Odd circle test results (Page 14 of 24)

Reviewing your work, it appears you didn't make any modifications to the side bearing/bushings.  Is that right?  Hard to see.

Note: In the first pic "image.jpg", your x-axis belt is twisted on the left side.  It may not matter, but worth mentioning.

I see you're still operating without that rear enclosure fan in the back right and you removed the left metal bracket for the filament fan on the left side.  Did neither of those fan prove useful?  I may relocate those fans as you have to the extruder and y-axis locations.

Glass is nifty as it has to be perfectly flat by design in mirrors to prevent deflection.  I've definitely found that my bed is not level (ever slightly bowed on the edges).

I'm also going to replace the black top with an acrylic clear sheet so I can see through the top. 

I've become SD4 mod happy so I must slow my role.

the fan at rear on SD4 etc is for fumes filtering/exhaust on ABS etc. not sure they where always included, rather an addon

Do you have a link to what you're talking about?  The Z-stop switch is pretty basic and functional, backlash testing on the vertical for mine has shown 2 significant decimals of accuracy on mine, good enough.  Did you have problems?

The e3d extruder height effectively lowers the vertical print space.  We all make sacrifices but print space isn't an area of sacrifice to me.

Not sure if anyone else covered this, but replacing the stock Solidoodle Z-rod is a very good idea. For a little extra you can buy a much higher quality rod.

I have no idea.

Claghorn you need to turn all of these little web pages into a SoliBlog.

I think I found everything by searching amazon, google shopping, and ebay:

http://www.amazon.com/gp/product/B00KHT … &psc=1 was the shaft coupler I got, but the M3 rod is so flexible by itself, this one might be better: http://www.amazon.com/dp/B00OK6ECV8/ref … 2HXJF6H4P4

I couldn't find any brass 5/16 bolts locally (and didn't want to try and machine steel ones), so boltdepot.com is where I got the brass bolts and I ordered various M3 and M5 bolts and nuts at the same time just to have them handy.

I got the M3 threaded rod from Ametric on ebay and cut off as much as I needed from a segment that wasn't bent too badly during shipping (not sure how you could ship something that long and thin without it getting bent or having the packaging cost more than the rod, in any case it was only bent near one end :-).

I thought of this test, not sure what it really proves though.

Okay, so since when printing infills the motor doesn't slow down and speed up, instead changes direction almost instantly, I thought if I print circles made up of only infills, that will prove that the issue is caused by backlash on the belts, nothing to do with bending rods, bearings instead of bushings, VREFs, etc.
Am i right? Or not quite?

This is how I did it

My glass was messy, so it didn't print very smooth, but you get the idea.

hi-res version: http://i.imgur.com/eJHlF3P.jpg

superimposed on perfect circles:

The circles STL is attached.

Did I prove anything with this test?

Would be much easier to test if I could increase the "shell resolution" (more lines, less thick), but I don't see anything for that.

So after assessing my lawsy carriage installation while I waited for a 625zz bearing to upgrade to the mk5, I found a ton of slop in various areas and also determined I was running my belts a bit too tight along Y. Here is my improvement after proper installation of the lawsy carriages. These are perfectly acceptable to me and I can only imagine bearings on the back rod will improve the situation further

Conclusions so far: the lawsy carriages are likely to improve your situation, and you should install them anyway because linear bearings are the proper way to use a smooth rod. The aluminum extruder block on the SD4 is also junk and can wiggle quite a bit. Get a 625zz radial bearing and upgrade to the mk5. You'll appreciate being able to adjust the tension anyway. Run your Y belts only as tight as needed to drive the carriage and no tighter. Loosening these belts actually decreased my backlash. I assume this is from drag in the bushings so they are the next thing to go.

I run that belt tight so if it is inducing backlash it is from drag in the bushings. I don't experience the same slop pirvan does, I don't think; my drive belt looks like it runs very smoothly and I can't find any play in it. I'm going to put bearings on that rod next, and try pirvan's solution after to see if it helps.

I'm actually more annoyed with my corners right now than I am with  my circles still a tad sloppy on Y from lash. Is there an easy way to measure my backlash? I've seen people do it on the forums before. To measure it along Y I could move in small increments and see when the Y rod starts moving--but how small of an increment? Less than 0.1mm?

Actually, the circle test seems to be a pretty good backlash measurement. Printing a 1 wall thick cylinder starts moving the head one direction to go through one side, then moves it the other way on the way back. If the ends don't join up perfectly, I'm pretty sure that is backlash (or something wrong that ought to have a name, anyway :-).

Now If only there was some way to tell the slicer exactly where you wanted it to start the circle wall, you could check both ways (for me, slic3r always seems to start in the very back, so I'm seeing Y backlash if it doesn't join up).