Ben's 3' x 6' Rebuild with pics

New Home Forum Mostly Printed CNC – MPCNC Your Builds – MPCNC Ben's 3' x 6' Rebuild with pics

This topic contains 129 replies, has 14 voices, and was last updated by  Smith Hayward 8 months ago.

Viewing 30 posts - 61 through 90 (of 130 total)
  • Author
    Posts
  • #76051

    Ryan
    Keymaster

    The dual firmware wants them wired normally closed, the regular firmware expects normal open.

    #76786

    Benjammann
    Participant

    what OS are you using? What is the specific error you are getting? what version of the Arduino IDE do you have installed? Have you tried a different port?

    Thanks for these questions. They led me to determine I was on the wrong version of the Arduino IDE. Also, I think I was trying to upload while connected in Repitier.

    So since then I’ve been tweaking my wiring and trying to wrap my head around the auto squaring calibration. I was getting very poor repeating position on cuts, then after a few hours realized my pulleys were loose. I’ve been breaking it in, getting everything square, retightening belts, going over everything with a fine tooth comb.

    About to go out and finalize auto squaring settings. Getting very close. Hoping to install some threaded inserts in tabletop soon, but unfortunately, despite being very square the X-axis is still .125″ too long over 60″. That’s going to mess up my spacing for the inserts considerably in my opinion. Does anyone have any ideas on how to get it a little closer, or a workaround for machining my insert holes where I don’t have to worry about it?

    Over the holiday weekend, the MPCNC got us inspired to clean and organize. Still a long way to go..

    IMG_2207

     

    Attachments:
    #76829

    Barry
    Participant

    Just create a gcode for a single row holes.  Make the home the center of the first hole.  Measure out the locations of each first hole and mark them on the spoil board, then just put the end mill on each row’s point.  When it finishes a row, move to the next spot, and rezero.

    #76857

    Benjammann
    Participant

    Just create a gcode for a single row holes.

    I might be able to do that with a row in the Y direction since at the near end it’s accurate, but at the far end, even the Y gets 1/16” too large. If I do it all manually I’m likely to end up even more off.

    Any way at all to shorten up an axis? Are my belts too loose or tight? I hear setting steps per mm doesn’t work on XY?

    #76858

    Benjammann
    Participant

    Just create a gcode for a single row holes.

    I might be able to do that with a row in the Y direction since at the near end it’s accurate, but at the far end, even the Y gets 1/16” too large. If I do it all manually I’m likely to end up even more off.

    Any way at all to shorten up an axis? Are my belts too loose or tight? I hear setting steps per mm doesn’t work on XY?

    #76877

    Barry
    Participant

    There’s no way you can stretch the belts enough to make it do that unless they’re not fiber reinforced.  The steps/mm are set by the pulleys, and not really set in software.  You’re an eighth inch off in one direction, and a sixteenth off in the other.  What happens if you draw a two foot by 5 foot rectangle?  Is it square?

    #76916

    Benjammann
    Participant

    The steps/mm are set by the pulleys, and not really set in software.

    What happens if you draw a two foot by 5 foot rectangle? Is it square?

    When I left off calibrating I was within 1/16″ square. Going to get it perfectly square tonight. From what I’ve seen during the calibration, I don’t think the 1/8″ over on 60″ is going to be corrected, unfortunately.

    Is there anything I can do or purchase to make an attempt at correcting this? I’ve seen someone on here use steel reinforced belts.. Are there any higher quality pulleys that would help? Anything I could change even if it meant spending more, or going a bit outside of the original design?

    #76920

    Raphael
    Participant

    I was the guy using steel reinforced belts 😉

    And they don’t work for the MPCNC. The steel reinforced belts are too stiff, so you have to tighten them more to wrap around the pulley correctly, which applies too much force on the motor shaft and makes them lose steps quicker.

    I have exchanged them for regular glas fiber reinforced rubber GT2 belt a couple of weeks ago and it runs a lot better now.

    If your X and Y axis are both off a bit but the machine is square then feel free to change the steps/mm setting. As long as it’s a linear error this should fix it (for example, a 100mm move is off by 1mm, a 200mm move is off by 2mm).

    Of course it’s also possible that you have a mechanical problem that needs to be fixed, check that the frame and the gantry are as square as possible, the belts and pulleys are thight, the rollers can move freely and that your motors don’t lose steps/your motor drivers don’t overheat.

    #76925

    Raphael
    Participant

    The glas fiber reinforced GT2 belt is what you get when you order regular GT2 belt, it’s nothing special, but it was an improvement over the steel reinforced belt.

    And you’re right, I can’t really think of anything that would cause a dimension to be too long except steps/mm…

    To calibrate the steps/mm move each axis as far as you can and measure the move. Then calculate how much you need to correct the setting, change it, reflash the firmware and test again, repeat if necessary.

    So, for example, your current setting is 100 steps/mm, you tell the machine to move 600mm, it moves 603mm:

    (100*600)/603 = 99.5 steps/mm

    That should get you pretty close.

