If you've been following along in the Improving Print Quality thread (https://www.forum.makertech3d.com/forum/3d-printing/improving-print-quality) then you know that the guide system included for the AXIS is sub-par, at best. We've had some very good attempts to address the issues with this system by @Gareth Owen . His additions and replacements are a great place to start with addressing issues with the guide system. However, I believe that if you really want to address issues with how the printer snags or wobbles along an axis then the best way to fix it is through a different mechanism.
My first attempt to address the guide system was to replace the guides with bearings fitted with cups to lower the friction between the moving body and the guide. This turned out to be fruitless as I was unable to design a setup that wouldn't interfere with other items on the printer.
What I settled on was cheap linear rails found on Amazon. I used two 250mm and one 300mm MGN12 linear rails. These rails are around $20/each. A set with a rail for each axis is about $55 as of the time this post. While this isn't an expensive proposition it does take the total invested cost in the printer up a significant amount. If you got all the upgrades when you got your AXIS and then add this on along with the metal coupler you're beyond the price of some budget entry level printers. I look at the cost of the printer and the upgrades I selected as more of a sunk cost at this point. Yes, I could have put that money towards a different printer, but hindsight is 20/20 and I can't go back in time to tell myself to do anything different. So in the end let's make this printer print better!
X-Axis Linear Rail
The first axis I wanted to address was the x-axis. It seemed like it was going to be the easiest. I ordered a 300mm rail. The first idea was to attach the rail to the base and the slider would attach to the platform. I decided against this as I felt that the length of the rail would stabilize the platform better. I'm not sure if it makes a difference. Ideally I think two rails spaced out on either side of the platform would make the assembly rock solid. The issue with this decision would be addressing the belt system for driving the platform. I thought that this modification would be too much.
Using the supplied files on MakerTech's GitHub for the AXIS printer I created sketches of where to place the slide on the base and where to put the rail on the platform. Designing and thinking this through was much easier than the actual execution. Looking at the hole layout on the base I selected the best place I thought for the slider. In actuality I was lucky I picked the spot I did. Guides for the belt system was only a cm or perhaps a little more from where it would mount. On the platform I centered the rail in both x and y directions. I had to revisit this and move the rail further toward the left side, away from the x-axis limit switch, to address travel issues.
Too low. The platform interferes with the pulleys.
This shows the rail centered on the platform. I ended up having to change this to move the rail more towards the right side (from this photo's viewpoint).
The slider on the MGN12 rails when attached to the rail sits low enough that the guide system for the belts has interference with the platform. A spacer is needed to raise the slider so that the whole system is raised. I guessed and created a spacer block 10mm high. This provided enough clearance but is perhaps a little higher than stock.
Z-Axis Linear Rail
After looking at the y-axis and z-axis I figured that it would be easier to tackle the z-axis second. It should also have a large affect on the quality of prints. I also figured it would be pretty simple as all it would require is to mount the rail to the Z-Axis Pillar (A) and the slider to the Y-Axis Gantry.
I went back to Fusion 360 with the models for the printer provided by MakerTech. I sketched out where I wanted to place the slider on the gantry. It was going to be simple, all it had to do was be center-aligned to the lead screw and placed vertically so that the mounting holes would be accessible. The rail should also be easy. I didn't need make a sketch for it. I just looked at how low the gantry would travel on the z-axis and then centered it on the pillar.
First appearances made me thing this was going to be great!
I thought that this solution was going to be simple and possibly the best upgrade when I was partially through assembly. I only wish I had "assembled" all the objects in Fusion before I got this far. I ended up with issues pretty quickly.
The belt, which normally travels between the gantry and the pillar, now had to get past the slider.
I was quite disappointed when I saw this. Some quick adjustments later and I came up with a different belt path.
The pulley on the stepper is just on there.
Add in a longer bolt with some spacers for the pulley, another longer set of bolts for the guides, zip tie the belt on, and we're ready to go!
At this point the y-axis appears ready to go. The machine is turned on and I home the axis. Everything works great, except for the z-axis! The distance between the slider and gantry is smaller than the original design this caused the lead screw to be pulled towards the pillar. As the gantry is lowered the lead screw cannot bend enough and it binds.
Doh!
To fix this issue I went back to Fusion and decided to take all the parts for the printer and assemble them in the correct orientation and at the correct location.
This showed me the difference I needed to fill with a spacer.
So, with this spacer printed I was able to get the z-axis moving without hesitation or wobble over the entire range of motion! Win!
Y-Axis Linear Rail
The final axis to tackle is the y-axis. This axis is turning out to be the most difficult. Finding a spot to place the rails so that the everything is in an acceptable alignment to the pulley and stepper is a bit of work.
Looking at the pictures above, you can see that the belt path is now different. The design will need to take this into account.
Current adapter for the tool carriage to the linear rail slider. The protrusions on the rear are for hex nuts to be inset and bolts threaded into for the attachment of the belts.
Further Improvements
I'm not a mechanical engineer. I can conceptualize things pretty easily, but getting that "down on paper" is an issue for me. This is especially true when working with CAD tools. I'm a software engineer by trade so for me this is the first time I've used these tools. That said, I solve problems all day (in code) and have to often come up with creative solutions. Sometimes they're not the best solution and I ended up meeting the "make it work" requirement. The same is true with my linear rail solution. I'll be posting all the work I've done in Thingiverse and providing links here. I'll do my best to provide any documentation. I hope others will read this, see my mistakes, highlight them, and we can work to address them. As I said in my opening, I have this printer and it isn't going anywhere. I might as well do my best to make it work. I believe that this solution is a cheap but highly effective way of improving the quality of what the printer can provide.
That said here are some things I think should be investigated and improved:
Address better x-axis mounting
I've offset the x,y plane towards the LCD. Come up with the best slider and rail position either by having the rail on the base and the slider on the platform are options as well as looking at better positions.
Better belt paths for x-axis
Better belt path for the y-axis.
There nuts on the back of the Z-Axis Pillar (A) are just enough out to hit the belt as it's traveling up and down the z-axis. This hasn't been an issue during smaller prints and is normally only noticeable when homing.
Re-evaluation of the tool carriage to linear rail slide adapter.
EDIT:
I realized I did not link to the linear rails that I'm using. I purchased them both from Amazon:
https://www.amazon.com/gp/product/B07GBNTJW3/ for the y and z axis
https://www.amazon.com/gp/product/B07Y2SNX4R/ for the x-axis.
If I was doing this again, I'd go for the 300mm rail for the z-axis and a 350mm rail for the x-axis. This would give more margin on how far the slides were able to move before exposing the internal bearings.
Yeah, the X-axis rails are T-Track. They're meant for woodworking tables and the like ... my design is a simple hack with what I happened to have on hand. The Z-axis is the same rail you have ... the 12mm x 350mm rail with 2x MGN12H carriage blocks. Here's another pic of the Z
(the giant pipe nipple is just a 2lbs weight to keep the thing upright when the bed is at its extents ... it also seems to help steady the z-axis a bit).
I borrowed your idea and successfully upgraded my X and Z axises. I took a slightly different approach on the Z-axis. I used a 2 carriage block system so the belt could go between them. My X-axis is made of a T-Track I had laying around and some T-bolt posts I managed to print and polish smooth.
I have models and drilling templates if anyone is interested. I made a few other mods along the way. I should probably put it all up in my own thread. But for now, here are some images:
There's two spacer pieces that attach between my two mount pieces and the carriages (the 2 aluminum pieces partially hidden by the gantry). My rail is mounted just like yours.
@Gareth Owen yeah sorry bad initial explanation.
The plate would only be helpful when mounting the linear rail @Michael G is using on the outside of the Z pillar.
Now that i have the dimensions and step files for the Zpillar itself. I can easily designing a new single component that could be used as a new linear guide. the only issue would be the accuracy and tolerances of the print itself.
Gareth if you want me to put something together im happy to do it. im currently just doing my first prints and trying to calibrate the printer as i have only just been able to finally extened the thermo wires.
Sorry i didnt explain my idea very well. The bracket i was suggesting in the image would mount onto the original linear guides that Makertech provided for their original build.
Sorry for the crude picture. i havent had time to put together full rending of the model in fusion 360.