LinuxCNCKnowledgeBase | RecentChanges | PageIndex | Preferences | LinuxCNC.org

No diff available--this is the first major revision. (no other diffs)
Quick Overview of our conversion of a generic HM45 mill into an EMC controlled, linear scale based CnC? Machine.

[High Res]

We decided from the outset that we wanted the ability do have decent accuracy without it costing a mint. Accurate ballscrews cost far too much and so we decided to use linear scales on the axies to get the best accuracy and repeatability, regardless of the tollerence on the screws.

Doing this however posed some problems, mainy all backlash must be eliminated or the servo loop might start seeking making the mill bed jump around in a noisy, scary and expensive fashion.

We decided that we didn't trust the chinese anti-backlash ballnuts and as such we made our own anti-backlash system with 2 ballnuts and some belville washers, the floating ballnut works against the fixed one to keep it pressed against one side of the screw, the belville washers providing the preload.

In addition to this we needed to take the backlash out of the drive system, so we opted to direct drive the ballscrews, this meant we needed some large torquey motors, after looking at magmotors and such like it looked like they would work but be too expensive. Instead we decided to try 36V scooter motors from oatley electronics, and they seemed to do the trick.

To drive the motors we decided on the mesa chain of products with a 5i23 in the PC and a pair of 7i40's to controll the motors. Unfortunatly the current limit they come with was set a little low for the use we were wanting, as such we "modified" the 7i40s to defeat the current limiting and to also keep the power off the PCB lest it get crispy, In use these mods don't even get warm when driving the non-counterweighted 100kg head up and down.

The Z drive we initally tried to CnC? by automating the quill however the off axis forces and the lack of a precision slide (restricint the rotation) meant that this was doomed to failure as the whole linear scale segment would oscilate as the quill went up and down.

So instead we put a ballscrew setup into the column and drive that up and down, unfortunatly in its un-counterweighted state the Z drive motor didn't have enough power to push the head up and down, stalling occasionally and pulling ~20-30A in doing so. As such we have put a temporary gear reduction on the Z drive, using more scooter parts to do so.

We hope shortly to be getting a beefy gas lift sruts to counterbalance the 100kg head on the mill and overcome the problems we are having tuning the PID to handle the unbalanced forces between going up and down.

LinuxCNCKnowledgeBase | RecentChanges | PageIndex | Preferences | LinuxCNC.org
This page is read-only. Follow the BasicSteps to edit pages. | View other revisions
Last edited February 19, 2010 3:29 am by Russell (diff)
Published under a Creative Commons License