The max X-Y accuracy stock is +-.011mm (88 steps per mm).

Yep.  This hack improved the surface finish of my prints.  Maybe I had a particularly wobbly bed, but the lateral shaking was noticeable.

Having measured it, I am now thinking of using a turnbuckle across the front to eliminate the variance (I get into the machine through the right side).  Heywhynot.

I would be interested to see it  measured though.

Since I started this ... :

On my SD3: Standing in front of the machine looking into it towards the z rods in the back, the front vertical rails are tilted to the right .0135" over 100mm of Z travel.  The verticals are parallel within 0.003" over that distance.  See attached pics.

2n2r5 wrote:

The stabilizers seemed to be helping on my test print last night until it got to ~125mm on the Z. Then it started having little layer offsets where the rollers must have hit some debris on the side rails. The offset was slight but very noticeable on the smooth walls of the print. I will try to post a picture at some point. Just for good measure I ran the print twice and got the same result twice. Luckily, to take the stabilizers out of service, I just undo one screw on the spring and move the plungers inwards. 

Results so far have been mixed. If someone has some extra time and wants to play with this then have fun. If nothing else, I have a good base for a printable roller limit switch.

2n2r5,

Nice idea!  I think perhaps why you are getting mixed results may be due to the fact that these still create a "floating" system being that both ends are spring-loaded.  Perhaps if you immobilized one side to ride hard on one rail, while the other side remained spring-loaded to compensate and always bias the bed to the other side.  Of course, you would need to find out which rail is more dead parallel to the z-axis, but it's worth a shot.

Yeah, that is the conclusion that I ended up with too. I have equal forces pushing on each side of the bed working to cancel each other out. The friction in the plungers might actually be helping a bit. Probably not enough to make it noticeable. I don't have any real issues so even if this did help, I would have a hard time seeing any evidence of it.  I think a shock absorber would work better than a spring but really, I only messed with this cause I was bored. I don't think that this will help enough to make it worth the trouble. The best thing would be a new bed all together that has supports wider apart and at all 4 corners. That falls under the realm of might as well build a new printer.

I think the important thing isn't if the side rail is straight so much as it is straight to the eye. I am wanting only to get an esthetic print. When I look at the side rails, I would be delighted to get something as straight as them. If you have a bent angle support then putting some flat stock on on the inside of each rail should cure that. I really don't think the 3/8" guide rods are significantly out of parallel with the sides of the printers. Or at least it didn't show up in my test prints. I would be interested to see it  measured though.

If you attach a dial indicator on both sides of your bed and sweep the steel frame from top to bottom you would find they are probably not perfectly parallel with the Z axis rails. Those rollers are going to follow the angle of the frame and not the rails, which introduces a slight push factor in one direction or another via the force of the springs. Might be doing more harm than it prevents.

This could be worked around by installing a secondary steel rail from top to bottom which is adjustable with screws so it can be dialed in via the indicator.  Naturally it would need to be of proper rigidity not to flex in the middle.

The stabilizers seemed to be helping on my test print last night until it got to ~125mm on the Z. Then it started having little layer offsets where the rollers must have hit some debris on the side rails. The offset was slight but very noticeable on the smooth walls of the print. I will try to post a picture at some point. Just for good measure I ran the print twice and got the same result twice. Luckily, to take the stabilizers out of service, I just undo one screw on the spring and move the plungers inwards. 

Results so far have been mixed. If someone has some extra time and wants to play with this then have fun. If nothing else, I have a good base for a printable roller limit switch.

2n2r5,

Nice idea!  I think perhaps why you are getting mixed results may be due to the fact that these still create a "floating" system being that both ends are spring-loaded.  Perhaps if you immobilized one side to ride hard on one rail, while the other side remained spring-loaded to compensate and always bias the bed to the other side.  Of course, you would need to find out which rail is more dead parallel to the z-axis, but it's worth a shot.

... little layer offsets where the rollers must have hit some debris on the side rails.

I have seen that as well.

You and I are among the folks trying to run these machines at accuracies beyond their design intention (you are farther along that curve than I).  While your vases require aesthetic purity, a single layer mismatch of a few thou just annoys me more than it affects the integrity of my part (most of my prints are tools for some specific purpose).  That being said, I now routinely wipe down the inside of the front vertical sheet metal - the "rails" for the bed stabilizer - to avoid becoming annoyed (about that anyway).  I also hit the inside of the sheet metal with scotch bright just cuz.

The risk of a transient layer mismatch is inherent in this mod.  Oh well.

Results so far have been mixed. If someone has some extra time and wants to play with this then have fun. If nothing else, I have a good base for a printable roller limit switch.

Hows the performance of the stabilizer? Notice much difference?

I got a variety pack of springs a while back and used a couple springs from that.
http://amzn.com/B000K7M36W

This configuration uses 625zz bearings. I have included the sketchup file if you wanted to use this with other bearings.

The arrow that I have on my filament drive gear is also added below.