www.欧美日韩,日韩av一区二区三区四区,青青青在线精品,欧美成人激情综合网

HI~歡迎來到數(shù)造云打印平臺!

設(shè)計(jì)師認(rèn)證

您的設(shè)計(jì)師身份尚未確認(rèn)

設(shè)計(jì)師身份確認(rèn)后方可提現(xiàn)。
現(xiàn)在去確認(rèn)嗎?

您的位置: 首頁 > 商品詳情

三維臺階測試件

3D打印機(jī)/測試件

926 0 0

模型下載詳情

  • 152800695614972.stl (0.01M)

全部1個子文件

材料選擇

暫無

后處理

暫無
顏色分類
  • 默認(rèn)色

訂單配置

  • 材料:暫無
  • 顏色:默認(rèn)色
  • 后處理:暫無
數(shù)量:

?預(yù)計(jì)交貨時(shí)間:暫無

3D打印老專家

為您推薦

這個簡單的試件可以測試XYZ方向的線性精度。每個臺階5mm高,XYZ方向25mm;


Description:
Object overall x, y, z dimension = 25mm.
Each layer of cubes is 5mm high.
Base layer cutout: 5mm high, 5mm deep, 10mm wide.
2nd layer cutout: 5mm high, 5mm deep, 5mm wide.

可檢測如下幾個方面:
1. x,y,z行走精度:
Once you have the correct E-Steps to match your mechanical setup, I found that the next two issues that affected linear accuracy were: Temperature & Extruder E-Step Fine Tuning. See the linked Calibration guide above. Fine tuning my Extruder E-Step calibration was the difference between an accuracy of +0.4mm to +0.5mm and +/-0.1 to 0.05mm.

2. 角度精度:
Once you have aligned your printers axis to be perpendicular, this object is of limited use. I found it easier to print out circular objects to test for angular accuracy, as now I was finding that one axis was sliding at a different rate to the other, resulting in ellipses, or squashed circles. Improvements from here focused on reducing the friction in the axis and improving the belt tension.

  • Reducing axis friction (specifically for Reprap Prusa Mendel):
    First - Ensure that your guide rods are parallel and your carriage doesn't bind up on any point. Your build instructions will have instructions on how to set this for your printer, but on mine there were little grub screws for adjusting the location of the rods. Use a treadlocking compound (like Loctite) or spring washers to ensure they don't come loose under vibration.
    Second - Reconsider your y-axis carriage. The Makergear Reprap Prusa Mendel y-axis carriage has a tendancy to bind up on the y-axis when you are leveling the bed. Rearrange the fastener/spring arrangement to rectify this. From the side, working from the top down, my solution was: Screw head (M4x40 socket head) - M4 Washer - Wooden Build Platform - M4 Washer - Spring - M4 Washer - ~6mm of spacers (I used more washers) - Plywood Y-axis carriage mount - M4 Washer - M4 Locknut. This arrangement removes the tendancy for plywood carriage to warp and bind when your are levelling the bed, provides a degree of give to the build platform should you drive the head into it, and gives sufficient tension to the mount that I haven't had to relevel the bed in 6 months since I switched over.
  • Improve Belt Tension: Add the spring from some old school wooden clothes pegs to your belt. Pic - http://www.iheartrobotics.com/2009/11/timing-belt-tensioner.html  - The new Makerbot Replicator has a similar spring to tension their belts.

Bridging:
The cutouts help with checking your small bridging settings, and slicing settings for overhangs: These were added more for interest than anything else. To properly calibrate your bridging/overhang settings, I recommend printing some dedicated bridging calibration pieces such as: http://www.thingiverse.com/thing:9804 or http://www.thingiverse.com/thing:12925

Slicing settings:
If you are finding that the top layer or two are being printed as solid cubes and bulging at the sides, see if you can adjust the minimum solid infill area setting on your slicing program. I use Slic3r (Current Ver. 0.9.8), and the setting is: Print Settings -> Infill -> Solid infill threshold area. This was set to 70mm^2 by default. Drop it down to 20mm^2 and the top cube should be printed like any other.

Top surface layer fill Vs perimeters & infill.
See Triffid Hunter's Calibration guide for how to adjust Extruder E-Steps on the fly to calibration top surface infill.

Vertical Walls
Wall building and the effect on print speed Vs the quality of the vertical wall printed.
See the temperature section of the calibration guide. I found that higher print temperatures cause too much oozing of filament on the corners of the cubes. Fine-tuning the temperature as per the guide will allow you to see the effects of extrusion tempeature changes on corners & extrusion quality.

Cooling settings on small layers (such as the top cube layer, as above).
Here you can adjust your small perimeter settings and minimum speed/layer time/fan speed settings. I am still trying to find a good combination here for my printer. My current settings for this are Small Perimeters - 30mm/s, with a speed reduction if the layer time is below 10s, and a min. print speed of 15mm/s.

Base Layer Height:
If your base layer is too low, then that will affect the first few layers of your print. I replaced the default z-axis stop on my printer with one which gave me much finer control over the setting of the z-axis height and calibrating your base layer heigh properly can be the difference between putting a perfect print off the desk, an aborted print, and spending time reworking the model to trim back the squished out base layers.

There are other calibration objects out there that do similar or more things, but I had fun printing multiple pyramids out to check the results of different settings.


猜你喜歡

與描述相符

0

0

1

非常不滿

2

不滿義

3

一般

4

滿意

5

非常滿義

模型下載

該模型為收費(fèi)模型

打包文件內(nèi)包含1個stl子文件

下載需要花費(fèi)您80元
確認(rèn)支付嗎?

400-690-8069

九九视频直播综合网| 双城市| 亚洲视频四季一区| 亚洲免费在线| 亚洲国产精品一区二区三区| 免费成人精品综合视频| 日本AⅤ第一福利| 日韩一区二区欧美一区二区| 开心色五月丁香五月亚洲综合在线| 玖玖久久| 日韩色av色资源| av爱爱| 亚洲日韩第一页| 婷婷色图| 久久三区| 一区二区毛片aaaaaa| 欧美一区二区三区精品久久久久| 亚洲福利在线| 中文字幕日韩精品人妻成人游戏| 大香蕉视频| 亚洲综合在线一区网站| 亚洲无码精品久久久久| 久久狠狠1卡2卡| 色综合久久久综合| 人妻熟妇乱又伦精品视频APP| 欧美日韩七区| 日韩福利电影| 在线精品国产| 狠狠干网站| 99久久精品费精品国产一区二 | china熟女熟妇乱老女人| 禁久久精品乱码| 嫩草嫩草嫩草嫩草嫩草| 国产aaa爽快片在线观看| 亚洲天堂成人在线| 女同互舔| 久久久一本精品97久久精品66| 91av欧美| a天堂在线官方在线| 免费欧美| 欧美多p|