I know that the latest S3 grbl controller comes with everything pre-configured for the machine. But I’m curious if anybody has tweaked the $100, $101, $102 and $0 variables with any positive result?
I suppose it’s a function of how consistent the steppers on each axis are, but curious what the experience of others has been.
Here’s a writeup @WillAdams did a while back. This should get you going in the right direction. Obviously this isn’t mandatory, but if you’re looking to get that extra thou from your setup, doing a ‘real world’ steps/mm is a good way to get that.
-Edward
Awesome! Thanks @edwardrford! I’ve got a stationary dial indicator, I wonder if using that and jogging a fixed distance, then calculating the error would accomplish the same thing, without the need to mill?
Absolutely. You can use a dial indicator for all 3 axis, calculate the difference between actual and GRBL then apply to the settings.
Would be interesting if you did a before and after. Maybe cut a square and a circle then calibrate and re-run the same job to see what differences it actually makes.
So, you’ve sort of seen through to my goal here: determining if doing straight bit inlay without having to offset the size of my pocket or plug is possible, given enough accuracy. I’ll run a job to create a basic pocket and plug, “tune it”, then rerun and see if the fit changes.
I’m anxious to see your results - keep us posted!
Just to demonstrate how brain-dead I am this morning, I thought the title of the thread was discussing improvements that cost $100, $101, and $102, respectively, and a freebie at $0.
And yes, I’ve adjusted all those parameters before.
The coffee just isn’t taking.
I have!
X and Y Axes:
I have my machine belt calibrated to cut bores (round) within a few tenths (refresher course: that’s 0.0001") of an inch, and oddly enough I purposely have my machine set up to cut small ID’s Large OD’s (about 2%) so that I can sneak up on tight (bearing) bores.
Z Axis: Never needed any adjustment, it was and still is spot on.
$0, never played with this setting.
I adjusted them once, then adjusted them back. Didn’t make a difference for what I was doing and I like periodically retensioning the belts, and would expect to have to recalibrate it each time if so.
I use my 6 Inch calipers to check the X and Y distances (since nobody has given me a set of Joe Blocks yet). What I do is clamp some of my beautiful aluminum 4" clamps at some set distance and then use a test indicator to check machine value vs actual (you can use a set of 123 blocks too). This is a lot easier (for me) than cutting a CDS every time I want to check the machine.
Hope this gives you a new and easy method for checking the machine.
This ^^^.
With a squared machine, trammed spindle, surfaced wasteboard, and final dial calibration my XXL has been accurized to where I can hold very respectable part tolerances.
Alright, results.
Measured with a digital dial indicator on all 3 axis. Did 3 sets of measurements for each and averaged, then used the average to find the ratio * 40 and am now using the following;
X = 40.226
y = 40.206
z = 40.619
Naturally, I failed to do a plug test before I dialed these in, but I can easily enough switch back to the 40.0 defaults and then test again.
I’ve got a big vcarving project in the works, I’ll report back if this calibration fixes any of the small oddities I’ve been seeing.
Thanks for the help all!!!
Getting slightly off topic here, but in doing the measurements needed for this, I found that a given movement had a .002-.003 variability. That is, moving X .100 didn’t always move it the same amount. Sometimes .101, sometimes .098, etc. Same occurs for all axis’. Could be inaccuracy in my $50 digital guage. Am I expecting too much precision out of the S3 and grbl?
There’s a finite point at which Grbl rounds off movement. Usually if you rig the movement to always be from the same direction and distance it will consistent round off.
If you’d provide the exact numbers and directions of movement and what you need we might be able to help more.
Some example X jogging numbers, all in mm, stepping to the right 1mm at a time:
0, .93, 1.93, 2.92, 3.91, 4.94, 5.93
Now incriminating from that position back to the left, 1mm at a time:
5.93, 4.96, 3.95, 2.95, 1.96, .96, 0
The above were taken with my “corrected” 101,102 and 103 values. Interestingly, the same behavior is seen with standard values of 40,40,40. Further strangely, grbl no longer reports a movement that matches the jog. If I zero out then jog x by 1mm, it instead reports a movement of .994mm (though this number varies, as do the ones above).
Any resolution with this? I’m going through a similar situation.
I have not 
I haven’t found much practical impact of this apparent lack of precision though. I wasn’t using a $200.00 dial indicator, so it’s possible my measurements were the things that were off, not the steps? No idea.
An inexpensive digital one will work — got mine at Harbor Freight either on sale or w/ a coupon:
(battery life seems to be hit or miss and random as to whether one gets a good unit, or one which eats batteries)
see https://www.shapeoko.com/wiki/index.php/Tools#Optional_tools
This one is popular and well-regarded:
To be honest though, I usually grab my Mitutoyo:
I’ve got a digital dial-indicator in that same price range - if that’s accurate enough, then I’m out of ideas for the random jog distances :-/
Random movements are typically caused by: