printed objects are always tilt

[adenovirux 0]

Hi, all:

I download some samples from thinkgiverse, but it seems all the sample I printed are tilt in certain direction. The bed was calibrated as the instruction in "run bed leveling wizard". Anybody has an idea on what happens?

Photo about the printed objects are here:

http://www.flickr.com/photos/9456942@N08/9430782688/

Also you can see that there is a "base" in the left object, but it is because the motor shift an unexpected distance druing the print. This problem happens from time to time. Is there anyone also has the same phenomenum?

Thansk a lot in advance!

Carol

 

[illuminarti 18]

The gentle slope in the sides of the prints is probably due to the pulleys slipping on the shafts. You need to tighten the screws really hard, and also use a permanent marker to put a mark on the pulley and shaft, so you can see if the pulley slips at all. Check all the pulleys, including the ones on the motors.

The larger shifts might be caused by the same thing, but may also be due to large fast moves failing, and the motor skipping steps. That's probably due to too much friction in the axes. With the machine powered off, try moving the head by hand, by pressing both sliding blocks on the axis with your finger tips. The head should move fairly easily, and with the same amount of pressure in both directions. If one axis is harder to move, look for anything that might cause friction, like the head rods not being level and at right angles to one another. Make sure the short belts aren't rubbing on the frame. Also, check that the end caps aren't pressing two tightly on the ends of the axes. You might also add a little sewing machine oil to the axes to help everything slide smoother.

 

[gr5 2,005]

What you are reporting (tilted parts) is very common for newly assembled machines and usually simple to fix. Illuminarti just gave you a list of the most common causes. There is a 98% chance it is one of those. I just want to elaborate on one:

 

Make sure the short belts aren't rubbing on the frame.

 

This is evident because everytime the x or y axis shifts direction (the stepper switches direction) the belt twists a different way (back and forth - back and forth). If you see this then it is touching the frame. Usually some washers as spacers fix that particular problem.

I selected this particular issue as it is probably more likely than other issues for this particular gradual tilt like this. Second most likely is just set screws not tight enough.

 

[adenovirux 0]

Thanks, illuminarti & gr5!

The gentle tilt (gradual shift) is rescured by the re-adjustment of the belt tightness, to solve the unequalness of front/back belt tightness. I have difficulty to distinguish the pith properly so I used a tuning app to help!

And you're right, my shorter belt is about to rubbing the frame, also some of the pulleys are too loose so I can even move them along the shafts by finger. Also the two pulleys on the left back side are too close, so they bother each other when moving. I had make markers and will print a lot to see if the sudden shift happens again.

 

[illuminarti 18]

The pulleys need to be absolutely tight. The gradual shift isn't caused by differences in belt tension - it's caused by the head not returning to the same physical place for the same coordinates on each layer. And that can only happen if the motor loses track of its position, skipping steps, or the pulleys physically shift. Adjusting the belt tension may have temporarily reduced the stress on the motors, or perhaps you retightened the pulleys better after. But belt tension isn't the root cause of shifting layers. Belt tension issues can affect exact positional accuracy within a layer, but cant cause layers to shift.

Having the two pulleys in the back left touching one another isn't a problem (assuming you mean the ones on the same x-axis rod). They're supposed to move in lock step with one another as the rod rotates. There should be no way for them to move independently of one another and 'interfere'. Indeed, both belt channels could be cut into a single pulley assembly.

Tighten all the pulleys. They must not move sideways, nor rotate independently of the shafts, or you will never get good prints.

 

[adenovirux 0]

I make all the pulleys very tight to the shafts, and all the markers show there are no, at least significiant, movements. The printed results also go very well, tilts are corrected completely.

However, the large shift still happens sometimes. When it comes, there is a sound every 0.5-1 sec or so. Does it means the motor skip steps? I have lubricated the rods, so the head move smoothly. The only thing remains uncheck is the end caps. In fact I don't really know to which degree is too tight, but I think it should be fine since the axis can always rotate without too much difficulty.

Thanks again illuminarti!

 

[gr5 2,005]

 

• [*]

[*]Which axis is shifting suddenly (large shift)? X or Y? Please give us more information. It's a little frustrating to have to say if ( a ) then do this, if ( b ) then do this, if ( c ) then do this when sometimes we have already asked the question. You only be trying to fix the axis that shifts at this point.

[*]Did you get all the pulleys? Each axis has 6 pulleys. SIX. Count them - perhaps you missed the one on the motor? It's usually the 2 pulleys that people forget about that are most likely to shift/slip.

[*]"my shorter belt is about to rubbing the frame". Did you fix this? It's very important to fix this. Not only does it cause the shift, it also will wear away at your belt. Make sure the pulley on the motor is as close to the motor as possible without touching - about 1/2 mm. Also add washers under the plastic standoffs.

[*]There is also a known slicer bug (not sure if it is in the latest Cura or the old Cura slicer) where the slicer tries to move 4 axes at once (normally Z axis moves without the other 3 moving). There is a bug where when this happens the acceleration and velocity are too fast for the stepper to actually move. If the "large shift" always occurs at the same spot of the same model it might be helpful to have illuminarti look at that spot in the gcode.

[*]You could try to lower your acceleration. By default, the acceleration in Marlin is 5000 mm/sec/sec. You could try lowering to 3000 mm/sec/sec. You should know that when you turn the UM off it will lose this setting unless you "save to eeprom". You can adjust this value using the ulticontroller or with Pronterface (free software) combined with gcodes. It would be bad if you had to fix this problem by lowering acceleration. Much better to fix the hardware.