Welcome to Our Community

Some features disabled for guests. Register Today.

Ovals when you want circles - What's loose?

Discussion in 'General Talk' started by Richard McWhorter, Oct 10, 2021.

Tags:
  1. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Hey Guys,

    So It started as an issue where we were getting ovals when we wanted circles. I pinged tech support and followed the punch list of tightening that they suggested.

    One thing led to another and we ended up moving from a stock LEAD 1010 to a high Z upgrade.
    I learned a lot about the machine and thought all was well. Along the way I tightened all axis and checked the lead screws and everything along the way. Now I have it all together and one of my users is reporting that the Ovals rather than circles has returned.

    So help me understand what causes this. Is it slippage only on the Z axis? Would the part of the circle that is misshapen indicate where the error is coming from? I can send photos tomorrow but thought that this might be a well know problem that gives off lots of trial signs.

    Thanks for your thoughts.

    Richard
     
  2. Alex Chambers

    Alex Chambers Openbuilds team
    Moderator Builder

    Joined:
    Nov 1, 2018
    Messages:
    2,020
    Likes Received:
    983
    It is a common problem, but there are several possible causes. Pics, labelled with the X and Y directions, will help narrow it down a bit.
    Basically you need, starting at the bit, to wiggle things using reasonable force and looking for anything moving that shouldn't.
    Alex.
     
  3. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    Yes, photos would help confirm of its slippage (more pronounced stepping during direction changes for example) or calibration

    Also let them check the circle-diamond-square calibration job - measuring the real world dimensions and writing it down on it will show what is out: Circle Diamond Square
     
  4. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    So I was able to run a check on the machine and grab a few photos and measurement. I put the data in a table:
    upload_2021-10-10_21-55-31.png

    The photos are attached below. I also checked the calibration on the Z Axis and got this:
    I moved one inch 5 times and got the following:
    1.034
    2.066
    3.104
    4.141
    5.002

    I checked the X and Y Calibration as well. I used a large ruler rather than the Digital Caliper. I started each with the bit aligned at the 2 inch mark and jogged the cutter 5 inches, checked the mark and jogged 5 more inches. Both x and y came within ~1 mm of the 5" and then the 12".

    I also felt around on the machine and did not find any major movement. I had a minimal amount of movement when I pushed hard on the side of the bit. But I don't know what is "Normal".

    Hope this is helpful in figuring out why we are getting ovals. (The ovals are much more apparent in real life than the photos)

    Thanks again for your insights!

    Richard
     

    Attached Files:

  5. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    The pronounced step at around 5'o-clock on the circle indicates something loose. Stop collars not tight up against the bearings or a slipping shaft coupler most likely.

    It jumps during the direction change
     
  6. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Would it be only on the Z axis?
     
  7. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    I have been focusing mostly on Z and the machine sits in a corner so the y stop collars are a bit harder to see. I will check all stop collars and shaft couplers tommorow. Kinda late here! Not sure where you are but thanks for the quick reply!
     
  8. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    No the Z axis would have up and down bumps. This is in X or Y
     
  9. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Very helpful. I will be back at the machine later today and see what I can find. Thanks for the trail signs!
     
  10. Giarc

    Giarc OpenBuilds Team
    Staff Member Moderator Resident Builder Builder

    Joined:
    Jan 24, 2015
    Messages:
    2,341
    Likes Received:
    1,265
    Once you figure out what is loose, you should calibrate the machine so it is dead on. 1 mm is quite a bit to be off at 5 inches, or 12 inches. When I command mine to go 1000mm, it goes 1000mm. Calibrate with as long of a ruler as you have.
     
  11. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Thanks for the feedback!
    • I used the inches side and estimated the amount off. I guess it would be more precise to just use the mm side of the ruler. I used a v bit thinking it would add to visual precision. What do you use?

