584,833 active members*
5,499 visitors online*
Register for free
Login
Page 1 of 2 12
Results 1 to 20 of 33
  1. #1
    Join Date
    Mar 2020
    Posts
    218

    Homebrew 4th axis build underway

    After spending way too much time debating whether I should try this or not I have pulled the trigger on building my own 4th axis for my 1100M. So far I have acquired a stepper driver (SteppersOnline equivalent to the Tormach piece), a NEMA 23 frame stepper, BS-0 dividing head, and the appropriate CPC connector to plug it into the existing outlet on the machine cabinet.

    Pictures and specifics of the build to come as I make progress but so far I have managed to get the stepper wired and responding to A-axis inputs from pathpilot without too much trouble. Unfortunately my stepper didn't last long before it started making a pretty awful noise (steady, high pitched beep sound) whenever the machine is taken out of reset. This didn't happen immediately but rather after 10-15 minutes of messing around trying to find a microstep setting that would work for the built in scale factor and the 40:1 head that I have. Not knowing what I could have done wrong but thinking perhaps I fried the stepper or the driver, I grabbed another motor I had laying around and plugged it in. This one was a NEMA 17 and rated for less current than the original. Within about 5 minutes it started smoking and making the same noise. Given the behavior, the reduced time to failure for the smaller motor, and the resulting notchiness of both motors (more so than before they were plugged in), I believe the motors were not capable of handling the current output from the drive. The NEMA 23 was rated for 2.4 amps which was within the range of the driver at the lowest setting but less than that setting's maximum output. The 17 was only rated for 1.something so it makes sense that it would burn up faster. I now have a 4amp stepper on it's way so I can run it at the drive's lowest setting and have some overhead available. Assuming I'm right about the cause of failure, that should solve the problem. It does look like I will have to alter the scale factor in the machine settings file which is a bit annoying but doable for now.

    In the meantime I am working out a mounting solution and drive coupler for the stepper to the dividing head. Nothing terribly complex but one that will allow for easy access to the worm gear adjustment lock screw so I don't have to tear everything apart to tighten up any backlash.

  2. #2
    Join Date
    Sep 2009
    Posts
    624

    Re: Homebrew 4th axis build underway

    Before you get too far into that, look up the Roger Caffin thread here at CNCzone.com/forums/linera-and-rotary-motion/261174-cnc-engineering.html

    And other threads, same forum.

    The first few posts cover some of the limitations of the approach you're taking.

  3. #3
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    While I appreciate the input, I am well aware of the limitations. Fortunately, my needs are not such that any extraordinary lengths will be required to achieve my goals. I have one specific project that has been in the back of my mind for the better part of 10 years that will require quite a bit of accuracy to complete (metric splined axle shaft) but even that is a 3+1 machining operation so as long as my build is accurate, precision isn't hugely important and backlash can be effectively ignored as I will only be turning the indexer in one direction. Long term I have my eye on the Micro Arc 4 but again, my 4th axis needs are minimal so I'd rather put the money into upgrading my machine to MX specs first.

    All that said, a minor update for now with more positive news coming next week hopefully. I received the new, higher current stepper the other day and plugged it in. As with the other two, initially it was working fine so I started playing with the scale factor to get the rotation set properly. After some time though, the new motor failed in the same way as the previous two. Excessive heat built up and the motor started making the same sort of high pitched tone (though lower in pitch than the other two). It still turned when commanded but clearly something is going wrong with the setup. I am giving it one more try though. A new stepper and driver are on their way to me now and should arrive tuesday. They were ordered as a kit which hopefully means they will work together properly and any issues of current or voltage in excess of what the stepper can handle will be avoided. Stay tuned.

  4. #4
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    New stepper and driver arrived last week as expected. My own foolishness almost put an end to the project for good though as after I wired up the new motor and fired up the machine, I could only get the stepper to spin in one direction regardless of what was commanded in the MIDI. Tried switching the direction of one coil and it started spinning the other way, but still only one direction. Tried a different stepper and got the same behavior so I finally quit and decided the driver must be bad. Started pulling everything out and that's when I noticed that the driver had 9 dip switches instead of the usual 8. Turns out the 9th switch was a mode selector to indicate to the driver whether the input it received was a step/dir signal or a cw/ccw signal. Of course I had it set wrong so I flipped that, put it all back together, and lo and behold, it spins both ways now!

