This part of the site is still currently under construction. For now please carry on using the original site at - caggius.wordpress.com
I planned (and pre-bought) a lot of upgrades after reading plenty of articles on the web. Some of which were good, others provide to be of little or no benefit and others are just a total waste of money.
In 18 months I have only implemented a few of them as the printer seems to function quite well out of the box and up until now I have not felt the need to change it too much.
So the changes to the original A8 that I have found worthwhile are:
Z Frame Stiffener
Definitely needed as the ANet shakes a lot from side to side. So I designed and printed my own. Why? Because I wanted to learn how to design, slice and print. There are plenty of better ones on Thingiverse, but I learnt a lot more by doing it myself.
Y Axis Stiffener
After a lot of research into printed stiffeners, and several failed designs using extra threaded 6mm and 8mm metal rods, I eventually bought a “LACK” table from IKEA as a permanent base and used a pair of long wooden blocks 25mm wide x 45mm high to run just to the right of the Y belt front to middle (222m) and middle to back (136mm) frames. I screwed these to the Lack tabletop and tightened the Y frames on to them by adjusting the two Y axis 8mm threaded rods. Surprisingly this simple solution seems to work really well.
Y Axis Belt clamp.
I downloaded several designs but they did not seem to fit the bracket or belts very well. In the end I created my own, based off the one linked below, then I had to refine it for the slightly thinner new belts.
https://www.thingiverse.com/thing:2347849
Y Axis Carriage Plate
I have replaced the 3 part bolt together frame with a single piece frame off eBay. Apart from looking more professional I am not sure that it has made any difference to the print quality.
Y Axis Belt Tensioner
Again a must have to keep the belt properly tensioned. I printed this one from Thingiverse, then I had to modify it to fit the slightly smaller holes in my frame (or it may just have been a slight over extrusion on my printer).
https://www.thingiverse.com/thing:1959208
Y Axis Belt
If you add a tensioner then you will need to buy longer belts. So I also switched to the more flexible glass fibre reinforced ones. This flexibility helps reduce backlash, but the thinner belts required reworking the Y belt clamp that I had already created!
Y Axis Bearings
The ball bearing ones provided can be a) very noisy and b) chew their way through the cheap chrome steel rods. So I switched to IGUS DRYLIN bearings. But be warned getting hold of genuine ones may prove problematic as there are a lot of fakes out there.
See my X Carriage Modification Section below here.
X Axis Belt clamp.
The two belt loops pushed over 3mm bolts works OK until you really tension the belts up and then they bend. I downloaded and printed several designs but they did not seem to fit the bracket or belts very well. In the end I created my own, a mix of ideas from several on Thingiverse and again I had to refine it for the new thinner belts.
X Axis Belt Tensioner
Again a must have, to keep the belt tensioned, I found this one on Thingiverse and it worked well enough first time.
https://www.thingiverse.com/thing:3105157
X Axis Belt
As per the Y Axis, new glass fibre belts…
X Axis Bearings
As per the Y Axis, IGUS DRYLINs…
X Axis Carriage Plate.
New! Jun 2020 – The fake IGUS bearings I bought from China wore out very quickly and generated a lot of backlash. So when I replaced the bearings for a second time (from a trusted UK supplier) I also bought some sheet Aluminium and created a wider X Carriage plate. This enabled me to use the 34mm gap between the hot bed and the X Min limit switch to position the top bearings further apart and reduce the rotation around the Y axis. See my X Carriage Modification Section below here.
I modified the Filament reel holder to add some some 8x22mm ball bearings using a conical holder that I created so that the reel does not bind. I reused the bearings that I had purchased, but not used, for the Z Screw Wobble.
External MOSFETs
Oct 2019 – I finally got around to adding the external MOSFET for the Bed Heater.
Apr 2020 – the external MOSFET failed! So I have I reverted to the SKR 1.3’s onboard MOSFET.
Cooling fan duct.
Plenty of different ducts out there but for what I have been printing the stock one seems good enough and with Slic3r the fan rarely switches on anyway !
New! May 2020 I have found it necessary to screw the Duct to the fan case to stop it slowly slipping out of the fan case and hitting the print.
New! May 2020 – After burning my fingers changing the filament I have made three further changes to the hot end assembly:
a) I fitted a heater block silicon boot and ran a fresh PID calibration cycle.
b) I removed the extruder motor heatsink and created a new mount for the motor fan 1 cm away from the motor itself. This allows a full view of the extruder gears and the filament throat entry whilst changing the filament.
c) I have fitted a PWM Motor controller to slow down the motor fan and reduce the noise levels further.
Note: Testing has shown that these changes have had minimal effect on the motors running temperature with my setup (low motor current, and UK ambient temperatures). If you make these changes you should validate this on your machine in your environment.
The following add on’s were bought but have not yet been used. They are pretty much on hold pending a catastrophic failure:
SKR 1.4 Turbo Board and TMC 2209s .
