Shifting in "Two story spec house" thingiverse model

[gadgetfreak 7]

Hi guys,

I'm trying to print:

http://www.thingiverse.com/thing:26244

and so far only the first floor (stl file "FIXED_Spec_House_1st_Floor.stl").

With bad results :(

I have an almost two-year-old UM (8mm rods) upgraded to v2 hotend / v3 bolt. A little bit slack in long belts (should be ok I think), no slack in short belts.

I've sliced it in both

Cura 12.12A / 0.2mm layer height / 50mm/s / Temp 210 / retraction (I forgot the clip on that print)

and

Cura 13.03 / 0.15mm layer height / 100mms /s Temp 210 / retraction

but with about the similar result.

Most parts of the house and furniture gets shifted in both X and Y direction, more so in the window parts or pillars.

Can anyone try to print it and see if it comes out better for you?

In this album I have pictures of the print and also of smaller calibration objects printed with the same plastic just after the second print. It's a 20mm test cube and "concentric circles" They look fine.

 

I'm guessing the house is pretty hard since it's big and because of the small pillars. Could it be done better with another slicer?

It would be nice to see someone else try.

Thanks.

Daniel

 

[gr5 1,998]

Well if the belts seem even slightly loose to you maybe you should tighten and loose belts would definitely cause what you call "shifting":

http://www.thingiverse.com/thing:24336

http://www.thingiverse.com/thing:10082

http://www.thingiverse.com/thing:16449

Or read the ultimaker assembly instructions - somewhere in there is a great video showing how tight the belts should be and how to tighten them without needing any parts (there's a screw you can adjust).

Also tighten all the X,Y pulley screws (10 of them?). Make them verry tight - you-should-be-nervous-you-might-break-something tight, but not so tight that you break something, lol.

Another thing to try is printing even slower. I dont' think it's the slicer.

 

[printedsolid 3]

This doesn't address your question at all, but that is a really cool model! I think I'm going to modify it so the door can be opened and give it to my daughter as a home for her family of Ultimaker robots : )

 

[gadgetfreak 7]

gr5, thanks for your comments and links.

I already have belt tensioners (springs from clothes pegs) on there but it seems that the belts have stretched since this was done (over a year ago).

I cannot fit more spring and at the same time have it print over the whole area of the bed easily.

I have a set of replacements belts but have not been wanting to disassemble the whole X-Y gantry and re-calibrate it, but if that's what it takes I guess I need to do it.

But Daid wrote that the tightness of the long belts are not as important as one might think. I did have other shifting problems and oval circle problems before tightening up the short belts.

But as can be seen inte the gallery of pictures, the concentric circles print and 20mm box show no shifting. But those objects might be "easy" in comparison to wide-area objects?

Perhaps this particular model where there are so many "pillars" and short segments are really putting a the long belts to the test?

I did re-tighten the pulley screws prior to the print so they should not be the source or the error.

One problem is that is takes such a long time before the error is shown.

Perhaps I can find a model with these type of pillars (or I design my own) as a test object.

If anyone else would care do to a test print I would be happy

PrintedSolid: It sure is a cool model...I just want to be able to do a nice print of it!

/Daniel

 

[gr5 1,998]

The thing that makes this print the most difficult is just all the skips: Where the extruder is expected to stop down to zero and then instantly go back up to extruding again.

What happens is you often get stringing where it is extruding when it isn't supposed to be (the extruder doesn't stop fast enough) or you get under extrusion during printing the skin, or you get over extrusion when the head slows down.

These can cause the walls to be thicker or thinner at different layers.

XY belt play can also cause layers to shift so I thought possibly that should be elminated since you mentioned it. But if it's extruder issues...

Some people try to fix this by having the extruder reverse directoin by 2mm or even 4mm. This is called "retraction" it can help, but sometimes it just makes things worse because the head pauses during the retraction.

Another thing that helps is to have very fast travel speed and very slow print speed so there just isn't much time to make thick strings. Say 50mm/sec for print and 250mm/sec for travel.

Another thing that helps is to have a lower print temperature. This is so the plastic is more like paste than like honey. That also makes it so that the extruding stops faster when it needs to stop and starts faster when it needs to start back up.

The fan can also make a difference - possibly if you turn the fan off you can go to a lower printing temp. If you go too low with the temp, then the pressure gets too high and you can get underextrusion (not enough plastic all over) which looks kind of like lots of holes everywhere evenlyish spaced.

 

[gadgetfreak 7]

gr5, that's a lot of really interesting information, thanks.

I have also in other models seen bad extrusion, like after a few layers the infill looks a bit sparse (regardless of PLA color). But since the intial layer always look good, I'm not sure where the root cause of that problem lies.

I will not have time to make more test prints of the house for some days now and I think I would most likely design some other test object with small pillars combined with walls to see if I can reproduce the problem without waiting a couple of hours into the print...

If somebody else [with well working mechanics, strecthed belts etc] would be able to make a test print of the first floor I would be really grateful.

