1. Hey guyz. Welcome to the All New Phlatforum!



    Sign Up and take a look around. There are so many awesome new features.

    The Phlatforum is a place we can all hang out and

    have fun sharing our RC adventures!

  2. Dismiss Notice

Step Resolution??

Discussion in 'Trouble Shooting - Support - Help Section' started by FozzyTheBear, Jul 19, 2009.

  1. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    Hi,
    This has probably been covered before, but I'm just doing some basic calculations just to see what numbers I need to have ready for when my kit arrives.

    I'm just working out the motion resolution and puzzled as to why anybody is running their steppers at anything other than full steps.... why would I want to run at 1/8th step or 1/4 step or even 1/2 step??

    This is how I work it out... assuming that I'm using the ACME 12 tpi lead screws with the stock motors....

    200 Steps per revolution from the motors at 12tpi is 200 X 12 =2400 steps per inch.
    Therefore 1 inch divided by 2400 = 0.00042" per step approximately. So that's less that half a thousandth of an inch per step. I don't really see a reason for cutting foam or wood at a higher resolution than 1/2 thou... OK, assuming that to be correct.

    The motors run faster at full steps than they will at 1/2 steps or 1/4 steps or 1/8 steps.... they also produce more torque at full steps, you also have less count to deal with because the count isn't multiplied, so they should also be less prone to missing steps.

    Forgive me if I'm being thick here, but have I missed something??.

    While I'm pretty new to CNC, I'm not new to engineering or electronics. so as far as I can see I got that right. Unless anybody can illuminate me as to what I got wrong there please??.

    Best Regards,
    Julian (Fozzy The Bear)
     
  2. Anonymous

    Anonymous New Member

    Offline
    Messages:
    1,280
    Trophy Points:
    0
    Yeah, I tried the math too, but on my MK1.5, I think I'm still running everything at 1/8 step increments. There were many options when I was running the 3/8-12 screw drive, but I was always finding binding spots that would cause missed steps unless I kept the resolution down.
    I remember posting some of my calcs at viewtopic.php?f=140&t=735&p=5859&hilit=ipm#p5859 and your numbers look correct.
    I think the best I ever reached was 60 ipm with the ACME rods (which is certainly "gettin with the program"). :cool:

    YMMV - Your Mileage May Vary :)
     
  3. gasmasher

    gasmasher New Member

    Offline
    Messages:
    402
    Trophy Points:
    0
    Location:
    Atlanta, GA
    Microstepping isn't all just about resolution. When running a stepper at certain RPMs resonance in the circuit can cause the motor to miss a step. Microstepping helps remedy the problem by turning the strict on-off of coils to a smoother curve that looks more like a sine wave. This reduces the chance for resonance and a missed step.
     
  4. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    Thanks Dennis and Gasmasher,

    OK that clarifies it a bit.... I'm certainly not after chasing down 60ipm Personally I think 40ipm is more than enough for me on fast feed rate. This is why I'm not modding for a belt feed. I might later mod up to ball screw feeds, but I suspect at that stage I'll be wanting to change out the steppers for something with a bit more torque. Fortunately replacement steppers seem to be widely available on e-bay, even on this side of the Atlantic (UK)....

    Isn't that what the filter circuit on the board is for? I can see the point of microstepping for that purpose now, but I'd have thought that it would have worked better to use full steps and just turn up the filter. Kind of less actual speed loss that way maybe?

    Best Regards,
    Julian (Fozzy The Bear)
     
  5. gasmasher

    gasmasher New Member

    Offline
    Messages:
    402
    Trophy Points:
    0
    Location:
    Atlanta, GA
    Sorry, you are getting out of my understanding when you mentioned the filter circuit. I don't have an EasyCNC board so I'm not sure how it is designed. Are you talking about the diodes used to protect the ICs from feedback voltage when a coil is de-energized in the motor?
     
  6. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    No I was talking about the PFD..... The easy-cnc boards, in addition to the V-ref control have a second pot that controls something called Percentage Fast Decay (PFD) As far as I know, this more or less knocks the corners off the waveform, until is more closely resembles a sine wave. It helps eliminate noise interference. With the penalty that using it slows the output down slightly.

    But if we're running at full pulse to start with, then the board is turning the motors at twice the speed it would be at half pulse. Hence, then using that PFD filter would I think be more of an advantage than running micro-steps to get the same result and should result in a higher actual output speed.

