. I made a plan to integrate all parts in a housing, but there is no room for 2 more TB6600. In the moment I spend no mechanical work in this housing, and I have the option for changing to a bigger one, but I’m not really interested to do this.
Hopefully, I get some more ideas from you to find the error, or points for solution.
Thank you, and best regards from Braunschweig/Germany
Thomas
The parts are printed with PETG, and I used high infill, I couldn’t say how much, because I used other parameters in CURA, but I’m sure more then 60% infill.
The 101.6cm and 101.7cm are the diagonals of the outside frame, right?
Are the lengths of the two X sides and the two Y sides the same? The two X sides of the outside frame are both 800mm?
I don’t understand, but you haven’t wired up the motors, so I wouldn’t be too concerned yet. When you get the motors wired up, use them to draw a big rectangle, and use the measurements from that to decide how to fix it. The motors will move in lockstep (X and X2 move exactly the same distance) so you just have to start the gantry square and it will stay that way.
Make sure you wire them in serial and not parallel.
Thank you for your help !
I forgot the housing from my tape measure, sorry. 
That means for all datas, 67mm in addition.
X0 and Y0 is a simple definition from my side, that means its the zero point from each axis in my mind.
Diagonal: Housing from the innercorners, in height (Z-direction) of the tube 101,6 and 101,7cm, plus 6,7cm each.
Y0-side: 62,7cm plus 6,7
Ymax-side: 62,8cm plus 6,7
X both sides 76,1cm plus 6,7cm
For my opinion good datas.
And Mr. “Phytagoras” says for the diagnoals nearly the same, okay 0,2mm less. I believe Phytagoras is right and my measuring is not perfect. 
When I drive per hand the middle part (in the middle of the X-axis) to the Y0-point directly to me, and measure on both sides the not used rest-tube on the Y-sides, I have 8-9 mm difference.
When I drive the middle part (in the middle of the Y-axis) to the X0-Point, in my case on the right side (unusual, from my feeling it must be left-side, that is my tribute to the tool carrier), and measure on both side I have 6mm difference.
Your recommendation is to use 3 drivers (that is my plan too), and drive the both X- and Y-motors serial. I saw the plan in the documentation and its clear for me. I have to invert one coil from the second motor.
I only repeat your answer, to show that it is clear for me. When I drive the steppermotors serial, they will make exact the same way, and so I’ll not get a squaring problem. That sounds really good
One more point, sorry, because I read in the past different parameters
I wanted to use these parameters in the TB6600
microstepping X and Y: 1/8
microstepping Z: 1/4
Current for X, Y and Z, I have two options with the switch-settings (no poti): 0,5A (peak 0,7A) and 1A (peak 1,2A), I thought to use the 1A, and feel the temperature from the motors.
On the housing from the TB6600-driver is noted that I can microstep down to 1/32, I’m not sure to use this (and not sure if the integrated chip from Toshiba can do this, because its not noted in the datasheet documentation from Toshiba).
Hopefully you understand my answers and questions.
Thomas
We use 1/16th microstepping on all axis.
You need to check your motor specifications to make sure you don’t exceed their rating. But I suspect the 1A is going to be better. Is there a good reason to use these over the drivers in a rambo board?
My plan is to use the ESTLCAM ARDUINO Steuerkarte from the ARDUINO-Club. This card has a lot of options, but no hardware-driver for the motors, that is the reason I bought 3 TB6600.
Thank you for your answer, I’m looking now for “nice” hard blocks, or I must design them for myself.
Best regards
Thomas
Jeremy, the pic you posted looks as if the Rollers are not positioned correctly. I did the same thing and had xy issues as well. This is from assembly instructions: * This is showing the pairs for the X and Y axis. It should be one Roller and one RollerM. It does not matter which way the gantry rail angle is facing as long as it is facing the same as its counterpart.
So angle needs to be on same side… not opposite.