    • I used a digital calipers on the Z and liked the precision. of course that limits me to about 6 inches of travel. Ok for the Z, but you seem to recommend longer measures on the X and Y. So I got a LEAD 1010 with 29" and 32" working area, would you suggest just pulling back from the physical limit an inch or so and testing?

    • Any difference in testing by using the jog multiple times ( 10 x 1 inch) vs using the command line and going 10 inches ?
    Learning as I go, appreciate the shared insight!!!
     
  12. Giarc

    Giarc OpenBuilds Team
    Staff Member Moderator Resident Builder Builder

    Joined:
    Jan 24, 2015
    Messages:
    2,341
    Likes Received:
    1,265
    I also used a Vbit. I personally like to jog once as far as I can within the limits of the machine work area. A little bit of error may not visually show in a 10 inch jog. When I calibrated mine, I was off by 0.3mm in 1000mm (it went 1000.3mm). I doubt I would have seen that in a 200 to 250mm jog. My eyes aren't what they used to be. I calibrated my lathe with just short 100 mm jogs because it was easy. But, I noticed error when I went out the full length of travel so I picked a longer measuring stick and recalibrated.

    I also used a caliper on the Z because -- like you said -- there is not much travel.
     
  13. Stwspoon

    Stwspoon New
    Builder

    Joined:
    Jan 16, 2018
    Messages:
    6
    Likes Received:
    7
    This is just my 2cents, but I had similar issues and after lots of searching and tightening, I found that my spoiler board was shifting under load in x and y. This was on a C beam machine n. Good luck
     
  14. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Thanks for the help !!

    • I tightened it all down ( X, y leadscrew stop collars, Shaft couplers ) checked the corner brackets, and was surprised that several screws that were tightened last week were loose? Do you guys advocated using the blue Loctite to hold things down?

    • Also the tolerances looked better than before but still out. I attached a photo hoping that would point to something. I found a video on calibration ( I think it was from Giarc long ago) and plan to tweak the calibration. If it was modestly off could that be a factor in this or is that only affecting the planned vs actual of axis travel?

    • Also noticed that the front wheels on the right side of the x gantry seem to have worked there way a bit loose. Could that be a factor?

    • Hope there's a trick to adjusting them from above?


    We have the spoilboard attached to the table and getting underneath is a major production. Any ideas on this issue and the ovals is appreciated.

    Thanks
     

    Attached Files:

  15. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    You are welcome yes. Vibration prone environment :)

    It still looks more like backlash than calibration. Leadnuts tight against the plates? No slippage there? See LEAD 1515 cut is not accurate, please help! (similar symptom)
     
    #15 Peter Van Der Walt, Oct 13, 2021
    Last edited: Oct 13, 2021
  16. Alex Chambers

    Alex Chambers Openbuilds team
    Moderator Builder

    Joined:
    Nov 1, 2018
    Messages:
    2,020
    Likes Received:
    983
    Re adjusting eccentrics on the Y gantries - take some thin (1.5 mm) mild steel, cut an 8 mm slot, bend to a right angle. It will also help to use a mirror - I use (used to use, I've got rid of the wheels on mine) an inspection mirror, but anything that helps you see what you are doing will do.
    I agree with @Peter Van Der Walt - a calibration error would produce a perfect ellipse, it does look more like a backlash issue.
    Alex.
     
  17. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Thanks for the ongoing help. I was able to get back to the machine and tighten the loose eccentric on the Y gantry.
    I led a orientation class for two new Maker Space members and everything cut fine. We made two really basic signs.

    I intended to go back and cut the circle-diamond-square calibration that Peter forwarded and see if the adjustments hade tightened the margin.

    Before I got to it another member asked me to help him run a simple thru cut on a pumpkin file. I did and all was well until about 5 minutes into a 12 minute job. Then as the bit was moving from one row of letters down to another the Z suddenly plunged all the way into the board. I hit the kill switch and when I tried backing the z out, the direction of travel was reversed ( Up went down and Down went up).

