Wednesday, 18 May 2011

Unexpected mechanical faliure

After my filament tangle and snapping incident printing the Octopus I started printing the great looking minimalistic spool (thing 8472) by whosawhatsis when half way through the print some odd noises started to happen and then my X axis Flexi-drive snapped.

After removal.

I was just about half way through the print when it happened, I shut the machine down and investigated.


Just after it snapped.


Old photo of before the failure - clearly bad motor-shaft alignment.
I found a photo of what is looked like before it snapped and the only thing I could see were two of the flexi-compression rings seemed to be very close together and the whole flexi-drive looked compressed.
So I expect that it was under some strain and eventually failed.

I still don't really know why it snapped. I'm not using acceleration so maybe along with the NEMA23 motors running with 24V  I'm pushing my machine too fast and hard?

New installed flexi-drive and silicon pad + washers to help get better horizontal alignment.
I had a spare flexi-drive so I fitted that and used washers to space the motor out to give better alignment and a small silicone pad under the motor mount to allow some freedom for any oscillation, I hope this will not make the problem worse, time will tell if this happens again or maybe I was just unlucky with a weak part? or it was just the bad motor shaft alignment.

Not wanting to waste a good hour of printing I wanted to restart it, so after a height measurement of the part I concluded it must have failed around Layer 62, so a quick edit of the GCODE to start at layer 61, I re-started the print, the first layer bonded well with the last and it was off again. 
First useful part printed. and one of the largest things I have printed so far.
The print finished later with no other issues and part is usable if not a little out of alignment where I nudged the opto-switch when taking the motor off.

I also checked the Y axis motor mount and that looks fine, so I hope this was just a one off event.

I will print the Filament mount arms tonight and hopefully assemble it, that will be the first useful thing it's printed and it should allow me to leave the machine for longer without worrying about filament tangle.

Thanks to whosawhatsis for the great model, I'm sure this is going to look great on my machine.

Rich.

3 comments:

  1. Glad you like it!

    WRT the layer alignment, my X axis skipped during an attempt to print this part, and though I managed to get it to where it appeared to be properly aligned after a couple of layers, I couldn't get the bearings to align properly. This prevented them from turning smoothly and made the whole assembly wobble, so I had to scrap the print.

    You should be able to print it faster by reducing the infill, though. Give it two shells and three or four layers for solid surfaces, and you don't need much infill at all. Making it too heavy won't help the functionality, and mine is plenty strong enough printed as I described.

    ReplyDelete
  2. Hi, Thanks.

    I have not experimented with more shells yet, but I need to take the skeinforge plunge soon so I'll have loads more options to play with.

    I just got the 3 arms printed with less infill, they work really great. the bearings fitted perfectly and alignment on the studding is very good. The filament holder is now in use and I'm just printing another Octopus. I'll post a photo of it fitted onto my machine tomorrow.

    ReplyDelete
  3. Looking at your photos it does not seem to be 'in alignment'. These devices are frequently Aluminium and it looks like classic Aluminium fatigue.

    Check the maximum permissible miss-alignment for the coupling and measure your miss-alignment with a dial indicator or equivalent home brew device. It's tricky to tell for sure but it looks to me like you have exceeded it. Couplings have to be aligned in three axis and the total miss-alignment accounted for. It may seem strange but the job of this coupling is to allow for 'minor' miss alignment and to prevent axial thrust transfer into the bearings of say a rather expensive shaft encoder.

    http://en.wikipedia.org/wiki/Fatigue_limit

    Simple rule, coupling life is directly related to degree of alignment.

    Think of it this way. Aluminium work hardens and you are flexing it every time it rotates!

    ReplyDelete