The lever on the tool to the module mount exaggerates the tool head mount lack of rigidity which is why I started there. But the size linear system I chose will work on other axis similar. Access to hardware should be acceptable… if painful
I imagine if I do every axis I will also brace the Z as well
Pug
I should also note 
I will not be doing the 14 x 4 holes in each Drylin rail with the Snapmaker (although hopefully after they are installed it will handle it 
)
You could probably use it to mark everything. Like using it as a centerpoint to get everything marked so you can then safely drill all holes in a drill press.
And just to be sure, I assume you did see this thread? Minimize rocking
It has a lot of possible improvements for the y-axis. (including how I did it myself) maybe there’s something there you can reuse.
I don’t think the SM2 can reach all of the hole coordinates for its own module lol!
And yes, the table is equally as bad for rigidity as anywhere on the machine sadly, the right bracket on a rail will help. How much? dunno (I suspect it will help a lot)
Hiwin rails are a bit overkill and not something I would consider doing to the SM machine… I’d be building my own cnc before using those rails on this one.
Pug
I fully agree. I personally went with a lot cheaper sbr16 rails and that actually works pretty well.
It’s a long thread and has many variations on the mod.
One thing though, when using the cnc there is quite a lot of dust that sticks to the rails and possibly gets in the bearings, So I am thinking of putting some kind of cover or “sock” over it. Maybe some old nylons from why wife or something 
Yes these would be ideal for the Y axis
If you raise the new linear guide off the top surface of the linear module with spacer blocks you (1) increase the size of the torque arm resisting rotation in the Y-Z plane, perhaps doubling it, (2) open up space for a hex wrench to reach mounting bolts, and (3) create clearance for a second bearing block. The mass penalty is well within the capacity of the two drive motors on the Z axis.
I know this is a slightly old thread but it seemed like a good place to post.
I have been throwing around the idea of not using linear rails for the X and Z axis and instead using the groves/channels built into the Snapmaker liner modules similar to how a lot of other printers use V guide wheels in the aluminum extrusions. I already successfully braced the bed rocking issue with some overkill HGR20 rails and wanted to keep the other axis’s cheaper, simpler, and lighter weight. I still have some issues to work out with getting them to fit correctly but here is what I have in CAD so far. Anyone else have luck bracing the X and Z yet?
Another great idea!
I think it would work well if it could be realized in an aluminum housing for Y-axis flexure countermeasures.
Yes for the Y axis (tool head) brace would ideally be made out of aluminum for stiffness and to support the tensioning adjustment bushings that come with the wheels. I have tried a fixed (non-adjustable) 3D printed version and it is very tricky to get perfect. So I have been racking my brain to come up with a design solution that I doesn’t have to be made by a CNC machining service so that it would be easy and cheap for anyone to make. The best idea I have right now is trying to embed a price of aluminum plate into a 3D print but I haven’t quite figured out what that will look like yet.
The bracing you’ve got pictured won’t be that effective. A full explanation takes a bit of explanation.
The sliding mount on the linear module starts off with 6 degrees of freedom. We need a coordinate system (this isn’t the same as the machine as a whole): X along the lead screw, Y across the top face of the rail, Z from bottom to top (rail is lying with its slider up). The lead nut constrains X position (more accurately, it’s a 1D constraint between X position and Y-Z rotation), and the width of the lead nut and mount constrain X-Y and X-Z rotations. Of the other three, only Y position is adequately constrained by the internal rails and the U-wheels. Z position and Y-Z rotation are only weakly constrained; it won’t get out of position too badly, but it’s mostly unconstrained in its “standard” position. The U-wheels exert zero force transversely until they start to bind. If they move the same direction, it changes Z position; if in opposite directions, it changes the Y-Z rotation.
The blue mount on the left of the picture braces only Y position, the one that’s already one of the good ones. The real need is to constrain Z positions on either side of the lead screw. The blue mount on the right of the picture does half of half of the job; it prevents Z position on the Y+ side (linear rail coordinates) from going negative, but three more such wheels would be needed for the other constraints. And even then, it doesn’t work for the transverse linear rail (“machine X”) without seriously restricting travel because of interference with the vertical rails (“machine Z”).
I don’t think there’s a good incremental fix for this on the A350 and kin. There are some non-incremental ones, but I’ve pretty much decided that these, however theoretically interesting, are better applied to a fresh machine. For the transverse axis, that involves “widening” the transverse into a frame table and removing the internal U-wheels entirely. At that point, though, the whole extrusion is also useless. I consider the linear rails a quite ineffective use of mass budget for a moving machine part.
@eh9 Thanks for the feedback and detailed explanation of the machines degrees of freedom! I’m surprised I didn’t realize it myself earlier as it seems so obvious now. My goal is much simpler now as I will focus my effort on only the (“machine Z”) linear rail with the stabilizer in the right of the picture. I was already planning on adding those braces on both the top and bottom of each side but just included a single example in the picture as I did not want to do the extra work to render all 4 of them yet. I am pretty close to a functioning result which I hope will help improve the end result of machining harder materials and increase the speed/depth of cuts.
Finally got a version of the Z-Axis rollers that everyone can easily make and use for cheap! The mod effectively reduced the slop/splaying out of the tool in the Y direction which should increase machined part accuracy and maybe even allow for faster feed rates. I still need to do testing to see how much a difference it actually makes but man it feels so much more rigid now!
Alright I’m back with another design that I think will finally work for bracing the toolhead. Would like to know what you think of this one as I am just printing out the first prototype and could used some feedback. The horizontal bolts are only there to add stiffness to the relatively thin plastic back panel/brace and the vertical bolts are to act as an axle for the roller wheels with a printed spacer between the wheels. The brace is 2 separate parts (a top and bottom) that will be bolted together by the axel bolts and held in place from the force applied by the wheels on the module so no mechanical connection is made to the toolhead or axis carriage bracket