    Since then I have been slowly dialing in the scale factor to try and get it as close to dead nuts as possible. For anyone else working on something similar, all the math that I found for setting the scale factor seems to be completely wrong. Everything I've seen posted indicates that (Steps per rev * gear ratio) / 360 should give you the correct scale. It works for the standard issue 4th axis from tormach (2000 steps per rev * 90:1 ratio / 360 = 500). Therefore, with the same motor step angle, the same microstep setting, and a 40:1 gear ratio I should have the following: 2000*40 / 360 = 222.222222. Started there and it wasn't even close to what it needed to do. Based on the actual behavior I then attempted to calculate a new scale factor based on how far the motor turned with the first one I calculated. That got me next to nowhere. It was then that I discovered that the scale factor is non-linear, eg. doubling the scale does not double the output. It seems to be roughly that doubling the scale results in half again more output (2x in = 1.5x out). So I started the long process of guess and check to try and get it right. I'm pretty close at this point. Over 40 revolutions of the stepper, I'm short of perfect by about 5 degrees. With the 40:1 ratio that means my actual output will be about .125 degrees over 1 full revolution (360 commanded, 359.875 actual). Not quite good enough but damn close. At this point every time I fire up the machine I check the accuracy of the a-axis and make a small change to the scale to be verified on the next boot up. I suspect this will be an ongoing adjustment until I either get the error too small to measure, or get tired of making adjustments and decide it's more accurate than the mechanics of the system to matter.

    As for the longevity of the stepper itself, though it definitely still gets hot, it's nowhere near as bad as the previous motors were. I had the machine running for about 2-3 hours yesterday running some other parts and had no issues from the a-axis motor being powered the whole time. So I still don't know what the real issue was, stepper or driver, but its behind me now. Next step is to finish the design for the mounting system to attach the motor to the dividing head. Real pictures will follow once that's done.

  5. #5
    Join Date
    Nov 2012
    Posts
    591

    Re: Homebrew 4th axis build underway

    It was then that I discovered that the scale factor is non-linear, eg. doubling the scale does not double the output
    That sounds wrong! It should absolutely be linear. There's no way that it would work correctly otherwise.
    So, what's going on is likely that you're losing steps, or your controller is doing poorly with micro-stepping, or speed of steps, or something like that.
    I suggest you debug what's going on first, and get to a system that has a linear scaling relationship.
    It sounds to me like your current system will change behavior under load, or air humidity, or phase of the moon.
    If it's not behaving linear, then no amount of tweaking will make it correct under a wide variety of conditions.

  6. #6
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    I had the thought about losing steps early on and have done some experiments to determine if that was happening. My simple experiment was to rapid the stepper in one direction then return to the starting point at a slow feed rate. I repeated the scenario backwards as well, slow feed to a point, then rapid back. In every test, the motor came back to the same point that it started from. Granted that is not a definitive answer as the speed could have nothing to do with it and I could be missing steps regardless of feed rate and accelerations. I have my controller set to 2000 microsteps per revolution which seems to be the standard setting that Tormach calls for on both of their 4th axis setups. I'm using a driver that is a variant of the same driver that Tormach uses so I assume that it's capable of doing what it needs to.

    If you or someone has any debugging suggestions other than what I've done, I'm happy to hear them so I can try them out.

  7. #7
    Join Date
    Nov 2012
    Posts
    591

    Re: Homebrew 4th axis build underway

    Do you have a logic analyzer? Or at least an oscilloscope? Hook it up and count the steps, to verify assumptions. (My background shows :-)

    My next suspicion would be the steps-per-rev setting, perhaps. If the scale is not linear, something clearly is not as expected. (Also, double-checking the math that says it's not linear ...)

