M2 Accuracy -- What's reasonable to expect?

General discussion topics
BBID
Posts: 36
Joined: Sun Jun 04, 2017 4:47 am

M2 Accuracy -- What's reasonable to expect?

Post by BBID » Mon Jun 19, 2017 11:35 pm

IMG_1841.JPG
I've been printing to some 20mm calibration cubes to see what kind of accuracy I'm getting out of my M2 Rev. E. The following settings were used:

Material: Hatchbox PLA, Silver
Extruder Temp: 200C
Bed Temp: 60C on Buildtak
Print Speed: 60mm/s with 50% underspeed for perimeters and solid infill (fastest I can print without warping).
Retraction: Default S3D retraction settings for PLA
Slicer: Simplify 3D
Nozzle: Standard 0.35mm (Brass)
Cooling Fan: ON @ 100%

So far, the best cal. cube dimensions I've got out of the M2 are 20.4 x 20.21 x 19.94 mm (x, y, z). Makergear states that the steppers are accurate within 1/80th of a mm, which would be 0.0125 mm. I'm not sure how this translates into dimensional accuracy for printed objects, but it would appear the calibration cubes I'm printing aren't within spec. Members of this forum have posted that commercial machines like the M2 have printed, dimensional tolerances no greater than +/- 0.20 mm. I guess my questions are:

1. What's reasonable to expect accuracy-wise from the M2?

2. What's with the repeated linear pattern on the faces of the calibration cube?

Thanks for any input!

User avatar
ednisley
Posts: 1188
Joined: Fri Apr 11, 2014 5:34 pm
Location: Halfway up the Hudson
Contact:

Re: M2 Accuracy -- What's reasonable to expect?

Post by ednisley » Tue Jun 20, 2017 12:53 am

BBID wrote:What's reasonable to expect accuracy-wise from the M2?
In round numbers, day-in and day-out, features will be within ±0.2 mm of their nominal dimension. You can tweak the results for better accuracy on any particular model, but don't expect that to transfer to the next model.

Until you get the hang of it, your models will have issues with small holes (they'll be too small), openings with overhangs (they'll be too short), and thin walls (they won't fill correctly). None of those are killer problems, they're just how fused plastic printers work.

Always remember: you're producing an object by squishing molten plastic into a flat layer, not machining metal.
What's with the repeated linear pattern on the faces of the calibration cube?
Perhaps a slight temperature variation as the PID loop hunts around the 200 °C setpoint, a slight variation in leadscrew polish, or a tiny periodic variation in the filament diameter.

You can see surface finish differences far smaller than you can measure: a perfect surface finish (well, apart from the usual thread layer lines) on an as-printed part simply isn't practical.

BBID
Posts: 36
Joined: Sun Jun 04, 2017 4:47 am

Re: M2 Accuracy -- What's reasonable to expect?

Post by BBID » Tue Jun 20, 2017 2:01 am

Thanks for the feedback ednisley!

I've gotten used to reaming holes to get accurate dimensions, which is why I routinely undersize them in CAD. I just wasn't sure about what to expect with dimensional tolerances in x/y with the M2. My Flashforge Creator Pro gets about the same +/- 0.1-0.2 mm x/y tolerances, as does the Original Prusa i3 MK2S according to the user base. I guess I was expecting to get tighter tolerances with the M2's linear rail system, but it looks as if the real advantage with the rails is less ringing as higher print speeds, not more accuracy.

I never thought of it before, but your PID explanation regrading surface finish makes total sense. The pattern seems too regular to be a lead screw defect, or filament diameter variation. I know from experience that printing this brand of PLA at higher temps produces a glossier finish in the final part. Watching the LCD I can see temps fluctuate +/-1-2C.

User avatar
ednisley
Posts: 1188
Joined: Fri Apr 11, 2014 5:34 pm
Location: Halfway up the Hudson
Contact:

Re: M2 Accuracy -- What's reasonable to expect?

Post by ednisley » Tue Jun 20, 2017 5:22 pm

BBID wrote:tighter tolerances with the M2's linear rail system
IMO, those rails buy you more mechanical stability: after you get the machinery aligned, it stays that way.
dimensional tolerances in x/y
The overall accuracy limit comes from dragging a thread of hot gooey plastic across the platform, due to both its squishiness and shrinkage as it cools. The platform holds the bottom of the part at almost the exact XY coordinates set by the stepper accuracy, the top of the part shrinks a bit more against free air, and the middle is, well, somewhere in the middle.

