Beta Firmware 1.16 - Public Test

BETA, DO NOT USE ON PRODUCTION SYSTEM
This firmware adds distortion adjustment to system setting via CompensateX / CompensateY. For most users, use recommended value is enough. Other settings are the same as 1.15.

System Setup:

X Size 900
Y Size 900
X Deviation 100
Y Deviation 100
Smooth 50
Z Motor Speed 8
PM Motor Speed 30
Advanced Setting

Laser Power 58
XY Speed Set 4
Z Reset Position 1877
PM Reset Position 40
Z Reset Position 40
Compensate X 100
Compensate Y 200

Firmware download:
https://drive.google.com/open?id=1acyCMmZVfjTXCHmD3HkSovGj2zGli0Nu

Firmware upgrade method is same as before

Firmware 1.16 calibration guide:

  1. Record current 1.15 settings, especially if you have any custom settings.
  2. Upgrade firmware to 1.16, refer settings to 1.15 IF you use any custom setting. However, keep X/Y size at 900.
  3. Check center using cross pattern. At the same time, make sure x/y galvo are set correctly, (pushed all the way in)
  4. Run calibration scan circle, ONLY match the outside 120x120mm circle. Ignore inner circle
  5. Print a cube on the outside to check how much the size is off
  6. Use x/y size to correct the difference.
3 Likes

I’ve used the firmware update, after going through advanced calibration and 120 mm I printed a 20mm cube with the default settings(x:900 y:900). The cube was(+/- .03mm). This differed from my previous settings for Peopoly resin (x:928 y:928) to produce the same results. With the previous settings the y would be 1mm larger on the tilting edge and on the left edge and .6mm smaller on the door and back sides. I then started printing with the beta firmware around the edges and noticed that the deviations were significantly reduced. All the cubes were again accurate on the x, but the y were not. With the new firmware th y: was reduced by .3mm +/-.05. I have not measured the new printing area to obtain accurate settings. The new firmware has definitely improved the results of printing accurate price. I’ll see how altering the x/y scaling changes 3rd party resins.

2 Likes

are you using 200 for compensateY? or 195? That is what is affecting the Y-axis (door to back)

The y compensate is set to 200, I believe that was the default value

I will adjust the y compensate to 195 and reprint

I just flashed my machine to 1.16. I can confirm that the Y Compensate value is defaulted to 200. I changed it to 195 per @peopoly original post. @peopoly might want to publish a corrected firmware .bin file if 195 is supposed to be the default value.

I reprinted my cube test with the y compensate: 195. The cubes were placed dead center of the build platform, and two centered on the edges of the the right tilting side and one in the front door.
The front cube x: 20.08 y:19.70 z:19.98
The right tiliting x: 19.45 y:20.41 z:20.03
center: x:19.52 y:19.68 z:19.98

These were printed with the default x:900 Y: 900 scaling. I am about to reprint with the settings I usually use scaling:X: 928 Y:928.

One thing that I noticed was the center was perfect with the default settings of 900 and y-compensate at 200. Lets see how these turn out!

I reprinted with the scaling I normally use for peopoly white, X;928 and Y:928. The cubes were placed dead center of the build platform, and two centered on the edges of the the right tilting side and one in the front door.
The front cube x: 20.48 y:19.99 z:20.03
The right tilting x: 19.86 y:20.74 z:20.01
center: x:19.98 y:20.07 z:20.02

The front and back of the cube is very precise, varying very little about +/-.05. The new firmware is promising, but still needs some work. @peopoly what would you suggest for use to work on to fix the issues?

2 Likes

thanks for the feedback, still need to tweak the algorithm. It took few months of work to get here and we will continue to optimize it. If you need quick fix, scaling in software is likely the easiest way until the firmware is further updated.

Thank you, should we expect the finalized firmware update in a few months?

@drdl1

one thing we found during the testing stage is that users may have some calibration issues and affect the results.

This is the steps we recommend:
Complete calibration guide:

  1. Check center using cross pattern. At the same time, make sure x/y galvo is set correctly, (pushed all the way in)
  2. Run calibration scan circle, ONLY match the outside 120x120mm circle. Ignore inner circle
  3. Print a cube on the outside to check how much the size is off
  4. Use x/y size to correct the difference.

We are going to finalize 1.16 to publish as it is while continuing to improve it and adding new features. The adjustments in 1.16 is enough to offer benefit to those who needed.

To test the correction in the v1.16 firmware, I printed 20mm x 20mm blocks (hollow, with letter label on each) arranged like this on the build plate:
A B C
D E F
G H I

NOTE: I originally calibrated X and Y sizes on my machine for 20mm blocks offset from the middle - so that the inside and outside differences would average closer to the intended size.

Measurements are min and max values measured on different sides of the block. (ie: range of the minimums of each edge)

	v1.15 X			v1.16 X
A	20.50 - 20.55	20.53 - 20.55	
B	19.52 - 19.60	19.58 - 19.60
C	20.82 - 20.84	20.72 - 20.77
D	20.49 - 20.58	20.52 - 20.56
E	19.54 - 19.61	19.61 - 19.62
F	20.73 - 20.78	20.73 - 20.76
G	20.56 - 20.62	20.51 - 20.58
H	19.57 - 19.65	19.62 - 19.64
I	20.75 - 20.81	20.76 - 20.80

Variation in X is somewhat reduced in v1.16, but most measurements are nearly identical (within expected measurement error) between firmware versions.
And there is still a large change from middle (19.6mm) to edge (20.7mm).

	v1.15 Y			v1.16 Y
A	20.47 - 20.77	20.35 - 20.38
B	20.24 - 20.31	20.25 - 20.35
C	20.52 - 20.80	20.33 - 20.35
D	19.75 - 20.05	19.60 - 19.64
E	19.51 - 19.56	19.49 - 19.51
F	19.75 - 19.86	19.53 - 19.57
G	20.28 - 20.54	20.05 - 20.11
H	20.01 - 20.07	20.07 - 20.10
I	20.39 - 20.49	20.10 - 20.12

Variation in Y within each block is significantly reduced in v1.16 (from 0.3mm in v1.15 to 0.1mm in v1.16).
There is still a large change from the middle (19.5mm) to the edge (20.3mm), but it is somewhat reduced in v1.16.

Variation in X (0.1mm) started out smaller than in Y (0.3mm), and is now about the same in v1.16 (0.1mm).
Absolute differences from center to edge are still about the same in X and Y ( around 1mm! ).

2 Likes

I followed instruction it looks like in Kingston upgrade doesn’t work ( https://peopoly.dozuki.com/Guide/Moai+-+Firmware+Upgrade/3 ) when I see comment it looks like I’m not the only one. After 20 seconds I still see 1.15

The Kingston card supplied with my printer works fine for transfering gcode files but not for updating firmware. For this I bought a cheap ordinary SD card (nonbranded) and this works fine for updating. I use this card only for this purpose.

I tried with 3 types of cards:
Kingston - original ( software responding, screen is blank for a moment, but no further upgrade )
ADATA 4g - no reaction
SANDISK 16g - no reaction

make sure you copy all files from the zip file, including that txt file.

You also process change log ? :slight_smile:

it worked for some users based on the feedback.

https://peopoly.net/pages/support still has 1.15 as the latest firmware. Since 1.16 is officially recommended, that page should be updated (it should probably just point to http://wiki.peopoly.net/doku.php?id=moai:firmware)

so every time you update the firmware you seem to have to recalibrate? wow. I don’t have time for that. just saying. always updating the form 2. just update and go