banner perso

Some links about me. Many of my 3D designs are free. I also post on Google+ and in another blog, oz4.us
review (32) general (26) issue (16) mechanical (16) replacement (14) software (13) addon (12) bowden tube (10) business (10) consumables (10) heat (10) feeder (8) hot end (8) weird (7) motor (6) off-topic (6) trick (6) electric (4) bed (3) cnc (2)

Saturday, November 17, 2012

Short belt tensioner for NEMA 17 stepper motors

How to tension the Ultimaker short belts once for all?

Upside version of the tensioner
The short belts that drive the main rods from the X and Y steppers must always be kept tight. Otherwise the print quality really gets bad. Two symptom that are easy to spot is non-round cylinders or "bouncing" waves in the surface of the print.

The usual fix is to unscrew the 4 bolts, push the motor downwards with a thumb and tighten the screws very hard while keeping the pressure...

I just could not stand this anymore, so I designed this thing and this derivative (with an additional spring).

Downside version of the tensioner
There are four reasons why the "official" or "recommended" procedure fails in my opinion:
  • Unlike the main belts, the short belts seem always to get sloppier with time! So I regularly had to check and re-tighten the steppers.
  • (hence) the screws need to be tightened up to the point they start to dig in the plywood to stay in place, i.e. you need to damage the plywood! May be my plywood is not hard enough for the constraints, as you may see in the first picture that it started to bend on the side...
  • (hence) once the washers or nut heads went deep "enough" in the plywood to stay locked for a while, moving them further becomes dirtier even, precisely because they just want to slide back to their initial holes. So you need to tighten them even more and damage further your case...
  • Finally,  in my opinion, no one can achieve a precise belt tension by pushing a stepper with one hand, while screwing with the other one!

Fine, robust and stable belt tension adjuster (at last) !

So I finally addressed the issue by designing my own tensioner, and stop using a procedure that damages my printer while not really fixing the issue in the first place.

Printed with 0.15mm layer height, 100% infill. With or without PTFE thermal insulators.

Having a long screw to slide the steppers down gives both force and precision. And no more damage to the case. I also made them thin so they do not get in the way when I move the printer around, since they are on the outside of the case.

Finally, nylock nuts are cool but they tend to rotate freely in PLA if their slot was not exactly adjusted to their side. This is tricky because it depends much on temperature and printer settings, especially with thin walls.

The holes here are conic. It is easier to slide them in place than with a lateral/partial slot, and the more you screw, the more they are kept in place from the 6 sides. No need to finely tune your printer (which is hard if your short belts have the backslash we're just trying to fix...)

Finally, the stepper sometimes become hot enough to meld PLA. I do not fear much that the heat could propagate via the mounting screw and through the plywood layers, but I designed a special version that uses two 3mm pieces of bowden tubes as thermal insulators on the two stepper screws (see above). I used bigger washers for these to dissipate heat further.

I made two versions of the belt tensioner, where the bolt is facing upwards or downwards

Pros and cons of a spring in the design

As a word of caution: do never over-tighten a belt or you will ruin the bearings!

The setup I described so far lets you go much farther than what is required. After some discussion about it in the forum, I made a derivative which adds a spring between the screw head and its washer.

Keeping the stepper loose enough to slide along the wall may not be a good idea because the stepper may start resonate when direction changes (and you'll get wavy vertical textures around the corners).

Spring-loaded tensioner with a loose stepper.

Anyhow, the spring better takes care of the tension and probably dampens excessive settings. Time will tell if the revision really is worthwhile.

Spring-loaded version of the same tensioner


  1. Dad the same problem ages ago. I put threaded rods in the original motor mount going down trough the bottom plate to fix this...

    never had to change it since.

    your plates are a little less invasive perhaps but do require more parts.

  2. Hey MoonCactus, Great blog, learning many cool tricks and your upgrades are great. I was planning to print your this tensioner when I just found about this "direct drive" upgrade, consisting in mounting the steppers in line with the rods, eliminating the short belts altogether:


    Any thought on that?

  3. I share the idea but sincerely, I do not think that any backlash is left out with tmy tensioners (especially on so short belts)... There is not much to be gained here, and there are other and much bigger sources of mechanical errors (most prominently due to the longer belts, bed shaking and so).

    A direct drive setup may enhance the print but it would surely not be in the top of my list once you fixed the short belt correctly. But to me, the cons are unrelated actually: the stepper motors protrude so much out of the ultimaker frame when direct-driven that I really would fear moving it around (and it takes much more room to pack). May not be a problem for some though.