Geared feeder upgrade for UM2 (GT2 belt reduction)

[meduza 191]

Hello

I have been working on a feeder upgrade for the UM2, that is intended to solve 2 problems with the UM2 feeder.

1.) Direct drive feeder with a Nema17 makes it quite weak, therefore i am doing a belt drive reduction of 20/36 = 1:8 (a reduction of 16/36 = 2:25 is also possible with the same printed parts).

2.) The motor shaft gets very hot, which in turn makes the feeder wheel hot, and when it approaches the tg of the plastic the filament gets soft and the feeder wheel tends to grind, the belt drive reduction decouples the thermal path between the motor shaft and the feeder shaft and removes this problem completely.

I do not want to design a new feeder tough, there is people that have put a lot work into that already, instead i want to focus on making a universal belt drive reduction that can fit all kinds of feeders already designed to fit the UM2, including but not limited to Roberts feeder.

Now i am on the point of the 5th printed protype and have something that mechanically should be working (not test-fitted it on a UM2 yet), but could use quite a lot of refinement which i will be working on in the next few weeks.

It uses only easy to source parts, the shaft could be a 5x40mm dowel pin, the pulleys are 36T and 20T GT2 pulleys, the belt is a 110T GT2 belt loop and there is also two 5x10x4mm bearings in there to take the load on the feeder shaft, other than that it is just some M3 screws and nuts and printed parts.

The print of the reduction base is not looking that good, quite possibly because i printed the prototype at 0.2mm layers and quite high speed

This is just to show how it should be fitted to a UM2 back, this picture is of a old revision.

 

[fabi580 0]

great idea

 

[fabi580 0]

It should work with a M500.

 

[Dim3nsioneer 557]

It should work with a M500.

 

You mean a M92 followed by a M500, don't you?

 

[meduza 191]

When you are saying more wheels, are you then thinking about multiple feeder wheels in line?

If so i think that would be a overly complex solution.

I do think that a solution with dual driven larger feeder wheels that drive the filament from both sides would be a much better solution than the one we use today tough.

A heatsink for the stepper motor could lower the temperature somewhat, the extreme would be forced cooling with a fan, but the stepper motor covers does not have space to do that, so you have to remove those.

Your suggestion of cool-down would require a quite time-consuming cycle, it would be something like this:

1.) retracting filament from the hotend

2.) turning off the heater

3.) turning off the stepper driver to the E axis

4.) waiting maybe 15mins (it is quite some mass to cool down)

5.) reheating the hotend

6.) turning on the stepper driver

7.) priming the nozzle

8.) resuming printing

This would add an impractical amount of time to your prints, so i do not think this is a way forward.

I still think that the best solution is to thermally separate the stepper motor axis from the feeder wheel.

 

[fabi580 0]

You mean a M92 followed by a M500, don't you?

 

Yes of cause, sorry

 

[meduza 191]

I am in the process of redrawing this prototype pictured in the first post in Autodesk Fusion 360 instead of Sketchup since i am trying to change over all my projects, and will probably have something ready to try in a few weeks (maybe sooner if i have the time), will then publish it on Youmagine and write in this thread.