    Best Regards,
    Julian (Fozzy The Bear)
     
  7. tvcasualty

    tvcasualty New Member

    Offline
    Messages:
    637
    Trophy Points:
    0
    Fozzy, I'm always interested in further tuning!
    I found this information on the easy-cnc board controller chip:
    http://www.chipcatalog.com/Allegro/A3977SED.htm
    See pages 10-14 on the data sheet (pdf link on above page).

    Currently, in the software I set my "Driver Microstepping" to 1 and allow the boad to do my microstepping. I spent a lot of time looking into the specs of the actual steppers, but found the board had more to do with it, specifically the chip. When I put in the correct numbers into the software the noise level of the pp went drastically down and I was able to increase my ipm.

    I'm not sure if Mach has these setting capabilities, I am using ECM2.

    Perhaps I need to fool with the motor steps switch on the board as well?
    I fully admit to still learning mode of operation!
     
  8. kram242

    kram242 Administrator Staff Member

    Offline
    Messages:
    6,311
    Trophy Points:
    13
    Location:
    NJ
    Tv not sure if EMC has a CV (constant velocity) mode but it may be that you are running into this problem. This was a problem I ran into with the MKII as the speeds increased.
    Say for example, you're cutting a 90° corner. With CV on, as you get to the corner, one axis has to decelerate, while the other accelerates. This is you're rounding. The faster the acceleration, the smaller the rounding will be. You can also play with the angle settings so that it reverts to exact stop at a certain degree angle.
    CV mode is the blending of lines and arcs into a Polly line, the smoother the original line the more accurate the CV line, sudden changes in direction result in large blend rads, thus the original shape is distorted.
    Exact stop mode does exactly what it says, it will come to a stop before it does the next line of code. Constant Velocity mode blends the decel of one line with the acel of the next line so the movement is fluid and smooth. This can cause problems on some setups and can give corner rounding etc. Usually CV is the one to use unless it is giving you corner rounding problems.

    CV feedrate is, as I have found, the maximum allowed feedrate in "corner cutting mode", that is in the exact parts of the cutting where CV feedrate/CV distance is working. CV feedrate is working at the end and beginning of line and arcs segments.

    CV distance is how far from the start and end of a line or arcs segment where "corner cutting" is allowed. A high value or not turned on gives deviation from the original toolpath on a relatively long section at the intersection of two line or arcs segments. A low value gives a short distance where the actual toolpath deviates from the original toolpath, but at the same time lower feedrate in those same sections.

    You may want to set the CV distance to a low value like 0.1 to 0.3 when you need reasonably sharp corners. Such a low CV distance value gives nearly as sharp corners as exact stop mode but still considerably higher feedrate. When more rounded corners is allowed a higher value like 0.5 to 2.0 may be adequate. A CV distance of 180 gives very large deviations from the original toolpath when the feedrate is high and/or acceleration is low.

    There are several moves where constant velocity does not function - and G2 and G3 is one - because the axis co-ordinates for these moves are continually generated in Mach3 as the move progresses


    The slower the feed rate the better the match to the original path (more time to change direction).

    Under general config several CV config options can be set, these can be set the cause the axis to slow before a corner and reduce rounding.


    1. Absolute Stop mode means that the axis - as they change direction (or speed) to give an apparent change in direction - come to an absolute stop - before setting off on their new setting.

    2. CV mode means that the acceleration required for the new setting is computed and mixed in with the deceleration of the old setting, so the axis never comes to a stop.

    C.V. mode should be smooth - Absolute stop is often jerky is, for example a curve is made up of many little straight lines.

    Yesterday I was talking with xtream because he ran into the problem where the machine would basically cut every line segment at 200 in/min an create a very jerky style movement sounded like a machine gun on small arch's :)
    As far as CV settings go it seems to be a per machine basis depending on weight of the gantry etc..
    But once we get it dialed in for the Phlatprinter everyone should be able to use the same default settings.
    This is always been a known problem for all cnc machines. Really its a simple matter of physics.
    If you think of it as a car traveling down the road and having to make a turn ahead at 90degrees there are a few ways approach this. You could slam on the brakes at the turn stop then make the turn that would be exact stop mode) Or you could speed up and slide around the curve (acceleration turned down)This gives you rounded corners. Or you could do what most of us do and slow down and slowly make the turn (that's what CV tries to accomplish).

