Two Trees Saphire E3D + Belt Tensioner

thingiverse

The Two Trees Saphire Sp3 is a great machine but lacks a good hot end and a belt tensioner. The attched files are for a new hot end design which uses an E3D hot end and a 40mm heat sink cooling fan. The revised hot end includes a simple attachment for the belts and uses an independent belt tensioner located at the rear center. All the parts can be printed without supports and 0.2mm layer height. M3 holes require starting with a tap and M2 threads can be formed by forcing an M2 screw into the provided holes. I like to see the nozzle and resultant print as it progresses and the design takes that into account. The belts are attached to the 'Top Mount' by simlply looping around a 3mm dia by 6mm rod with the toothed sides together and inserting into the provided slots. It is relatively easy to obtain acceptable belt tension but the separate belt tensioner will take up any slack and make it easy to fine tune the tension. The part cooling is particularly effective and I have opted to run the fans at 65% which allows another 35% for small layers. And the part cooling fans are arranged in a way that does not reduce the printable area (many designs do reduce the printable area). The belt tensioner uses F623ZZ bearings, two required for each idler, and 'Gt2 16T 3B Toothed Idler Pulleys'. The design was developed in OnShape (https://cad.onshape.com/documents/6d57404aef3a42faf014913d/w/4829faaf0e06f8885349cca3/e/ac61b732be56eb6913587198?renderMode=0&uiState=639481c0899ac21de95fcca1) Belt Tension, Ringing and Acceleration. Published information suggests that increasing belt tension increases resonance frequency but perhaps more importantly significantly increases the magnitude of resonance. Testing that I have done supports this information. Clearly belts need to be tight enough to ensure good positional accuracy but over-tight belts have a detrimental effect on resonance and ringing on prints. Acceleration tuning shows that over-tight belts encourage ringing and an acceptable value of 1000mm/s2 or lower is possible. Reduced belt tension virtually eliminates ringing and allow 2000mm/s2 or even greater. Jerk or junction deviation are also impacted (I am currently using junction deviation of 0.03). I have added a belt tension gauge stl which works extremely well.