Hello, I had a post on this forum about one year ago after buying my workbee. After a lot of issue about the fact that I never have something precise, I had received an upgrade from belt driven cnc to screw driven cnc. I am using a 1500mm x 1500mm. Now that I have a little bit more experience with it I have been able do some stuffs like game board with it. But for real wood working, I have never been able to have something precise. I have a shift of at least 1-2mm so it stay unusable until now. I have rerun a serie of test today and I have notice that the movements without machining anything or when I am machining some stuffs are really different. I have done the same test on the 3 axes but I have the same issue. My previous setting was: M92 X399.44 Y399.44 Z403.836 ; Set steps per mm With this parameter I have the following results when I am moving on the X axe without any load: Theory | Practical 100 | 100 500 | 500.4 700 | 700,6 So I should just adjust the parameter a bit lower to be perfect but I am almost there But when I am machining some stuffs (I have attached the nc file it is just a serie of squares) I have the following values: Theory | Reality | Delta 10 | 9,66 | 0,966 20 |19,7 | 0,985 30 | 29 | 0,966667 40 | 39,2 | 0,98 50 | 48,16 | 0,9632 60 | 58,4 | 0,973333 So here it is also pretty consistent but unlike the previous test I need to set the parameter to a higher value. I have done a second test with the following values who should correct the machining values: M92 X410.478 Y408.147 Z403.836 ; Set steps per mm and indeed the values are now correct during the machining Theory | Reality | Delta 10 | 10,08 | 1,008 20 | 19,99 | 0,9995 30 | 29,98 | 0,999333 40 | 40,17 | 1,00425 50 | 49,54 | 0,9908 60 | 59,69 | 0,994833 But of course the result without load are very disappointing: Theory | Practical 100 | 103 200 | 205,8 500 | 514,2 700 | 720 I have use some lockit and be sure that the the motors do not slip as advised within my previous issue. Any ideas of what I could do for it? REgards, Benjamin v.
Hi Ben, I have a SPAM 1500 X 1000 Workbee using the Open Builds black box controller.....and am wondering if values for the steps should remain 200 on my lead screws is this what you are referring to?
One thing that is important when actually cutting is knowing the exact size of your endmill. The seller may say it is a 6mm endmill, but in reality it may be larger or smaller. If so, that will affect the accuracy of your parts. I measure and record the diameter of every endmill I have before I use it the first time. I had a cheap Ebat 10 pack of 3.175mm endmills. They ranged in size from 3.155mm to 3.17mm. None of the was 3.175mm. This lead to the holes I was cutting being to small so the bolts would not fit through the hole. Have you calibrated the machine? If using OpenBuilds Controller, there is a calibration wizard. If you are using some other controller do this:
Hello, About the calibration itself, I use the second method. I have improved it by adding a webcam on my phone and zoom in for a better precision but it is about the same method. About the bit diameter I have measure it and it is 4.97mm in place of 5mm so not too bad. I have rerun a test with this correction to be sure (I have reset the M92 parameter to the first test values because it is where I have the best precison without load on it) M92 X399.44 Y399.44 Z403.836 ; Set steps per mm Theory | Reality | Delta 10 | 9,6 | 0,96 20 | 19,52 | 0,976 30 | 28,92 | 0,964 40 | 39,14 | 0,9785 50 | 48,15 | 0,963 60 | 57,95 | 0,965833333333333 So the average delta on X was 0,972366667 and it is now: 0,967888889 so even worst. On the Y axe it was: 0,978202778 and now: 0,974319444 so about the same. For Stacey Miller, Sorry but I don't understand exactly what you mean here... REgards,
If you are sure its not mechanical in nature, still not willing to try a Grbl based controller? Can't say for sure it will solve it, but i will say we don't see Grbl users with issues like this (except mechanical reasons) or if it doesnt at least it rules out the electronics side of things (and puts focus on mechanical) I recall your old thread and I think you did check the mechanics pretty good by now. Docs.openbuilds.com/blackbox?
Well, I am not sure at all that the issue is not mechanical. I have bought my machine at Ooznest and apparently they are selling machine with non Grbl based controller without those issue so I guess that it doesn't come from the software. I guess that if the software say to the controller turn from 2 turns it will turn for 2 turns it will stay the same whatever the controller right? If I where pretty convince that the blackbox could solve my issue I will go for it quickly but for the moment I don't think that it is my issue. I am really hesitate to invest more in this machine as long as it stay useless because of this precision. About those 200 you mean as parameter of the M92 command? Like this: M92 X200 Y200 Z403.836 ; Set steps per mm
At least that mean you are feeling its leaning toward mechanical. Always hard to determine from here (far away) so good thing then RE 200: The screws are manufactured using a rolling process so you still need to calibrate. 200 puts you in the ballpark, but do calibrate
Well I was a bit skeptical about the fact that I should set the parameter to 200 but I have tried it but as expected it is now moving from the half of the distance. So when I tell him to move from 200mm it only move from 100mm. I have double checked the default configuration from ooznest to see if I had missed something and the default parameter for it is 400. If I am right the following command say the number of steps available to perform a full rotation of the stepper right? M350 X16 Y16 Z16 I1 ; Configure microstepping So I guess that your default configuration with the M92 paramter set to 200 correspond to a M350 command set to 8? Am I right? Any other idea of stuffs that I should test / adjust to fix my in working issue? Would it help if I tighten more the wheels or the opposite or it isn't the source of my issue?
The 200 rule is for our default configs (1/8th microstepping) You probably then run 1/16th (so 400 steps)