  8. #8
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    I thought I had an oscilloscope but it was a cheap thing I got off amazon for basic circuit testing and proved to be a worthless pile of junk. I do have to get another at some point so perhaps now is the time to do so. I have verified the steps per rev is set to 2000 so that's not an issue unless the driver is not outputting the steps according to it's own datasheet (insert need for oscilloscope). Even if that were the case though, the non-linearity is based on observation of behavior. I don't have my data points handy so I'll have to dig them up this evening or recreate the experiment to verify what I believe I saw before. Of course, now you have me thinking about it, my method for measuring the rotational output wasn't terribly precise at first and got better the more I played with it so perhaps my belief in the non-linearity of the scale value is nothing more than poor testing methods. Regardless of that, the basic issue of the published math to determine scale not working does remain. The math tells me very simply that I should have a scale of 222.222222 (unless I screwed that up?) for a 1.8° per step motor, with 10 microsteps, spinning 40 times for 1 revolution of the dividing head (200*10*40 / 360). I am currently at about 2840 to get the appropriate response from the stepper. This is the same behavior I got from the previous driver and steppers before they all burned up so it doesn't appear to be limited to my current setup.

  9. #9
    Join Date
    Nov 2012
    Posts
    591

    Re: Homebrew 4th axis build underway

    Yeah, I can't think of anything more than that.
    Other than 10 microsteps being ... weird. Most micro-steppable controllers I've used have used powers of two -- 4x, 8x, 16x, 32x, ...
    And the division between 2840 and 222 is about 12.8, which with your microstep assumption of 10 makes 128, which is a power of two.
    No idea whether that's relevant or a total red herring, though.

  10. #10
    Join Date
    Jun 2020
    Posts
    8

    Re: Homebrew 4th axis build underway

    BLMA is a known tube/pipe bending machine manufacturers, and all Sheet Metal Bending Machines have conducted the strictest quality inspection before delivery. Moreover, as one of the world's leading companies in the manufacture and development of tube and pipe bender, we offer you custom-made solutions for your individual tube bending machine.
    https://www.blmamachinery.com/pipe-a...-machines.html

  11. #11
    Join Date
    Jun 2020
    Posts
    8

  12. #12
    Join Date
    Jun 2020
    Posts
    8

    Re: Homebrew 4th axis build underway


  13. #13
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    So, indeed, my belief in the non-linearity of the scale factor was totally wrong. Last night I checked it again and what I was seeing as non-linearity, was in fact, just an inverse relationship between scale and output and that's what got me all confused. Accepting that my method of measuring is only accurate to about 1 degree of physical rotation and allowing for that error, there is a definite linear relationship between each step. So that settles that. The next confusion is why is the math that is posted everywhere to determine the scale factor not working for me. In fact, not only does it not work for me, based on what I've observed and measured, it shouldn't work for tormach either because at the preset scale of 500, the stepper will turn once for every 51 degrees that is commanded. Their rotary table is a 90:1 reduction which means 1 rotation of the stepper should be 360/90 = 4° of commanded angle.

    Well, I'm pretty sure I've figured it out. The answer to all of my problems is actually quite simple. I'm a moron. Yup, I said it. I'm not fit to run a toaster, let alone hack a cnc machine. Jwatte, your comment about microsteps being powers of 2 sent me on a quest to see what all was out there. What I've found is that many steppers are in fact all powers of 2 for microsteps, but there are many more that ALSO have settings for various multiples of 5 (5, 10, 20, 25, etc). So ok fine, whatever, not really important. What IS important, is that while I was looking at pictures of stepper drivers online (sorry mom, I'm a degenerate) I happened to notice that one of them has the ON position for the dip switches as down, or away from the labeled side of the case. I've been doing everything as though ON is TOWARDS the label. Which means that instead of 2000 steps per rev that I thought I was setting up for, I was actually setting up for 25600 steps per rev. Plugging 25600 into the published math results in a scale factor of 2844.444. Well boys and girls, after a whole lot of experimenting I've come to a scale factor that results in about 1° of error in 230,400° of stepper rotation. Guess what my scale factor is currently set to? 2844.43

    All of a sudden, my issues make sense because I was doing the same thing for the amp settings AND the idle current halving which means all of my motors have been getting driven at substantially higher levels than I thought which explains why they were burning up. The current motor is rated for something like 6 amps so even at my incorrect setting, I'm still in a safe zone for the motor to operate in.

