Calibrating your railcore for accurate print dimensions

questions and tips related to specific components

Topic author
Posts: 26
Joined: Mon Dec 31, 2018 8:36 pm
Printer Variant: RailCore II 300ZL
Been thanked: 3 times

Calibrating your railcore for accurate print dimensions

Post by justinsgray » Mon Jan 07, 2019 12:40 am

Step 1) Calibrate your extruder. Its helpful to have a fixture of some sort to aid in measuring the filament. I like this one from thingiverse:

The goal here is to get the steps per mm for the extruder correct. These depend on a variety of factors like gear ratio and design of the hobbed gears. You can get close with just some dimensions math and a steps-per-mm calculator, but you have to do the fine-tuning for your specific setup.

Step 2) You calibrated your extruder right? If not, go back and do it. The rest of this is pointless if thats not correct.

For my 300ZL with a bond-tech extruder and 1.75mm PETG filament I ended up with 884 steps per mm. You can expect a little variation with different filaments, but not that much except if you switch between rigid and flexible filaments. There you might see a decent variation.
Just command a certain amount of extrusion (say 50 or 100 mm), and measure how much you really get. Then multiply your current steps-per-mm by the ratio of what you expected to what you got.

You can start with a really really wrong steps-per-mm and still get there. Here is the history I had, starting with a horrible value of 490 steps-per-mm:

Commanded 50 mm extrusion:

Actual | Steps-per-mm
--------- ---------------------
28.63 490
41.14 699
46.8 823
48.98 876
50.8 893
49.47 883
50.2 884

Once you find the right value, add it to your config.g

Code: Select all

M92 X200 Y200 Z1600 E884	    ; steps/mm, X/Y may be more around 201.5 for accuracy
The comment in that line is very interesting. The reason you have to fine-tune the extruder steps is that filament is squishy and when it gets squeezed in between a roller and the hobbed gear (or two hobbed gears for a bond-tech) you end up with slightly different radii than you might have used in a computation based on the diameter of hobbed gear alone. But for the x/y/z axes this issue shouldn't come up.

The railcore uses 1.8 degree steppers, with 1/32 microstepping and a 2mm GT2 belt pitch with 16 tooth pulleys for x and y. According to ... spermmbelt that gives 200 steps per mm. The only reason I can see why we should not have exactly 200 steps per mm is if the belt pitch is slightly off for the cheaper pulleys we use (at least cheaper compared to the real gates stuff).

Belt pitch is the distance between low-spot to low-spot or high-spot to hight spot on a belt. There could maybe be some very slight variation in that along the length of the pulley that would contribute some inaccuracy. It can't be too far off or the pulley and the belts wouldn't mesh up. Alternatively, the pulley's could have a variation in the pitch on them. Or it could be some combination of the two effects. One last possible effect could be the impact of belt stretch...

What I should probably should do is make some kind of fixture to mount my calipers to. Then I can move the axes a precise amount and measure what the calipers say the distance was... im thinking something that mounts to the linear rail and clamps to the caliper. That way I could check x and y.

Short of that, the next best thing to do is to print an object of expected dimensions and measure it. Since this is really what you care about anyway, you might think this is the right place to start. But, IMO you'd be wrong :) There are many factors that effect the final print dimensions. The two most important are your steps-per-mm for the x/y/z axes and the nozzle extrusion multiplier. (The steps-per-mm for the extruder are really important too, but you already calibrated that... right??) By just measuring a print, it can be hard to pin point what is going wrong and why.

Also, FDM 3d printing is basically using a robot arm to accurate position the equivalent of a tiny hot glue gun as it squirts out plastic. There is only so much precision you can expect from this. Anything from variations in the filament diameter, to changes the air temperature, variation in the hot end temp, or small impuritied in the filament, will all have an impact. If I can get within 0.1 mm, Im happy. For any machinists out there, thats about 0.004 inches. Hitting that kind of tolerance on a mill or a lathe takes a bit of effort!

So even though I think measuring the print is the wrong approach, I did it anyway. :) I'll still go back and double-check the calibration of the x and y axes on their own. But since I don't have a model of a clamp for the calipers yet, I just started by calibrating the extrusion multiplier and using the data to sanity check the axis steps per mm.

I printed out a series of 1 wall thick boxes of various dimensions (20x20, 40x40, 20x60, 60x60).
Screen Shot 2019-01-06 at 1.43.27 PM.png
Screen Shot 2019-01-06 at 1.43.27 PM.png (83.86 KiB) Viewed 1326 times
So I printed all these various boxes and then measured both their dimensions and the wall thicknesses.

On my first test, you can see that the physical dimensions we're pretty close but the wall thicknesses were way off. More importantly, the error in the dimensions did not seem to be proportional to the model size. There is some variation, but it doesn't directly correlate with model size. If there was fundamentally an incorrect steps-per-mm on the x/y axes then you would expect the error to be proportional to size. Since it wasn't here, Im fairly confident that the correct steps-per-mm is 200.
Screen Shot 2019-01-06 at 7.32.25 PM.png
Screen Shot 2019-01-06 at 7.32.25 PM.png (108.73 KiB) Viewed 1326 times
On subsequent test cases, as I lowered the extrusion multiplier the wall thickness error comes down, but the measurement accuracy stays roughly constant. Thats about what I would expect if my axis is set up correctly. I ended up with a really low extrusion multiplier of 0.82. This seems tiny to me, but its what the calibration says and it gives good results.

Posts: 4
Joined: Wed Aug 21, 2019 11:52 pm
Location: Syria
Printer Variant: RailCore II 300ZL

Post by BeJimmiekes » Tue Sep 03, 2019 6:18 am

SO terry, if I wanted to have wireless access, and I got one of these...
I would still need to pick a wireless router right? they have some of these on sale today at woot as well.

Posts: 4
Joined: Tue Sep 03, 2019 10:02 am

Post by 247pharmacymart » Tue Sep 03, 2019 10:18 am ... 3FCB4?n=30 ... nzs9-dwfmq ... pular.html ... etic_drug/ ... =undefined ... on-section ... 6875780097 ... ysfunction ... p1AFtWq6EQ ... og%2F13077 ... tic%20Drug!M2NkZTYy:impotence-tre ... erm=button ... th-fildena ... 1b4c137c91 ... afil-100mg ... afil-100mg ... afil-100mg ... afil-100mg ... afil-100mg ... afil-100mg ... nal-brand/ ... nal-brand/ ... nal-brand/ ... nal-brand/ ... nal-brand/ ... 05e7c88dbf ... isawillson ... hetic-drug ... nal-brand/ ... inal-brand ... -mg-online ... -mg-online ... -mg-online ... -mg-online ... -mg-online ... -mg-online ... -mg-online ... -mg-online ... -mg-online ... -mg-online ... ildenafil/ ... id=1267680 ... 3&t=573441 ... 4&t=573442 ... le;u=81267 ... &uid=13775 ... ?tid=46089 ... al-health/ ... g-tablets/ ... fault.aspx ... &u=1106219 ... on=profile ... Sildenafil_ ... tid=130895 ... tid=130902 ... comments/1 ... ry;u=37989 ... file&u=609 ... e&uid=3422 ... ile&u=4028 ... le&u=26222 ... &t=2717095 ... &t=2717096 ... 82371.html ... 30094.html ... sage236165

Post Reply