@PrintedSolid, are you up for that? ;-)

 

[illuminarti 18]

If extrusion declines over time, it may be due to trying to print too fast with too thick of a layer height/wall thickness combo... leading to ever increasing pressure in the print head, slipping and a risk of blockages forming in the hot end.

 

[gadgetfreak 7]

illuminarti, any advice on what is considered "safe" or "too extreme" settings/combinations?

Is this model just to hard to print with Cura defaults? i.e. 50mm/s 0.2mm layer height (in my second try with similar results I used 0.15 lh) and 0.8mm wall thickness?

Thanks

[illuminarti 18]

I'm actually doing some testing of exactly that at the moment, and will post some results in a day or two. In general, 50mm/s with 0.2mm layers should be ok... it really depends on the plastic. You also need to make sure that your nozzle width is set to 0.4mm (if it is), because if the wall thickness isn't an exact multiple of the nozzle size, Cura changes the effective nozzle size... and you can start to run into problems then...

At 210, i wouldn't think you'd have problems... but it's possible your temperature calibration is off a bit... try raising the temp to say, 230 and see if that makes things better.... and bear in mind I mostly talking about the issues you mentioned where prints start well, and kinda tail off... I'm not sure that the problems in the house model are related to that.

Anyway, stay tuned and I'll post some more results in a few days detailing the safe range of temperatures, speeds, etc.

 

[illuminarti 18]

Yes... smaller things have the advantage that they typically involve smaller continuous print moves, compared to larger prints. As a result, on each move the head has less time to accelerate up to speed, so the head is typically moving slower on average. This means that the volume of plastic being extruded per second is less, so problems with limits on the extrusion rate are less likely to show up.

 

But as can be seen inte the gallery of pictures, the concentric circles print and 20mm box show no shifting. But those objects might be "easy" in comparison to wide-area objects?

 

 

[gr5 1,998]

You issues look more like you are printing too hot than too cold. I see no regions of under-extrusion. If I tried to print at 210C you would see lots of under-extrusion but I guess my thermocouple is off or something (maybe too much air from the fan hitting my tip).

Anyway, here are some samples of too low and too high temp (not my pic but it's very helpful):

http://umforum.ultimaker.com/index.php?/topic/1763-bad-surface-on-printed-parts/?p=12318

 

[illuminarti 18]

Just to clarify again... my comments about temperature and throughput were addressing gadgetfreak's comment about objects that start printing well, and then get sparse later on. I'm really not sure what the deal is with the house model...

The shifting of things seems to be more of an x-y mechanical or slicing problem than related to extrusion/temp etc.

 

You issues look more like you are printing too hot than too cold. I see no regions of under-extrusion. If I tried to print at 210C you would see lots of under-extrusion but I guess my thermocouple is off or something (maybe too much air from the fan hitting my tip).

Anyway, here are some samples of too low and too high temp (not my pic but it's very helpful):

http://umforum.ultimaker.com/index.php?/topic/1763-bad-surface-on-printed-parts/?p=12318

 

[gadgetfreak 7]

Related to this and the excellent picture from Jake showing different temperature effects on a print:

It would be so good if Ultimaker (the company) could select (with help form the community) a set of different calibration models that could be printed in Cura project planner with different temperatures (range of 5-10 degrees perhaps) to aid in calibration and troubleshooting.

Then Ultimaking could/should put up pictures of all of these test prints so that you could tell what a good print would look like (when mechanics are ok).

Do you think this is feasible? Or are there just too many combinations?

I think if there are 5-10 different models with pictures for different type of objects that would improve learning for many users.

Even one, two or three models that were standardized with pictures and temperatures / slicing profiles would be a big improvement in my eyes.

 

[illuminarti 18]

It could probably be done for some very simple objects, which would be helpful in getting the basics ironed out, and some confidence that that mechanics are ok - but there really are a lot of variables - not least the variations between different types of plastics. I think a lot of it just comes down to practice, and getting a sense of what settings work well for what types of objects, while minimizing the print time. In general, faster is not always better - the key is to print as fast as you can, while still getting acceptable quality results.

A big part of becoming an experienced 3D printer operator is realizing that there isn't some holy grail of perfect settings... it's a continuum with a lot of dimensions. The challenge is finding the sweet spot between not good enough, and better (slower) than I really need, while meeting the challenges of each model that you have to print.

 

[joergen 2]

I'm with illuminarti on this one... there isn't a perfect setting: each plastic is different, each heater element is different, each nozzle/barrel is a bit different, each model is different, in addition to different speed settings and layer settings:

extrusion temp is easy to find: for each plastic: heat up to 180C (for PLA) and start turning the big wheel by hand, and observe the extrusion. increase the temp in 5C steps until the extrusion looks shiny, not matte anymore. rinse&repeat for each of your PLA reels.

speed&cooling depends on the model and your experience: it is mostly a print, observe, identify the problem, fix it, and print again process. most problems in this regard are coming from not letting the previous layer cool down under the Tg of PLA (65C): printing on gooey PLA never yields a nice print.