Welcome to Our Community

Some features disabled for guests. Register Today.

Questions about accuracy and how to increase it

Discussion in 'General Talk' started by Stahlschlüssel, Oct 20, 2015.

  1. Stahlschlüssel

    Builder

    Joined:
    Oct 19, 2015
    Messages:
    2
    Likes Received:
    0
    Hello everyone,
    I'm new and I'm looking forward to this world of mini CNC milling machines, but I'm a polymechanic and have been instructed on CNC milling machines and lathes, although I did not use it for several years.
    I would like to inform me about the accuracy of these machines (especially the C-beam that I think that it's the most accurate) and if there is some way to increase it.
    Looking in the forum I found that the C-beam can reach an accuracy of 0.04mm, there is a posibility to improve?
    Is there anyone who has installed a system of coolant?
    Nobody works the steel?
    Sorry but I don't speak English, maybe this information is already in the forum, if so sorry.
    Thanks for the help
     
  2. Rick 2.0

    Rick 2.0 OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Dec 20, 2013
    Messages:
    2,891
    Likes Received:
    1,545
    Please understand these are hobby grade systems and are not designed to supplant the machinery found in industrial machine shops. They are designed for cutting wood, plastics, and soft metals and are not really built stout enough for cutting anything any harder. As far as greater precision, this can be achieved with smaller microstepping increments and higher powered stepper motors and power supply to compensate for the power losses of the increased microstepping.
     
  3. Ronald van Arkel

    Staff Member Builder Resident Builder

    Joined:
    May 7, 2014
    Messages:
    472
    Likes Received:
    245
    What Rick said, but would like to add that you can work steal with a plasma torch ;). I understand that's not what Stahlschlüssel is asking but I can say that accuracy depends more on what kind of tool/drill/head you are using. A not so well balanced setup can be out of range for over 0.5mm from what I've seen. The precision, distance difference, is what Rick said and must not be confused with accuracy (language barrier, I know the feeling (Dutch here)).

    -Ronald
     
  4. Stahlschlüssel

    Builder

    Joined:
    Oct 19, 2015
    Messages:
    2
    Likes Received:
    0
  5. Ronald van Arkel

    Staff Member Builder Resident Builder

    Joined:
    May 7, 2014
    Messages:
    472
    Likes Received:
    245
    Steelkey,

    The first one is sturdy, and it should be able to do 0.1mm accuracy on soft materials and metals, not on steel. It's expensive and to tell you the truth you as a polymechanic should be able to build one easily yourself! I would start with a C-Beam machine, improve it, build another one with larger improvements when you are used to the C-Beam.

    The second one I would not touch even with a stick. Sure there will be people happy with it but it's just made to cut wood as the frame and base are just wood. They talk about resolution, not about accuracy nor precision, and is nice in sales talk but practically useless as it only applies to the electronics; next step will be that they say it can do 0.000001 μm as the software can do it. For example, that small square of 1"; when I build the cheapest router I could build with OpenBuilds hardware I got 99.7mm square in both directions cut while the drawing said 100mm. I used a Dremel tool as a head and a 1/8" 4 flute round point bit/drill that isn't made for wood so I got a 3.6mm wide cut while it should have been 3.175mm, a difference of 0.425mm, divided by 2, 0.2125mm plus the 99.7mm makes it 99.9125mm... and that on the first run without tweaking! This is great sales talk but it totally depends on how well all is build and how good the head is ;).

    Now, what do you want to do with the small CNC you are going to buy or build? But what ever you do, build it yourself! :thumbsup:

    -Ronald
     
  6. R_B

    R_B New
    Builder

    Joined:
    Jul 18, 2015
    Messages:
    70
    Likes Received:
    12
    Accuracy, precision, repeatability - as the old joke goes, pick any two.

    Here is an odd rounding/approximation thing I found with my C-Beam machine in combination with grbl-panel.
    It isn't at all clear (to ME) where the discrepancy arises, although I suspect that it is in grbl-panel.

    I have been working with a cheap vice (vise) from harbor freight, mounting it directly on the Y-axis carriage plate.
    I decided that a square bolt pattern would be a good idea, since that would allow me to rotate it 90, 180, or 270 degrees.
    OK, so since the vice is sold in the USA and the carriage plate is also made in USA it seemed to make sense to switch to imperial measurements - inches.
    The plate is 8 1/2 inches square, the mounting slots on the vice are just about 5 1/8 inches apart.
    I thought it would be a good idea to mark the centers for the holes with a drill bit in the router and to use the machine to mark the hole centers.
    I "touched off" at the bottom left corner of the plate, cleared X, Y and Z to zero, unchecked metric in grble-panel.
    I then advanced to X-8.5 Y-8.5 just as a sanity check.
    It was approximately a mm "long" in both directions, the panel read 215.90 in both X and Y.
    I brought it back to 0,0 even slower, it came back to exactly the touch off point.

