3D Genius

By July 17, 2013 3 Comments

A bridge too far



Download this thing We all know that 3D printers can’t print onto thin air – you need supporting structure to hold the printed filament up from beneath. So for example, when printing a figure of a person, you can print up the legs, bug you can’t print down the arms, unless they’re folded or the hands are resting on the hips. But what you can do is to persuade your printer to stretch filament between surprisingly large gaps. This enables you to print boxes with lids, and many other objects that have unsupported roofs.

To test the bridging capabilities, we’ve used a model adapted from triffid_hunter‘s Bridge Torture Test. All we’ve done is to lower the pillars by half, so reducing the printing time. You can download our versions, in a variety of bridge lengths, here.

But how do you get decent prints? The distance between supports is a key consideration, of course, but the speed of printing is also a major issue. First, let’s see how exactly bridges are built, so we can understand the processes involved.


bridge03The bridging process

After printing the supports, your printer will stretch thin strands between them, all parallel to each other. It’s at this point that sagging is likely to occur, and if anything is going to go wrong, this is the most usual place for it to happen.

The first couple of strands mark out the edges of the bridge. We used a bridge length of 50mm (just under 2 inches) for the initial test, and most printers should be capable of stretching this distance.


The printer will then go back and forth between the support posts, dragging a strand of filament each time it passes between them. At this stage there’s no real structure here, just a network of fine strands.

Depending on your print speed, as we’ll see later, you might find one or two strands that droop down below the others. This occurs when a strand is too hot, or not under enough tension because too much filament has been extruded. It only takes a tiny variation in temperature or extrusion speed to produce a noticeable sag. In practice, these odd loose strands can easily be broken off the finished print later, so aren’t too much of a cause for concern – except when we’re deliberately pushing the printer to its limits.

bridge05After all the parallel strands have been laid down, the printer will then go back over them printing a zig-zag layer across these strands. This is the beginning of the strength of the bridge construction.

Zigzagging across those fine strands can produce unwanted bends in the structure, depending on the length of the bridge. As the filament is extruded it will necessarily tug on the existing strands to some degree, and may force them to bend or even break as it pulls across them.


The speed difference

The photograph at the top of this page shows the result of printing our 50mm bridge at four different speeds: a leisurely 25mm (~1 inch) per second, a typical 50mm (2 inches) per second, a speedy 75mm (3 inches) per second, and a super-fast 100mm (4 inches) per second. The faster speed prints are towards the front.

When we turn the prints over, we can see more clearly how ragged the first printed layer really is:


As you can see, the faster the print, the more tendency there is for bridges to sag and produce unwanted hanging strands. When it comes to bridge printing, you should definitely err on the cautious side: it seems that print speed has a lot to do with finished quality.

Note that all these tests were performed on our Ultimaker printer, which is known for its speed of printing; you may find your printer exhibits unwanted sagging at lower print speeds.


How far can you bridge?

Now that we’ve established that successful bridging depends mainly on print speed, the question remains: just how far can you persuade a printer to stretch a filament?

In addition to the 50mm (2 inches) of the original speed tests, we experimented with gaps of 75mm (3 inches), 100mm (4 inches) and even 150mm (6 inches). To our surprise, the printer managed, to some extent, to draw filament between all three gaps, even the largest one. Here’s the result of those three tests:


There’s clearly some major sagging going on in the 150mm (6 inch) bridge, but we were surprised it managed to print this one at all. The 100mm (4 inch) bridge showed much less sagging, although there were a number of loose strands hanging down; and the 74mm (3 inch) bridge showed very little sagging.

All of this is very good news for 3D print designers, as it means we can cheerfully span surprisingly large distances without too much concern about losing the odd strand here and there. Certainly, it opens up the field for much more ambitious, detailed designs.

But if you find yourself faced with a bridge that isn’t working, we recommend you reduce the print speed and try again.

Share it!
Posted in: Design, Technique

3 Comments on "A bridge too far"

Trackback | Comments RSS Feed

  1. Marcelo says:

    Good experiment. So the other option is to put support columns but spaced much farther apart than Slic3rs default 2.5mm. We could put tham at 50mm and minimize the sagging. What about filament temperature? would more or less affect the bridge? Also how about cooling with a fan? PLA vs ABS vs Nylon?

    • Steve Caplin says:

      Filament temperature, surprisingly, doesn’t seem to make a lot of difference. I use a fan with PLA and I do think the cooling helps, though, as it stiffens the strands as they’re being extruded.

  2. Sylvain says:

    I tried this bridge, I still have to do some work, 2 things that helped me was:

    1. My filament multiplier was to high, decreased, tested until the filament was not outputting to much filament while making a bridge…

    2. My fan was blowing to much air on the filament, so you can disable or decrease the fan speed while bridging, this can help

    3. Decrease the speed

Post a Comment