Hole tolerances depend on their orientation: horizontal is different than vertical, diagonal is a whole 'nother conversation, and low-vertex-count polygons work better than lovingly crafted circles. It's not simple!

BBID
Posts: 36
Joined: Sun Jun 04, 2017 4:47 am

Re: M2 Accuracy -- What's reasonable to expect?

Post by BBID » Sun Jul 16, 2017 3:40 am

PID Tuning Results.jpg
PostPIDFanOFF.jpg
For those of you that are interested, I was able to get a little better dimensional accuracy in the calibration cube by adjusting my extrusion multiplier till I was getting exactly the extrusion width I was specifying in S3D. The thing that's driving me insane now is trying to get rid of those lines in the print walls. I've tried everything: ran PID autotune, tried turning the cooling fan off, tried changing layer height, tried increasing print speed to 80 mm/s (which just produces warped prints)....nothing works. I've reached out to Makergear for a solution and hopefully they'll get back to me next week.

I really am getting frustrated with the M2. I thought with a little tuning this machine would produce much better prints. While I would agree that no FDM printer produces perfect prints, I have printed cal cubes on a Flashforge Creator Pro, Ultimaker 3, and a Markforged Mark II, and have seen that they produce excellent quality prints at much faster speeds. With this in mind, I have adjusted my expectations for the M2 accordingly. The M2 is appears to be capped at 40 mm/s for PLA if you want useable prints. Have other users experienced a similar limitation?

User avatar
zemlin
Posts: 467
Joined: Thu Feb 04, 2016 9:32 pm
Location: Indianapolis, Indiana

Re: M2 Accuracy -- What's reasonable to expect?

Post by zemlin » Sun Jul 16, 2017 8:13 pm

I developed a scaling process which enables me to consistently get within +/- .1mm, usually closer, on large and small parts.
On large, heavy parts I believe the weight of the part causes the platform to deflect a bit, so I have had the Z-dimension vary by as much as .75mm from one side to the other on a large .65KG part. With what I've been building that has not been a problem. On smaller parts I don't have any issue with Z accuracy.

viewtopic.php?f=3&t=3885

BBID
Posts: 36
Joined: Sun Jun 04, 2017 4:47 am

Re: M2 Accuracy -- What's reasonable to expect?

Post by BBID » Sun Jul 16, 2017 10:46 pm

Thanks for the link to the calibration cross Zemlin. I'll definitely give that a try. I haven't had too many issues with z-accuracy, other than the fact that it's always a bit undersized because the first layer is slightly squished to get good bed adhesion. Calibrating my extrusion multiplier actually went a long way to improving dimensional accuracy and cleaning up my seams/zippers. What I haven't solved are (1) the lines in my print walls and (2) printing at higher speeds while achieving good dimensional accuracy.

Let me ask the question a different way...has anyone been able to print calibration cubes at 80 mm/s on an M2 with stock cooling and still been within a +/- 0.1mm spec? I'm trying to figure out if there's something wrong with my M2, or if this is just the limitation of the hardware.

User avatar
zemlin
Posts: 467
Joined: Thu Feb 04, 2016 9:32 pm
Location: Indianapolis, Indiana

Re: M2 Accuracy -- What's reasonable to expect?

Post by zemlin » Sun Jul 16, 2017 10:55 pm

I print ABS at that speed (no cooling at all) and get good results.

BBID
Posts: 36
Joined: Sun Jun 04, 2017 4:47 am

Re: M2 Accuracy -- What's reasonable to expect?

Post by BBID » Sun Jul 16, 2017 11:40 pm

I print ABS at that speed (no cooling at all) and get good results.
Are you printing ABS with an enclosure? Can you print with PLA at the same speed and accuracy as ABS?

User avatar
zemlin
Posts: 467
Joined: Thu Feb 04, 2016 9:32 pm
Location: Indianapolis, Indiana

Re: M2 Accuracy -- What's reasonable to expect?

Post by zemlin » Tue Jul 25, 2017 5:55 pm

BBID wrote:
I print ABS at that speed (no cooling at all) and get good results.
Are you printing ABS with an enclosure? Can you print with PLA at the same speed and accuracy as ABS?
I do have my printer in an enclosure.
I do very little with PLA, although when I have there hasn't been any issue with accuracy. I have printed large parts without issue.

Post Reply