    You can change the setting with M92 if you don’t want to reflash the firmware, just make sure you use it right:

    M92
    So for X it would be “M92 X99.5”, then after you are finished with the calibration don’t forget to save the values to the EEPROM with “M500”. But I’m not sure if the EEPROM is even activated on Ryans default firmware…

    Attachments:
    #76928

    Benjammann
    Participant

    So, for example, your current setting is 100 steps/mm, you tell the machine to move 600mm, it moves 603mm:

    (100*600)/603 = 99.5 steps/mm

    That should get you pretty close.

    Excellent. Can’t wait to get home and finish squaring, then set steps, then finally get these inserts installed. I changed my engineering day job to first shift so I get out now at 2:30 and go directly to the garage. Lol.

    #76930

    Raphael
    Participant

    I wish you good luck 🙂

    #77133

    Benjammann
    Participant

    Just wondering where to set the auto squaring X or Y offset in the marlin firmware. Is it under skew?

    Also, I’m seeing that my Y axis draws the rectangle right to size near the home position, but at the other end of the table it’s 1.5mm too big. Any ideas how to correct this? Belts are exact same tension. Auto square is complete.

    #77188

    Ryan
    Keymaster

    Also, I’m seeing that my Y axis draws the rectangle right to size near the home position, but at the other end of the table it’s 1.5mm too big. Any ideas how to correct this? Belts are exact same tension. Auto square is complete.

    That should not be possible. How different are they?

    Just wondering where to set the auto squaring X or Y offset in the marlin firmware. Is it under skew?

    config-adv.h you will see dual endstop stuff with an offset value.

    1 user thanked author for this post.
    #77201

    Raphael
    Participant

    That sounds so weird and like Ryan said should not be possible…

    Can you post a video of the machine drawing those rectangles?

    If the error is not linear over the entire work area then it’s not the steps/mm that is wrong, but right now I don’t have any idea what could cause this.

    #77212

    Benjammann
    Participant

    Can you post a video of the machine drawing those rectangles?

    If the error is not linear over the entire work area then it’s not the steps/mm that is wrong.

    That should not be possible. How different are they?

    So at the moment, I’ve got the machine square via setting it each time in Repitier. Tonight I’ll set that number (X 0.75) in the firmware, then make a video of drawing a rectangle.

    The last time I drew a 24″ x 60″ rectangle the Y dimension near the home position was perfect, but the far side was around 1mm bigger. The X dimensions are equal, both being around 3mm too big on the 60″ sides.

    I’ll draw another rectangle tonight to confirm this. I’ve noticed multiple times that the far side is bigger, so I’m not sure it’s operator error.

    I almost said “screw it” and drilled my insert holes, but I think it’s worth the wait if I can get it closer.

    #77392

    Benjammann
    Participant
    So as promised I recorded a video last night of drawing a rectangle. The results were the same, with the far end of the rectangle being about 1mm larger in Y direction than the side close to the home area.
    As I was thinking of what you guys would say or think a possible solution to the issue occurred to me. Turns out that since I’m drawing just the corners of the rectangle, and the first corner is drawn the opposite direction from the rest, is it possible that the slight enlargement of the far end of the rectangle is due to some slack being taken up as it moves in the opposite direction to draw?
    Please let me know if you think this is the case, or see anything else that raises a red flag. Really interested to see what you think about my machine travel and such. Could I go faster, despite having the slight jerking motion as I travel? Is this normal? How do the motors sound? Don’t mind the unclipped cable tie ends or the frayed wires, that will all be tidied up soon.
    I’m having a really tough time posting the video. Any tips? My post keeps not submitting..
    #77395

    Benjammann
    Participant

     

     

     

     

    #77406

    Benjammann
    Participant
    So as promised I recorded a video last night of drawing a rectangle. (See previous post) The results were the same, with the far end of the rectangle being about 1mm larger in Y direction than the side close to the home area.
    As I was thinking of what you guys would say or think a possible solution to the issue occurred to me. Turns out that since I’m drawing just the corners of the rectangle, and the first corner is drawn the opposite direction from the rest, is it possible that the slight enlargement of the far end of the rectangle is due to some slack being taken up as it moves in the opposite direction to draw?
    Please let me know if you think this is the case, or see anything else that raises a red flag. Really interested to see what you think about my machine travel and such. Could I go faster, despite having the slight jerking motion as I travel? Is this normal? How do the motors sound? Don’t mind the unclipped cable tie ends or the frayed wires, that will all be tidied up soon.
    I’m having a really tough time posting the video. Any tips? My post keeps not submitting..
    #77412

    Barry
    Participant

    Just to make sure, that side isn’t longer is it?

    #77415

    Ryan
    Keymaster

    1mm over that distance is going to be hard to track down.

    The first most obvious is, are the Y and X rails the same length as each other…you can see if you are 1mm off well the square will be 1mm off.

    With those corners drawn how are the diagonals measuring?

    Yes the direction of travel can have that small/large of an effect with a pen.

    #77416

    Ryan
    Keymaster

    Also from that video your home corner pushes the pen down 5mm or so, and the far corner barely seems to touch, that suggests the corners are not at the same height.

    #77419

    Benjammann
    Participant

    Just to make sure, that side isn’t longer is it?

    Hmm, I’m sensing some humor here? You mean the X vs Y axis, or the far side of the rectangle? Far side is 1mm taller in Y direction

    Are the Y and X rails the same length as each other…you can see if you are 1mm off well the square will be 1mm off.

    Are you talking about the Y rails being the exact same length as each other? Actually, I never checked this due to them being a stock 3′ from McMaster.

    With those corners drawn, how are the diagonals measuring?

    Diagonals are measuring perfect due to me offsetting X 1.5 in the auto squaring firmware.

    Yes, the direction of travel can have that small/large of an effect with a pen.

    Ill correct this and test again. I think I did last night but I was getting tired.

    Also from that video your home corner pushes the pen down 5mm or so, and the far corner barely seems to touch, that suggests the corners are not at the same height.

    You saw my original Z tubes were cut on a lathe to within .001″ of each other. Then I bottomed everything out, supposedly setting the height. I’ll look into this as well. Thanks!

     

     

    #77425

    Ryan
    Keymaster

    It has to be one of the three things I mentioned, I am strongly guessing it it the trust of the 3′ cut being -1mm tolerance.

    #77426

    Benjammann
    Participant

    Thanks! I’ll be checking all these things tonight. Already extremely happy with how accurate and precise things are.

    #77428

    Jason
    Participant

    Was your legs off??

    #77429

    Benjammann
    Participant

    Was your legs off??

    I cut the legs to within .001″ of each other. I’m going to check if there is a height issue when I get the chance.

    #77522

    Benjammann
    Participant

    I think we’re finally there. Last night I got the machine calibrated enough that I could not detect any errors in the large rectangle pen test with a tape measure.

    I first measured the 3′ Y tubes and their lengths were spot on, so I loosened up the parts a bit and bumped the side with the longer dimension in 1mm. This immediately fixed the issue. Then I used a height gauge to test and set leg lengths, I guess there might have been some remaining support or issue with the parts because the far end was significantly higher despite having the leg tubes the same size. Then I used the height gauge on each end of the 6′ X axis to set the end stops and make it parallel to the X rails. Then used a square to set the Y axis end stops. After that, I didn’t even need to adjust the auto squaring settings. I also adjusted steps per mm with m92 using the formula (100*length programmed)/acutal length = new steps per mm) Everything is sufficiently perfect at the moment. Ready to bore the table for inserts tonight. Thanks for your help everyone.

    Almost seems like a “pre-flight” checklist would be handy with items similar to what I listed above all put in order. I know it’s mostly written in the instructions, but I was just listening to a podcast about how checklists, even for seemingly simple straightforward procedures drastically reduce errors. It’s easy to miss something buried in a paragraph of text. (or a lot of things in my case. Lol) This would cut down on forum support as well. It seems like what took me weeks, could have taken days. But that could just be because it takes me a while to catch on.

    #77542

    Ryan
    Keymaster

    Everything is sufficiently perfect

    I love this statement, I will be borrowing it frequently.

     

    “pre-flight” checklist

    Yes! The Last section of the instructions has not been updated in a very long time. Seeing your original issues I went through them again and obviously need to do some revisions to make this more clear. I was teaching a friend how to use the CNC and he kept leaning on the rails and setting material on them and realized I need to make it clear that is not a good idea and why as well.

    #77598

    Benjammann
    Participant

    Everything is sufficiently perfect

    I love this statement, I will be borrowing it frequently.

    “pre-flight” checklist

    Yes! The Last section of the instructions has not been updated in a very long time. Seeing your original issues I went through them again and obviously need to do some revisions to make this more clear. I was teaching a friend how to use the CNC and he kept leaning on the rails and setting material on them and realized I need to make it clear that is not a good idea and why as well.

    Thanks! Glad you approve. Don’t think I’ve ever used that phrase before, but I also like it.

     

    So my joy was short lived. I got home and decided to do a final test before I bored insert holes. I went back to drawing the 3″ square. It came out over 2mm short on one side. So it looks like I’m stuck with it being only dead on at one end of the machine. It will either be perfect at 3″ and 1/8″ off on 60″  Or perfect at 60″ and way under at 3″

    I can’t see a way to accurately bore inserts over the span of the entire table now… :/

    EDIT: False alarm. My motor pulleys were loose. Getting faster at finding the root cause tho.

    #77635

    Benjammann
    Participant

    So I finally returned to the router after two weeks of pen. Tried boring some 10mm test holes and no matter how slowly I went, or how many finishing passes I took, they all came out .25mm too small.  Is it just a case of any slop or play in the machine adding up when it tries to do that small helical motion?

     

    Do I just need to manually increase the hole size in the file?

Viewing 30 posts - 61 through 90 (of 130 total)

You must be logged in to reply to this topic.