    OBVIOUSLY the next step was to set metric and repeat with measurements of 215.9, 215.9 - - that brought me EXACTLY to the corner diagonally opposite the touch off corner.

    I don't know where this happens, I "assume" it is due to some rounding UP in the conversion from inches to mm - - I further "assume" that the machine is working TOTALLY in mm and that the imperial measure is just a convenience front end panel "feature".
    Anyway, point is there seems to be about a 0.5% error when using inch measurements with grble-panel and the suggested settings on C_Beam plate maker.

    This is more, MUCH MORE than the error of the 25.4 mm/inch approximation.
    It is enough to make parts that won't fit together - it probably doesn't matter too much for carving signs though.
     
  7. Ronald van Arkel

    Staff Member Builder Resident Builder

    Joined:
    May 7, 2014
    Messages:
    472
    Likes Received:
    245
    R,

    This might be a bit offtopic here but you get 215.90 in the GRBL pannel and 216.90mm with the machine? I would say you have to adjust/tweak your resolution by a decimal more, eg. change steps per mm from 200.0 to 199.08 (example, 8mm Acme and 1/8 step dev.).

    -Ronald
     
  8. R_B

    R_B New
    Builder

    Joined:
    Jul 18, 2015
    Messages:
    70
    Likes Received:
    12
    Accuracy, precision, repeatability - as the old joke goes, pick any two.

    Here is an odd rounding/approximation thing I found with my C-Beam machine in combination with grbl-panel.
    It isn't at all clear (to ME) where the discrepancy arises,although I suspect that it is in grbl-panel.

    I have been working with a cheap vice (vise) from harbor freight, mounting it directly on the Y-axis carriage plate.
    I decided that a square bolt pattern would be a good idea, since that would allow me to rotate it 90, 180, or 270 degrees.
    OK, so since the vice is sold in the USA and the carriage plate is also made in USA it seemed to make sense to switch to imperial measurements - inches.
    The plate is 8 1/2 inches square, the mounting slots on the vice are just about 5 1/8 inches apart.
    I thought it would be a good idea to mark the centers for the holes with a drill bit in the router and to use the machine to mark the hole centers.
    I "touched off" at the bottom left corner of the plate, cleared X, Y and Z to zero, unchecked metric in grble-panel.
    I then advanced to X-8.5 Y-8.5 just as a sanity check.
    It was approximately a mm "long" in both directions, the panel read 215.90 in both X and Y.
    I brought it back to 0,0 even slower, it came back to exactly the touch off point.

    OBVIOUSLY the next step was to set metric and repeat with measurements of 215.9, 215.9 - - that brought me EXACTLY to the touch off corner.

    I don't know where this happens, I "assume" it is due to some rounding UP in the conversion from inches to mm - - I further "assume" that the machine is working TOTALLY in mm and that the imperial measure is just a convenience front end panel "feature".
    Anyway, point is there seems to be about a 0.5% error when using inch measurements with grble-panel and the suggested settings on C_Beam plate maker.

    This is more, MUCH MORE than the error of the 25.4 mm/inch approximation.
    It is enough to make parts that won't fit together - it probably doesn't matter too much for carving signs though.


    Thanks for the feedback Ronald,
    Yes, it is approximately a mm of extra travel - I didn't measure it exactly.
    If I intended to work ONLY in inches it would be worth doing, in fact it would be THE way to go for anyone working in inches (with grbl-panel and C-Beam machine).
    Perhaps grbl-panel should load out those settings when metric is deselected ?
    Unless/until it does I suggest/recommend that anyone working occasionally in inches do all the conversion arithmetic OFF the machine and work in mm ON the machine.
    My point is that the machine with grbl-panel is inconsistent between metric and imperial measure, i.e. just checking/unchecking the metric box in grbl-panel won't get you accurate parts.
    A 1 inch feature on one part will be unlikely to fit into a 25.4 mm opening in another part, etc.

    Given that the movement mechanisms are based around a metric lead screw, I will stay in metric.
    Odd that openbuilds used 8 1/2 inches for the plate (-:
    Accidental that I grabbed an inch rule when I measured the plate (-:

    Lazy individual that I am; it was quicker, simpler, easier to (jog 1 inch 8 times then 0.1 inches 5 times) than to (jog 1 cm 21 times and then a mm once, then 0.1 mm 9 times).
    Lazy, but awake enough to notice that it wasn't quite right.
     
    #8 R_B, Oct 22, 2015
    Last edited: Oct 23, 2015

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