@Sven_Eric_Nielsen didn’t know about that CoreXY design but didn’t the fixed points on the carriage are in different highs? So instead of crossing i now have a static imbalance. And this could be archived also with the left design just by bringing one belt into a higher level. From the drawing of the right version: left red is above and on the right side below the blue line - how without crossing?
(and you also can cross belts http://corexy.com/implementation.html ) not sure which one is better - i don’t like both versions as they are both not symetrically (OCD trigger )
I mean the coreXY is a very good approach and is working well. But you don’t have the double precision which should halve errors (unround pulleys, streched belts)
The remote extruder need a seperation or adaption of the Firmeware as XY will influence the position and need to be taken into account.
printing pulleys is for weight reduction and my approach is to limit parts you need to buy. ( I can buy a CNC system for 20k if i just need a working solution )
Crossing CoreXY belts was never a problem, even back when people crossed belts. There was never any downside whatsoever with letting them rub at the crossing. People who never tried it just think it looks wrong. I have done it both ways and there is no difference.
Now, nobody crosses CoreXY belts anymore, because staggering the heights also works 100% fine without any problems. It looks wrong at first sight to have the belts not line up, but again, people who haven’t tried it only think it looks wrong. Why? No belt drives ever apply all belt tension and acceleration forces directly through ALL of: 1) the center of mass, 2) the center of bearing friction, and 3) the center of bearing stiffness. That is practically impossible to do. So the linear hardware always must resist some minor moment loads. Staggered belt heights in CoreXY is a VERY minor load compared to corner jerk forces on a moving extruder motor.
Belt stiffness in a drivetrain is a function of 1) the belt elasticity (obvious) and 2) the length of the SHORTER of the two belt runs between the motor and load (slack side or tight side). It’s a bit complex to explain why, but basically a short run of belt has much higher stiffness than a long run of belt, so the short run tends to change tension faster in response to motor action, and thus is more responsible for moving the load.
So, to maximize belt drivetrain stiffness, you want at least one side of the belt loop to be as short as possible. In CoreXY, one side of each belt loop ranges from ~0 length to ~X+Y length, which is not all that long.
Compare your design to a typical Replicator 1/2 or FlashForge style gantry where there are two separate Y belts to drive the two sides of the bridge. The load is split between two belts (that have minimum possible length) so it is really very stiff. Yes, the 2:1 pulley reduction on your design is good for motor torque, but will be less stiff in the belts than a FlashForge style design. (This is a very minor point and I suspect the 2:1 will do more good than the difference in belt length does harm.)
Another benefit of CoreXY that has not been mentioned is that belt preload tension inherently forces the gantry to be square. No particular rigidity is required for the bearings on the ends of the bridge. Of course we all know that the HBot drive forces cause racking, but even if HBot X-axis acceleration did not cause racking, the H-bot design does nothing to stabilize the two sides of the gantry to stay square. It all relies on the linear bearings.
Likewise, here, there is nothing but linear bearings to hold the gantry square. So when your XY carriage is on one side of the bridge, and there is more mass on that side, the belt loops can’t apply more force to that end of the gantry, and the linear bearings will have to resist the moment load. It’s not as bad as HBot (where X acceleration loads actively push the gantry crooked) but your design still does rely on linear bearing stiffness to stay square during certain moves.
In my design the center of force is at the center of mass of the carriage And also of the bridge for the other axis - and as the carriage run within a tri-angle (the two dimensional bridges show more flexion). And the difference in belt length will balance itself over the dividing pulley. I think this is a benefit especially if build with cheaper parts or on bigger scale.
But as already told the coreXY works great - but i am not so sure what happens with the additional load of the belt extruder, and i am hopping that my design can take those additional loads better.
The blue belt forces on the bridge+carriage don’t have any way to apply more force on the side of the bridge that has the carriage and thus more mass. So you’re applying a “balanced” force to an unbalanced load. That will try to rack the gantry. See drawing.
The extruder drive belt forces will also try to rack the gantry.
Even if these loads are small, they’re still trying to make your gantry crooked. Why use a belt arrangement that relies on the linear bearings to keep the gantry square (like an HBot) when you can use an arrangement where belt forces help keep the gantry square? Every popular XY gantry has belts keep the gantry square. (CoreXY, vanilla XY Cartesian bridge gantry, Ultimaker cross gantry.)
Is that two pulleys on one motor for two belt loops? (Doable, just wondering.) Looks like it will work and get you the 2:1 reduction. I might suggest using two separate motors just to reduce the idler count and belt routing complexity. Not a big deal though.
It’s only one pulley on the motor used, the path of contact with the belt will be lower (~120°) - maybe require bigger pulley.
Wonder how the X-belt layout look on the “maukcc” cartesio - the Y is more or less identical to my design (same on the plotter). But if the X -axis doesn’t uses a gun tackle i get a printer with different XY resolutions (doesn’t like that)
probably right, much more importand to have no backlash, low clearance - no slack. So low resolution is ok if the precision is high (and without precision - high resolution is useless)
In my opinion, the pulley reduction is more important for actuator stiffness than the “resolution” of the print. (Motor step resolution is already higher than the nozzle size or path detail decimation of modern slicers, so who cares if it’s 10um or 5um microsteps?) But when you do a 2:1 pulley reduction you get half the belt stretch, and half the load on the motor. Half motor load on a 2:1 pulley reduction means ~1/4th as much motor overshoot error.
)