    Of course, this particular epiphany happened while I am nowhere near my machine to test my new theory but everything I've been dealing with now makes sense. I'm truly kicking myself for the lost time and wasted money for such a stupid mistake but such is life I suppose.

  14. #14
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    Yup, I have confirmed my own stupidity. All the dip switches were reversed. On was Off and vice versa. Flipping them all to the correct orientation and setting my scale factor based on the published equations got everything working exactly as it should have been. Clearly I need to do a better job of reading the destructions before I do a thing. Oh well. ONWARD!

  15. #15
    Join Date
    Nov 2012
    Posts
    591

    Re: Homebrew 4th axis build underway

    Feels good when the impediment clears, doesn't it? :-D
    Good for you. I wish you the best, shiniest, parts!

  16. #16
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    Indeed, and thank you. With that nonsense behind me, I'm back at working on the rest of the build. Towards that end their are a couple smaller purchased components on their way to me now. Upon arrival I will verify their dimensions and finalize the design for the motor mount, then fabrication can begin in earnest. In the meantime, here's the mount I have come up with. It's a direct adapter between the stepper and the dividing head which requires no modification to either part and allows for access to the backlash adjustment on the dividing head with the stepper still attached.
    Attached Thumbnails Attached Thumbnails BS-0 Stepper Mount v7.png  

  17. #17
    Join Date
    Nov 2012
    Posts
    591

    Re: Homebrew 4th axis build underway

    Fancy!
    I'd probably make that out of two flat parts and four struts, though ... for practicality.
    But when you're making one, anything goes :-)

  18. #18
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    That was the original plan since it would be simpler to machine and make more efficient use of material. But I decided it's boring so I'm doing it the fancy way

  19. #19
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    Mechanical assembly is complete. Forgive the surface finish on the adapter bracket, I did most of it with a shear hog and didn't bother going back over it with a finishing tool since it just needs to be functional. I'm considering making another one as one of my first projects with the new fourth axis but we'll see if I end up doing that. You'll notice in one picture the clearance cut I added to allow for adjustment of the worm screw without having to remove the stepper or bracket. At some point I will need to add a splash shield or enclosure for the motor as it is not sealed so flood coolant probably isn't going to be a thing until I'm able to protect it from the spray.

  20. #20
    Join Date
    Mar 2020
    Posts
    218

    Re: Homebrew 4th axis build underway

    Long overdue to post an update but it's been more than a little crazy lately. Short version: SUCCESS!!

    Once my stupidity was discovered and corrected, suddenly things started coming together quite nicely. Sorry for the facebook video links, hopefully everyone can see them.

    https://www.facebook.com/ian.vivero....6842521763635/
    https://www.facebook.com/ian.vivero....9446244836596/


    I still need to come up with a splash shield for the stepper so I can get the coolant going since I don't have anything else to keep tool temps down or clear chips out but overall I'm pleased with my initial results. Really, this ended up being fairly simple once I figured out how to properly set DIP switches

Page 1 of 2 12

Similar Threads

  1. my 5th 3rd is underway (2500x1600x400 v-groove)! what shall I call it?
    By posix in forum CNC Wood Router Project Log
    Replies: 7
    Last Post: 09-28-2009, 11:55 AM
  2. HOMEBREW Z AXIS WIERD BACKLASH
    By dgalaxy in forum Bridgeport / Hardinge Mills
    Replies: 4
    Last Post: 05-03-2008, 04:05 PM
  3. My Homebrew
    By Danka in forum Community Club House
    Replies: 0
    Last Post: 12-31-2007, 06:59 PM
  4. Homebrew CNC Lathe
    By Deviant in forum Uncategorised MetalWorking Machines
    Replies: 24
    Last Post: 03-06-2005, 04:38 PM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •