Welcome to Our Community

Some features disabled for guests. Register Today.

Y axis not syching at random.

Discussion in 'Motors' started by Delco, Nov 29, 2018.

  1. Delco

    Delco New
    Builder

    Joined:
    Oct 23, 2018
    Messages:
    22
    Likes Received:
    8
    Hi , just finished a leadscrew 1000x1000 workbee build using nema23 motors , Arduino uno , protoneer shield with drv8825 drivers.

    it seems to sometimes have one y Stepper lag behind when jogging , it doesn't happen everytime but mostly if I do a large 100mm jog.
    I have set driver current to 0.9v running a 24V power supply , this is where they seem happy and quiet , I had it at 1.4V but the motors would not drive properly . I can up the power supply to 36V , would that help ?

    Motor specs
    • Step Angle:1.8°
    • Rated current:2.8A DC
    • Rated Voltage: 3.36V
    • Winding dc resistance(25℃):1.2Ω±10%
    • Winding inductance:2.3mH±20%
    • Holding torque:≥1.26N.m
    • Shaft diamater: 6.35mm or 1/4″
     
  2. Giarc

    Giarc OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Jan 24, 2015
    Messages:
    2,898
    Likes Received:
    1,613
    I had problems with my DRV8825s due to overheating. They would lose steps on larger jobs. Are you cooling them real well with a fan, or fans?
     
  3. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    Problem temporarily resolved dropped ref voltage to 0.8v and added better cooling fan .
    Think i need to reconsider the drv8825s and find a better stepper driver solution .
     
  4. Giarc

    Giarc OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Jan 24, 2015
    Messages:
    2,898
    Likes Received:
    1,613
    I switched to the DQ542MA drivers and have not looked back. My fear was that I would be running a long job and at 90% complete they would overheat on me. For the DQ542MA drivers I used an arduino screw shield to make the driver hookups easier.
     
  5. Delco

    Delco New
    Builder

    Joined:
    Oct 23, 2018
    Messages:
    22
    Likes Received:
    8
    Thinking about going to the tb6600 drivers , cheap and can still hookup to my protoneer shield . did a 6 hr test cutting air last night, cooling made a big difference no steps lost but I am also worried about long term , its running at a lot lower current than is desirable.
     
  6. Giarc

    Giarc OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Jan 24, 2015
    Messages:
    2,898
    Likes Received:
    1,613
    I believe others have used those with success. Here is a post/thread that may help you. Stepper motor won't move, any suggestions? (Solved)
     
  7. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    Well i fitted up the tb6600's to the arduino using a 36v 11A power supply . Using nema23 1.26nm motors .
    Instantly better setting current limit to 2.8A
    Found on the 1/16th microstep the motor would stall if i jogged faster than 5000mm/min. Dropped microsteps back to 1/8 and then was able to achieve 6000mm/min with a scceleration of 100mm/sec2 without stalling .
    Happy with the outcome .
     
  8. Delco

    Delco New
    Builder

    Joined:
    Oct 23, 2018
    Messages:
    22
    Likes Received:
    8
    Well after a long break I started to do some testing , dropped max speed down to 2000mm/min and acc at 40mm/sec
    The motors aren't stalling but I have noticed the Y axis goes out off synch by a few mm after a period of air cutting just driving x axis back 400 and then back to zero then repeat a few times.
    I have the slave Y TB6600 dir and step jumpered from the Y axis so sharing same control signals but using separate drivers.
    I have tried changing microsteps from 1/2 to 1/16 with no change , I have changed max amps from 2 to 4 amps with no change apparent.
    Starting to pull my hair out on getting this thing to be accurate.
    Are these motors too small or low on torque for workbee , The slave Y seems to loose steps more often but I have noticed the main y axis wheels are a little tight with the eccentrics loosened all the way off.


    Any ideas ?
     
    #8 Delco, Mar 9, 2019
    Last edited: Mar 9, 2019
  9. Christian James

    Christian James Journeyman
    Builder

    Joined:
    Jun 8, 2018
    Messages:
    434
    Likes Received:
    210
    I originally intended to use 175 oz motors something like you have, but upped them to 270 oz in case I wanted more power for a future build. Having said that, 175oz motors should be powerful enough. I would say though that a 32v power supply would be more suitable than 24v
     
  10. Delco

    Delco New
    Builder

    Joined:
    Oct 23, 2018
    Messages:
    22
    Likes Received:
    8
    I already upped the power supply to 36v when I added the TB6600 drivers and removed the DRV8825's
    I would of thought air cutting would not need much torque as its only moving the axis.
     
  11. Giarc

    Giarc OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Jan 24, 2015
    Messages:
    2,898
    Likes Received:
    1,613
    How tight is tight for those wheels? Which wheels on the plate? All of them, or just a couple? Is it just on the slave Y axis side? The TB6600 drivers are 4 amps each so that should be more than enough for the Workbee. Is there any chance you have a loose signal jumper wire from one Y to the slave?

    Caution! The next two paragraphs are not for the feint of heart, but for the impatient. If this issue is in the same spot on both sides, you should contact the Openbuilds Parts store directly via their email. Or, if you think it is defective in any way contact them:

    Now for the dangerous stuff. Do you have a 13/64 inch drill bit? How bout a 19/64 inch bit. The reason I ask is that maybe - if it is only one or two wheels that are too tight on the top (for example the middle two) but the others are adjusted to the proper tightness - the holes in the plate may not have been cut to the 5.1mm size in the design or cut "too low" towards the beam. The 13/64 bit will make this 5.1mm hole into a 5.15mm hole. This may be the "looseness" you need.

    If it is the bottom wheels that seem too tight, maybe the 7.2 mm design for the eccentric holes was slightly smaller, or cut slightly too close to the holes on top. The 19/64 bit will make this 7.2mm hole into a 7.54mm hole. That is borderline pushing the "too much" increase. If I were to go this route, I would only try it on one first. Just in case...


    When I built my CNC router, I cut and drilled all the plates by hand. I had an issue where I cut a hole on both plates (they were fastened together when I drilled the holes so they would all line up) too low with my 5 mm bits. I used the 13/64 bit and luckily, it took care of the problem.
     
  12. Christian James

    Christian James Journeyman
    Builder

    Joined:
    Jun 8, 2018
    Messages:
    434
    Likes Received:
    210
    Maybe worth swapping the slave driver with the X to see if problem changes?
     
  13. Delco

    Delco New
    Builder

    Joined:
    Oct 23, 2018
    Messages:
    22
    Likes Received:
    8
    Giarc the tight wheels are not hugely tight just not freewheeling and that axis seems to loose steps at random as does the looser side. X also looses steps although I haven't spent a lot of time testing that one yet. I think its more a tolerance issue in the cbeam as if I use it on the other side its fine.

    Chistian I have a spare new tb6600 so will try that and see if problem goes away.
    If it was a loose connection I would expect that drive to always be out not all of them being inaccurate but I will redo all wiring and see if there is a change.

    Maybe best to sort out x axis first as there is only one drive involved
     
  14. Giarc

    Giarc OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Jan 24, 2015
    Messages:
    2,898
    Likes Received:
    1,613
    Not to be a driver snob, but they are $6 drivers vs the $38 DQ542MA drivers. I now use the DQ542MA drivers and have never had issues since I switched. My guess is the higher quality is why OpenBuilds sells them vs the cheaper ones. I had similar issues as to what you describe while using DRV8825 drivers which is why I upgraded. ALthough, some people here have had great luck with the TB6600s.

    I agree that trying the spare is the way to go first.
     
    Peter Van Der Walt likes this.
  15. Peter Van Der Walt

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

    Joined:
    Mar 1, 2017
    Messages:
    13,751
    Likes Received:
    4,070
    It very much is a You get what you pay for.

    Heads up as the chineses saw TB6600s sell well, they often describe their TB6560 (the worst stepper IC ever) boards as TB6600s. Sometimes you also get lucky and buy TB6600s that are actually the even better TB67S109s. Just because the description said its a TB6600, doesnt mean it is. Thats why some report success, some absolutely love them, and most report problems. Depends on the actual chip used...
     
  16. Christian James

    Christian James Journeyman
    Builder

    Joined:
    Jun 8, 2018
    Messages:
    434
    Likes Received:
    210
    I use the uncased 6600"s and they perform perfectly well - not to say there may be bad ones about though.
     
  17. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    I bought them from the australian openbuilds store so assummed they werent the junk ones . Will take apart tommorow and try and see what i have got . Never expected them to be as good as a gecko but wasnt expecting lost steps either.
     
  18. Peter Van Der Walt

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

    Joined:
    Mar 1, 2017
    Messages:
    13,751
    Likes Received:
    4,070
    Official Distributors can be verified by emailing http://support.openbuilds.com/support/home
     
  19. Delco

    Delco New
    Builder

    Joined:
    Oct 23, 2018
    Messages:
    22
    Likes Received:
    8
    Well after hours of testing I can honestly say my issues are the TB6600's
    My setup is a Arduino v1.1f with a CNC Shield CNC Shield v3.51 - Latest Version - Maker Store using the Tb6600's Stepper Driver - TB6600 4A 40V - Maker Store and using the stepper driver cable adapters Stepper Driver Adapter Module - Maker Store , a 36v 11A power supply and Stepper motors Nema 23 Stepper Motor 1.26N.m (6.35mm shaft) Version B - Maker Store
    .
    The problem I was having was with the std openbuilds setup for my workbee 1000x1000 the motors would stall and loose steps when speed was set at 2500mm/min as per standard settings.
    My testing consisted of concentrating on the Y axis as it was easier to measure position with the dial indicator.
    I zerod the machine and commanded y400 then y0 and saw where it came back to but it stalled part way through the acceleration . I could not get it to move reliably until I dropped settings below 1000mm/min and 30mm/sec then the machine would move semi reliably only loosing steps but not stalling , I tried every microstep combination as well as every amp setting , as the amps below recommended settings dropped it started to stall again.
    In frustration I found some old drv8825 drivers and plugged them into the CNC shield and ran wired to the stepper connections on the TB6600 but unplugged to facilitate testing.

    I reloaded the firmware settings back to 2500mm/min and 150 mm/sec as per standard and success , I was able to ramp it right up to 5000mm/min with no issues.
    Testing repeatability over numerious cycles of y400 y0 iterations and it came back to zero every time ( well within 0.02mm) and as yet I haven't dialed in the backlash by adjusting the nylon blocks.

    To make absolutely sure I reassembled it again with the TB660 and it was back to being busted. I have 5 TB660's so I rotated them around through the Y axis motors with the same poor result. the only change needed to GRBL was $4 needed to be disabled which may or may not be relevant.

    In conclusion TB6600's are a waste of time for me , I went to them from the drv8825's as I was having intermittent heat issue on my last CNC and expected the TB6600's to be superior which they are NOT.

    Hopefully this helps someone else in the future.

    Time to look into a gecko 540 I think .
     
    #19 Delco, Mar 10, 2019
    Last edited: Mar 10, 2019
    Rick 2.0 likes this.
  20. Christian James

    Christian James Journeyman
    Builder

    Joined:
    Jun 8, 2018
    Messages:
    434
    Likes Received:
    210
    But you said you had this problem in your first post when you were using 8825's.
     
  21. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    Yes i had heat related issued loosing a occasional step and stalling on very high soeed jogs in excess of 5000mm/min with the dr8825s . Fine on a cold night but at 40 deg days it was a issue .
    Hence why i am not going back to drv8824s
     
  22. Christian James

    Christian James Journeyman
    Builder

    Joined:
    Jun 8, 2018
    Messages:
    434
    Likes Received:
    210
    No not you Daniel - I was referring to the original poster Delco.
     
  23. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    I must gave a different log in on my phone Christian lol
     
  24. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    The strange thing is the tb6600s were fine on the first night when i tested with just a x carriage on a vrail . Went to crap when i assembled it all again . And yes i have checked it runs free with no lead screw installed and the lead turns freely when installed
     
  25. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    The fact that everything was fine got me thinking so i reached out to David from Sharp CNC who uses tb6600 in his builds .


    TURNS OUT I HAD THE VOLTAGE TOO HIGH AT 36.2V

    when i originally set it up i set voltage to exactly 36v and it all worked. Somehow voltage crept up .2v and it all went south from there.
    I dropped voltage back to 24v and everything is working fine again .
    Aldo found one board would not switch microsteps to 1/8 but i can live with that .
     
  26. Christian James

    Christian James Journeyman
    Builder

    Joined:
    Jun 8, 2018
    Messages:
    434
    Likes Received:
    210
    My 6600's receive 32v from the power supply which is the recommended setting for the drivers I believe. They work just fine.
     
    #26 Christian James, Mar 11, 2019
    Last edited: Mar 11, 2019
  27. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    Must be heat related , below 32v was fine at 32.2v it starts having issues .
     
  28. Christian James

    Christian James Journeyman
    Builder

    Joined:
    Jun 8, 2018
    Messages:
    434
    Likes Received:
    210
    Hey Daniel, don't know if you have already but they have to be cooled. I have 4 fans on mine - one for each driver.
     
  29. Daniel Collins

    Builder

    Joined:
    Oct 17, 2018
    Messages:
    13
    Likes Received:
    0
    I have them mounted back to back with fan blowing between them down the cooling fins . Havent noticed a performance drop yet dropping voltages down .
     

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