For now I have stuck to the SKR 1.3 and TMC 2208s taht I installed in 2020.
I played with the 2209s and soft endstops but after driving the X&Y carriages into the frame a few times I decide to stick with the switches for now.
BMG Style Extruder Assembly with J Head and heater
For now I have stuck to the OEM setup. But the addition of my X Carriage Plate modification means this is so much easier to trial that I may give it a go.
Auto Self Levelling Inductive sensor.
To be honest I have found the ANet bed setup to be mechanically very stable, and my own print bed is pretty much flat (not overly distorted). So much so that I find re-levelling is rarely necessary and when it has to be done is actually very quick to do. If yours is slightly out then use a raft and/or print a 0.3mm first layer
I always monitor the setup whilst the skirt is printing and decide at that point if I need to stop and adjust it (if one side of the printed skirt is thicker than the other). When I compare this effort to the number of people struggling to get BLTouch working and in my mind this upgrade becomes very low priority.
The following add on’s were bought and will never be used:
8mm Lock Nuts.
Initially my Y frame nuts were shaking themselves loose. Once the X, Y and Z axis stiffeners were installed the Y frame bolts no longer shook themselves loose. So once the frame is stiffened they are not really needed.
Hot End and Bed Heater MOSFETs – in my experience the OEM MOSFETs on the ANet V1.5 Board do not get hot enough to worry me and are operating well within their design limits! The same can also be said of the SKR 1.3’s MOSFETs.
Stepper Motor Diodes – A total waste of money – they are only needed for the DRV8825 stepper driver chip, the ANet boards use A4988’s, and my SKRs have 2208’s and 2209’s…..
Z wobble bearings. Another total waste of time and money as they can actually make printing worse not better.
The Z Carriage is designed to slide freely on the steel shafts. As the backlash collar has only 2 grub screws (set at 90 degrees to each other) for each shaft the lead screws will always be attached to the motor shaft eccentrically (off centre). So by design the top of the lead screw will wobble, more so at the start of a print. If you do the Maths’ this wobble really has minimal effect on the Z axis accuracy. However if you then fix the lead screw at the top you are forcing the lead screw and steel shaft to bow around the carriage T nut, increasing running resistance. To counter this you will need higher Z motor currents to overcome the additional resistance and if it is severe this can introduce skipped steps which will push the X carriage out of level every time you print – not a good idea.
ANet SoftwareCompleted as part of the December 2019 Major Upgrade.
I moved to Marlin 2.0 Gold and later to 2.0.5Two tips here a) never install a new update until someone else has done the hard work of debugging it for you b) if your machine is working then there really is no need to go upgrading to every Marlin Level released - this will take all of your time.
Slicer
A very down level version of CURA is provided with the printer, which to be frank I could not get to produce prints. However Slic3r worked out of the box so I have stuck with that, until I moved to the Prusa in-house variant of Slic3r in 2020.However I occasionally find the odd problem with generating limited supports and so I have recently created a working configuration for CURA 4.4.1.
https://github.com/bcarteruk/ANet_A8_BTT_Marlin
Control from a Pi or PC – OctoPrint, Pronterface or similar.
I did not find this necessary to set up the A8 as shipped. To be honest I find the use of an SDCard to move gcode quite simple and given it takes just a few seconds and prints take hours why bother to spend ages getting the connectivity to work and tie up your laptop when you could be designing something else whilst the printer is working….
Having said that it was absolutely essential to have this function to debug the move to Marlin as there was too much scope to screw things up with so many things changing at once in the Marlin Configuration files. So PrintRun aka Pronterface was installed on both my Macbook and Windows laptops.
Jun 2020 – I recently found that the backlash on my X axis had increased significantly and it was not a loose belt. The IGUS bearings on the X axis had worn badly and the carriage was loose on the steel shafts causing the entire carriage to rotate about the Y axis.
I suspect that the bearings were cheap fakes and not originals. So I compared the new and old bearings. If you roll the bearings over, then on the new bearing the writing on one side is inverted compared to the other; on the old they are the same way up. So the two bearings came from a different mould and either the new or old are fakes !
That made me think, how can I reduce any Y roll caused by bearing wear. The simple answer is to place the top two bearings further apart. The ANet A8 has 30 to 35mm space to the left of the carriage, so it is possible to move the left hand bearing 30mm to the left and then in Marlin change the X axis endstop offset.
So that is what I have done – I used a 10cm x 8cm x 2mm Aluminium plate to mount the 3 bearings and the X Belt clamp. Then I mounted the existing X carriage plate to that.
The advantages/disadvantages are:
Plus: Wider bearings, reduced backlash
Plus: Easier to take the hot end off. e.g. to clear blockages, without dismantling the entire carriage, reducing recalibration on reassembly.
Minus: Carriage is a few grams heavier
My first attempt at the new plate has the carriage mounted in the middle. However it really needs to be mounted to the right hand side to keep access to the full 220mm build width. But I can run that way as a trial as I do not need the full width for my flight instrument prints.