Combating Warping

Look for guides here.

Moderator: CrazyIvan

Post Reply
User avatar
CrazyIvan
Posts: 133
Joined: Mon Jul 10, 2023 10:25 am
Has thanked: 10 times
Been thanked: 18 times

Combating Warping

Post by CrazyIvan »

Okay, this isn't so much a guide/tutorial as a discussion, but I figure it comes under the general category of education, so...

What Is Warping?

Well look, I only have a few months experience at the sharp end (hot end? :lol: ) of FDM printing, but I'm already pretty familiar with warping. It seems to affect mainly the underside of horizontal surfaces.

Depending on the geometry of the print:
  • The first layer might lift off the print bed around the edges;
  • The magnetic plate might lift (Tina2);
  • An overhang might lift off the support layers.
Here are some examples:

Image

In the above, the warped part was the other way up when printed. I cured it by re-printing with the opposite orientation, so that any warping would then be confined to surfaces which don't matter.

Image

This roller was printed with the left side downwards. The warping is in the underside (as printed) of the flanges. What really surprised me was the short scale over which this happened: the flanges are only a few millimeters.

I have also had problems with warping lifting the first few layers, which then catch on the nozzle or proximity sensor and shifts the print, and then catches again later on and shifts the print multiple times. The result is a mess of printing into free air. I now watch the first two (at least) layers going down, rather than come back after several hours and find a ruined print and wasted filament when I could have aborted it.

Granted my Tina2 printer has no volume or bed heating, but I'm only using PLA (which is supposed to suffer least) and the build area is only 100x120mm, so warping will be much worse with ABS or bigger builds.

How Does Warping Happen?

I'm open to discussion, but it seems to me this is entirely about shrinkage when the extruded filament cools. A layer goes down, and shrinks slightly as it cools. The next layer goes down and also shrinks as it cools, but it is on top of (and bonded to) a layer which has already shrunk. The consequence is to build in strain, and if the strain exceeds the grip of the layers onto the build surface or other layers below (eg a support roof), they will flex like a bimetallic strip. It only takes a very small differential contraction for this to happen.

Presuming the flexure does not obstruct the print head, subsequent layers will print on top and add further strain, but this effect eventually peters out, and as the print head is where it should be in Z rather than tracking the layer underneath, after a few layers there is no further warping.

How Can We Stop It?

When the filament comes out of the nozzle, it is liquid. There can be no strain in a liquid. The problem is between the temperature at which the filament first solidifies and the temperature it is at when subsequent layers come along. If the model was maintained at the solidification temperature until the print is finished, then the whole model could be cooled uniformly post-print... but that temperature is rather high.

The purpose of a heated print bed and/or heated build volume is to slow the shrinkage and therefore reduce the strain between the layers. A heated bed improves adhesion and heats the build volume indirectly. Making the temperature very high (as might be optimal) takes a lot of energy, but also is very bad for the printer mechanics and electronics.

The Tina2 has neither. I have achieved some benefit with a difficult print by applying a hot air gun while the first few layers went down, but it made me very nervous that I might be overheating the printer components.

Other Strategies

Can we design out warping? Vertical surfaces don't warp, and top surfaces tend to be flat, so the appropriate choice of orientation in the slicer should help. A cube, for example, could be printed standing on a corner, and not require much support if the overhang angles are slight enough.

Non-solid layers should cause less strain; where an internal volume is hollow or contains support, it will conform to the layer below better. But that doesn't help for the skin layers.

–––––––

Feel free to comment! I throw the floor open...

Post Reply