    As for the Phlatprinter we really do not need to worry about this to much because of the gantry being so light. We can turn the acceleration up and turn CV off and we get good cuts.
    However I am still experimenting to see what can be accomplished in using CV. If any of you guys want to jump in and experiment as well that would be helpful to all.
    FYI A lot of this post has excerpts from the Mach forum and are not my own just thought that they would help explain.
    So far I am running my MKII with the CV off and getting great results. But I think it would be nice to see if we can push it, if only to learn more about it :)
    Mark
     
  9. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    Hi Mark,
    This is all really interesting stuff......
    I'd love to get involved in that. The more I can learn, myself, about what makes the beast tick the better. If that then helps other people, that makes it even more worthwhile.

    Any thoughts on the suggestion that running full step and dialling in some PFD on the easy-cnc board might actually help to run the machine at faster feeds?

    Best Regards,
    Julian (Fozzy The Bear)
     
  10. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    Thanks for that link... That also makes interesting reading. I think we might be able to get a lot from that.

    Best Regards,
    Julian (Fozzy The Bear)
     
  11. 3DMON

    3DMON Moderator Staff Member

    Offline
    Messages:
    2,380
    Trophy Points:
    38
    Location:
    Sebastian, Florida
    Mark I am so glad you posted this. I cut at 100ipm because I was getting rounded corners at 150 and higher. I didn't realize I could do anything about it LOL.
    Thanks a bunch!
     
  12. tvcasualty

    tvcasualty New Member

    Offline
    Messages:
    637
    Trophy Points:
    0
    Mark I do not believe I have a CV mode in ECM, I will have to look into it. The issue I had was that the scream of my steppers on long straight cuts went to a nice hum when I adjusted those values in the software.

    The idea was to feed the board / steppers the type of signal it expects.
    This results in better speed and less error correction / missed steps.
     
  13. kram242

    kram242 Administrator Staff Member

    Offline
    Messages:
    6,311
    Trophy Points:
    13
    Location:
    NJ
    Erik, I really need to install EMC2 so that I can see what its settings are and if nothing else see how it works in conjunction with the output from the Phlatscript to help to troubleshoot problems.

    3DMON, your welcome and if you find anything out that would help please post, thanks man


    That's what its all about and thank you for your help.
    Mark
     
  14. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    Ahhh! well..... there's a few bits of Aluminium that I'm cutting to support the flexible shaft drive run from a 750Watt rotary tool, and some ball raced linear bearings that I picked up yesterday "really cheap" for just a few pennies in fact, off e-bay, going in there now. Should run smooth as silk by the time I finish fiddling with it. :D Like I said though, more on that later... I will post pics and descriptions. I'm still kind of wrapping my head around the best way to fit them.

    Best Regards,
    Julian (Fozzy The Bear)
     
  15. kram242

    kram242 Administrator Staff Member

    Offline
    Messages:
    6,311
    Trophy Points:
    13
    Location:
    NJ
    Sounds great Fozzy! Looking forward to your build log
    Sounds like its going to a mean machine! :D
    Mark
     
  16. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    With a bit of luck and based on the torque calculations I did, it should manage to cut 4mm Ply at a sensible speed. Using Titanium bits.

    Best Regards,
    Julian (Fozzy The Bear)
     
  17. kram242

    kram242 Administrator Staff Member

    Offline
    Messages:
    6,311
    Trophy Points:
    13
    Location:
    NJ
    man that would incredible! I really need to see you get this accomplished.
    If you use the foam carrier idea I think you will be able to.
    I was test cutting 1/8" ply the other day that way with a fish tail chipper bit on the MKII and it worked! I had to go back in and make sure the spindle mount did not move at all because there is a lot of torque.
    Mark
     
  18. FozzyTheBear

    FozzyTheBear Member

    Offline
    Messages:
    261
    Trophy Points:
    16
    Absolutely.... That's where the linear bearings and aluminium Carrier for the Z Carriage come into their own. By allowing everything to move freely but without any excess play to distort the cut. It should work fine. I polished up the tool steel hardened rails we got off e-bay today, they're looking good :cool: !!

    I really must start a build log and get some photographs posted.

    Best Regards,
    Julian (Fozzy The Bear)
     
  19. kram242

    kram242 Administrator Staff Member

    Offline
    Messages:
    6,311
    Trophy Points:
    13
    Location:
    NJ
    Yes please do. I am really curious to see how this build turns out.
    Thanks
    Mark
     

Share This Page