Offset Layers

So after setting up my Robo and printing out a filament feeder I tried printing out a playing card case, and it will print fine but will seemingly shift whole layers over about 1mm or 2 on the x axis. I tried reprinting it and it started to mess up in the exact same spot as before. I had noticed this on some of my other prints but it would be for only one layer then realign itself. I tried searching but could not find any similar issue, I am currently reprinting it in PLA with much slower speed settings to see if that helps.

Settings for 1st and 2nd try:
Perimeters: 50
Small perimeters: 50
exterior perimeters: 70%
Infill: 50
Solid Infill: 60
Top solid Infill: 50
Support Material: 60
Bridges: 60
Gap Fill:20
Travel: 100
Type: ABS
extruder temp: 235
Bed Temp: 85

Settings for current print:
Perimeters: 20
Small perimeters: 20
exterior perimeters: 50%
Infill: 20
Solid Infill: 25
Top solid Infill: 25
Support Material: 35
Bridges: 35
Gap Fill:15
Travel: 65
Type: PLA
extruder temp: 190

Pic to help explain my issue:

 

How tight is tight? There doesn't appear to be any obvious slop. How far/much should you be able to move it by hand?

Given the fact that the Z axis rod was completely broken free from the drive cam when we got it (see: http://forums.robo3dprinter.com/index.php?threads/gotcha.1080/) I am willing to believe anything, but I want to get it working. Thanks

 

Thanks, I will have to check that when it finishes the latest attempt at (really) slow speed. It still has another 5 hours or so to go. I want to do a direct compare with the speed at normal (as posted) and this (which I would consider as slow as is possible).

From memory though, I would say that sounds about correct for the belt (i.e. how they currently are).

I would also think the stepper pots being mis-set to cause more frequent (as in more often) layer deformities rather than once every 70 layers or so, but I am a beginner at the 3d printing side (just not at electronics). I will need to get some advice on how to even begin to set that.

 

I don't think it's a mechanical issue. Look at how consistent it is. Also look at how it is happening at changes in geometry. I also don't think it's speed related.

I had a similar issue on an ABS part and it actually turned out to be due to the part bowing. I built gussets into the model and it fixed the problem. http://www.thingiverse.com/thing:68094 You can see how that looks like an axis shift, but upon closer inspection, the change below the shift occurred gradually.

Your switch to PLA will probably fix it.

 

Thanks for the data.

The new print we are running in slo-mo is using PLA (we switched to PLA when we slowed it down simply because of the curling/bowing that was occurring on the corner of early layers--bottom right in photo above). I have seen already in our attempts a number of issues with ABS depending on the model shape/size so we thought that might be an issue.

BUT...It (the slow-printing PLA one) also has a lot of the same inconsistencies. NOT as bad as the original, but at approximately the same Z-axis offsets. I will post a side-by-side photo in a few hours when it finishes.

I had high-hopes that the speed adjustment would do it, but no. It did affect it as the offsets are not as gross, but they are still there and at approximately the same spots vertically.

 

OK, print completed.
Pic attached left-to-right: ABS, ABS, PLA (low speed)

As you can see the BLUE (PLA) version has the same aberrations, just not as extreme. I would not call this a good print since it is in no way straight. The X axis is still all over the place.

 

that looks like the enitre bed is shifting or sliding badly is the bed secured well to the rails underneath
can it be secured with tape on the sides or the front along the length or width

 

I am not sure how to adjust how it is connected to the rails, but the glass and the bed itself are well attached (see photos, bot ends have the same clips--4 total) and do not move in anyway with a moderate amount of pressure applied by hand. The calibration also works (head travel seems consistent X/Y and Z as far as I can tell). The original clips provided always (always!!) came lose mid-print, but this method has not budged during a print.It moving back and forth (as well as up/down) shows no discernable movement in the bed (other that what is intended)

 

looking at the pictures those clips are probably not the answer either metal on glass will always slide somewhat glass is to smooth all it takes is slight movement and things can start to slide I am not saying it isn't a lot better but to make it even more solid I would actually use tape around the entire perimeter of the bed.

As far as securing to the rails on yor bed the method is the screws and the small white stand off posts. The head of screw that holds them in place actually is supposed to fit inside that bottom rail and sometimes it is not in it very well but it should be tight all the time if it wavers in anyway the bed will shift. So you need to know that the bolts holding it in place are nice and tight and that the head of the bolts is fitting securely in that lower rail.

 

Those clips are really firmly clamping the glass to the bed... I can't move it by hand so I feel certain that is not moving (unlike the clamps that came with it that almost always moved when printing). Like I said I can't move it by hand (at all). If it is moving during printing it is certainly REALLY consistent about where and when it is moving since those prints are always messed up in the same way (same amount of offset) at the same layers.

Let me get to the screws for the rails and see if those appear to be firm.. given the (apparent lack of) testing that went into the unit before I got it, anything is possible. I mean the Z-axis threaded rod on one side was not even attached (to the coupling at the bottom of the threaded rod) so that could not have been tested.

 

Another weird (to me) thing is that at the really super-slow speed, even more X-axis screw-ups happened, just not to the same degree. In the faster case 5 sections had the X-axis mess up by 10 mm or so (each section maybe 10-20 layers 'tall') in the slow case a lot more offsets happened, but only a few mm each in the X axis. The faster print had several large sections that were perfectly straight.

SO anyway I am not buying that speed is the core problem here at this point.

 

Speed only had to do with the ringing aspect it sets up a kind of harmonic that causes the echo I have read about it on other forums as well and that was their conclusion as well

 

I can now confirm that this is true. I checked the screws and the spacers to those rails...I can't tighten them further. They are plenty tight. So the connection from the bed/platter to the rails is not moving and the glass plate and heating bed is not moving.

The belts seem tight enough as well. I am not sure how to judge that. You can force them to touch by squeezing them firmly, but they are not loose in any sense. You can feel the tension when you squeeze them together. The pulley in no way spins free or seems loose.

 

the last thing I cna think of is perhaps the gear on the motor that controls the y movement is slipping. Checking that though will if I remember right will entail removing the entire bed to get to the gear to see how well it is on the flat spot of the motor and then maybe even removing the motor if you need to make the flat area more pronounced so the gear would not slide or slip. I had to do the exact thing a few months back.
Since it is happening as Y increase or decreases that could be the case.

 

ouch!

OK, that will take a bit of research.

 

yeah mine was not good by any stretch I laid it on the table and set my dremel to the right level on blocks then moved it back and forth on the motor spindle until it was perfectly flat previously any gravel roads was flatter the the spot on the spindle but now it is back nice and flat and working great

 

Actually, the pictures my be a bit confusing. They print vertically like attached pic. I laid them down to better show the offsets on the X axis.

So as the Z goes up the X deflections are happening. Y seems pretty stable.
At this point I am clueless.

 

In that attached photo (and in the model) the sides of the card box should be flat (straight up and down).
Instead it looks like a funky 'S' curve