    It is soo frustrating to spend this much time on the machine rather than using the machine.

    We were using Easel to create the file and send it to the control software. Is it possible to go back to a log in the Control software and see where the job went bad and determine what caused it. We had this problem several months ago and never really got it nailed down.

    Are there a broad array of things that might cause this "Z" plunge......
     
  18. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    also, right after the plunge and the emergency power out I filmed the BlackBox and saw the network light flashing furiously.( see attached) Does this indicate something of note? Thanks
     

    Attached Files:

  19. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    Symptoms says Wiring: see docs:blackbox:faq-identify-motor-coils [OpenBuilds Documentation]
    Sudden onset of symptoms would likely be one of the 4 wires pulled out of a terminal, someone forgot to tighten a terminal or wire broke off inside insulation (tight bends, strain, back and forth bending, etc)
     
  20. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Thanks I will begin the check there.

    On that note, we had random EMI a while back and we spend forever chasing them down. We added ferrite cores and went back to our Dewalt ( from the Makita) and things seemed to settle down.

    Recently we got the newer limit switches and received the wrong wiring. Openbuilds sent us a complete wiring package for our LEAD 1010. I would like to rewire all steppers and limits switches and do the most possible to stabilize the system. Would you recommend moving up to a shielded cable to get the most stability or its it only more stable in theory.

    I have the new stuff from Openbuilds on hand, but the real effort is the labor and routing. If shielded saves me from a single instance of EMI of "Unknown" it would be worth while. Thanks for your thoughts.
     
  21. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    The Makita needs an Earth wire. See Wiring - how to handle router power

    Shielding helps, but less so than addressing the source of the EMI. Router most likely, seperate its power cable out into a Swing Arm so its not near any low voltage wiring

    Adding a second cable chain and seperating motor wiring from limit switch wiring is also a good upgrade, as motor wiring can also radiate EMI.

    If you don't have Xtension Limits, they are a must have upgrade

    More EMI tips here: docs:blackbox:faq-emi [OpenBuilds Documentation]

    I agree some labor is in order. It appears to have been put together a little too carefully - with some fastners and terminals already loose - indicating there may be others that weren't tightened correcty, and the backlash issues etc. It does appear to be in need of a little TLC to reach its full potential, both mechanically and electrically it seems. At the end of the day - it is only as good as you build it :)

    The Build videos doesn't go as in depth - but the more time you spend during the initial assembly squaring, making things perfect. Sanding down protrusions where they interfere with fits, carefully adjusting wheels so they are preloaded, but not squashed. Setting the anti backlash leadnuts properly. Spending some time on doing clean reliable wiring. Adding some additional cable chains / corrugated tubing. Adding extra Earthing. Building a solid base for the machine to be mounted to so it has a solid foundation. Spending a little effort on workflow in your own setup and then deciding what works best. Even if it means mounting things totally opposite of our video suggestions is fine too (I like mounting my electronics off-machine personally for example, and I put my XY homing switches at Axis Maximas for example (grbl default, our videos does it XYmin, Zmax). The base kits are a solid start, but there is always room for improvement - so if you are going to do a teardown and partial rebuild, look for places where you think you can do better and DO IT! :)
     
    #21 Peter Van Der Walt, Oct 15, 2021
    Last edited: Oct 15, 2021
    David the swarfer likes this.
  22. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Hi Peter,

    Thanks for taking time to detail out your response. It is quite helpful.

    • So I checked on the machine and specifically the Z connections and they looked and felt tight. Now admittedly the BlackBox is connected up under a wooden table and it makes it hard to see and feel at the same time. I plan to move it to an off machine location ( Like you mentioned) where I can easily test the full run with a multimeter.

    • On the issue of corner brackets and such loosening. The approach thus far has been to tighten the screws tightly and until there is a bit of flex in the ball driver. So short of some kind of torque wrench I am not sure how to improve ? I am leaning towards Loctite approach I mentioned earlier.

