Belt Tension Tester Using a Dial Indicator

This meter measures the displacement of a belt when a force is applied to determine the tension in that belt. Using a 30mm or 50mm span to support the belt and a dial indicator to press down in the middle of the supports, the belt tension can be calculated from the basic force vector diagram of the belt being depressed with the 2 legs as the supports.

• Tension=(Span x Force) / ( 4 x Deflection)

A recommended tension for a GT2 belt is 111N for a 1" wide belt, for a 6mm wide belt this tension would be reduced to 26N. 

Another recommendation “Gates, the creator of the GT2 belt, recommends a tension of 6 to 8 pounds of tension when using a 6mm GT2 belt for registration like in a printer.” 6 to 8 pounds is 27N to 36N for a 6mm belt. 

for example (working in MKS units)

• 50mm span for the wider tester is 0.05m, 
• 0.7mm deflection read from the DT is 0.0007m
• 150g weight that causes the deflection will apply a force of 0.15kg x 9.8 = 1.47N

Using these values in the equation above for tension in the belt

• Tension=(0.05 x 1.47) / ( 4 x 0.0007 ) = 26.25N 

To use the tester

• Assemble the DTI into the holder so that the test end is about the same height as the top surface of the legs. A 3mm screw can be used to hold the DTI but mine was tight enough without the screw.
• Move the head or bed so that a long section of belt is exposed
• Lift the belt and slide the 2 legs under the belt at the same time as lifting the DTI probe so that it will push down on the belt when in position
• Zero the DTI when the belt is in position
• Add a weight to the top of the DTI movable probe, I used 150g which depressed the belt about 0.4mm. 

Measuring with the tester: -My first attempts at this measurement and for my printer the tension was 28N and 18N for 5.7mm wide belts. 

Comparing the readings with several weights and a nylon string in place of a belt I found the measurement to be between 10 and 20% over the correct value (it reads higher than expected). 

I also tested with a weight attached to my Ender 3V2 Y axis tension adjustment. Larger weights give less error so as you get close to your desired tension, the error will be less. The likely culprit is friction preventing it from going the full distance when the weight is applied. The belt resistance to bending will also detract from the reading's accuracy. From the measurements I know that, using the wide holder, I need to see a deflection of 0.7mm when applying a weight of 150g to the DTI plunger for a belt tension of 21N.

Having checked out some commercial meters it seems to me they calibrate for the belt width and material, some have a selection of belt types to choose from built into the meter.

Let me know if you think it can be improved or, even better, do your own remix. 

I tried to keep it compact for the small dimensions accessible on a 3D printer. I've added a “wide” version if you have the space. Also added a tip for the DTI probe to make contact across the full width of the belt, not really required.

An improvement/remix would be a better system of adding weight to the top of the DTI. I added a “cap” that fits over the DTI plunger and with the ¼" x 20 threaded hole you can add your own design to hold a weight. I showed 2 examples that I used, and old chuck that weighs 227g and causes a deflection of 1.26mm. The cup with screws can be adjusted to any weight you want, shown is 150g.

The Prusament tension meter looks like a good alternative but they only supply g-code and require that you use a particular filament - all quite reasonable if you want repeatable results. Tom Anderson made one and included Fusion360 files. This one works but you may want to calibrate it yourself. Chip tells you how to calibrate. Mine reads a little low but not by much.

I have recently searched for “belt tension indicator” and found several 3D designs on Thingiverse, In particular:-

This one by kferic looks similar although they give no explanation. I think they use the DTI spring to supply the deflection force.

This one by calviniba (remixed by Digitalpaul)is the same idea but has the scale built in and doesn't need a DTI. It is calibrated as long as you get the same spring and don't have too much friction.

The frequency method. The most accurate way.

Knowing the belt density you can determine the tension in a belt by plucking it like a guitar string and recording the resonance frequency (Mersenne's equation 22). Using my phone and a reference setup (a single 6mm GT2 belt with 26N of tension - just below 6 pounds) I checked this method.

My belt was 5013mm long and weighed 56.2g so the density was 0.0112 kg/m

For three different lengths between contact points I measured the frequencies of the belt as I plucked it. They were less than 1% under with the most error being when the length of belt resonating was the shortest.