    • Thanks for the idea of separating Motor wiring from Xtension Limits wiring. We did upgrade to those and I can certainly run them separately. We have already implemented the Swing arm to separate the router power cable and the vac hose from the other wiring. I also picked up this EMI shielding tape off of Amazon. The idea was to wrap the new wiring kit to reduce emi probability. Now the question is, is that good enough or do I just buy some shielded cable and be done with it.




      upload_2021-10-15_15-2-30.png ( https://www.amazon.com/Conductive-C...RL4K/ref=cm_cr_arp_d_product_top?ie=UTF8&th=1)



    So after all of this, I went by the makerspace and the user that I was helping when we had the Z plunge last night was back at it. He had run the exact same job and it had cut as expected. No plunge at all? So I guess it is the intermittent stuff that is the most frustrating. It leaves you scared to look away !

    Again Thanks for the help, I promise I am working to become less needy .......!
     
  23. Alex Chambers

    Alex Chambers Openbuilds team
    Moderator Builder

    Joined:
    Nov 1, 2018
    Messages:
    2,020
    Likes Received:
    983


    Don't worry about it - we get bored here if no-one comes with problems to solve.:D:D:sleepy:

    Alex.
     
    Richard McWhorter likes this.
  24. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    Well, if its not wiring, other random causes include:
    - Max Rate / acceleration tuning in Grbl Settings just a little close to the edge of stalling for your specific machine - and sometimes during a rapid move Z stalls out and looses position. If you catch it in action you may hear the Z motor whine a high pitched tone for a moment. Random because the time it ran fine, the dust hose was just not weighing it down as much for example
    - Operator error like forgetting to home or zero (but usually relates more to crashes at the start of the job - would have been good to check the Serial Log tab right after the crash to see around which commands were sent - was there any unexpected coordinate changes, etc at the time
    - Mechanical: bit gummed up and got bogged down with uncleared chips in the bottom of a hole when it was supposed to retract, skipped steps, next Z-down move was from incorrect location because of an earlier skip
    - Bit pulling out of collet - for example with a 6mm bit shank in a 1/4" collet or insufficient tightening, it can pull out during the job - bit now extends deeper than toolpath planned for it to, thus crashing
    - EMI based serial corruption (G1Z-1X10 became G1Z-110 when EMI knocked the X out of the string)


    EMI tape / shielding - which EMI issues are you dealing with specifically - false hard limits, serial corruption, etc - sticking shielding on things isn't cover-all magic - shielding specific components for a specific reason, may help. To do that, we need to know what EMI issues you see.

    And then still, Shielding is low low down on my list of addressing EMI issues - nothing wrong with adding it - but usually not needed. Thus if you have something emitting so much EMI that none of the other tricks helped and you NEED shielding - then I would begin to worry about what the source is. If it was a HF plasma or a cheap chinese VFD, fair enough, we expect a lot of EMI. But not from a Dewalt or a (properly earthed) Makita
     
  25. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Ok Thanks, I discussed it again with the user and he felt that perhaps the center of a small letter cut out had come loose from the cut immediately prior and perhaps bound the bit? I don't know, When I heard it go bad I was headed for the kill switch and did not notice.

    Would that possibly be a cause ? kind of like your mechanical example?

    He is cutting several pieces and promises to report back results. Maybe it was a one off and I am a bit paranoid of EMI?
     
  26. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Resident Builder Builder Project Maker

    Joined:
    Mar 1, 2017
    Messages:
    8,715
    Likes Received:
    2,852
    Sounds very plausible. Remember to add Tabs to those inside toolpaths to prevent them popping out :)
     
  27. Richard McWhorter

    Builder

    Joined:
    Mar 31, 2019
    Messages:
    72
    Likes Received:
    12
    Thanks! I will double check that on future cuts!
     
    Peter Van Der Walt likes this.

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice