Drill on UM2+ with Mark2 magnetic mount

[conny_g 251]

Wanted to try that for a long time, now at least the mount is ready:

Dremel flex extension mounted on the UM2+ head using the magnetic mounting plate of the Mark2 dual extruder project.

 

Currently printing a modification of the plate (needs to be 2mm thicker and the magnets are too far in, used a version of the file that had this "bug" for the plate), then I will write a g-code generation script for drilling and test it.

Depends on the pressure required for drilling if that's going to work, the mount will not withstand very much pressure, maybe some more once the magnets stand a bit more out so the magnets of the plate on the head the plate on the drill can physically touch - currently they are like 0,5-1mm apart and the force is significantly reduced.

 

Edited July 2, 2018 by conny_g

[foehnsturm 969]

Cool!!?

[geert_2 557]

Cool indeed. It would turn the printer into a light CNC machine.

 

Can the holder withstand the occasional getting stuck or slamming of the drills? If not, maybe you could consider fixing the clamp with wing screws, or so, instead of magnets? This also makes fast assembling possible.

 

Just make sure you level the bed correctly before starting to drill.  :)

 

[Dim3nsioneer 557]

While I like the versatileness of that mod I recommend to protect the bushings and bearings from dust.

[conny_g 251]

4 hours ago, geert_2 said:

Cool indeed. It would turn the printer into a light CNC machine.

 

Can the holder withstand the occasional getting stuck or slamming of the drills? If not, maybe you could consider fixing the clamp with wing screws, or so, instead of magnets? This also makes fast assembling possible.

 

Just make sure you level the bed correctly before starting to drill.  ?

 

 

Yes, that’s to be tested if this kind of mount holds that. I wanted to try the magnetic mount first as the magnetic plate is already on the head. But of course there’d be other possibilities. 

As the Dremel does not need automatic docking and undocking there could be some „nose“ or „edge“ taking the vertical forces instead of the magnets.

From manual drilling of PCBs I know that the perfect-quality drills need little force in and out. While other drills can be a bit nasty.

 

Bed leveling: I was planning for air below the PCB or some wood panel ?

 

4 hours ago, Dim3nsioneer said:

While I like the versatileness of that mod I recommend to protect the bushings and bearings from dust.

 

Yes, I‘ll see how much dust there is and I planned to mount a small vacuum nozzle as well.

Manual drilling of PCBs actually doesn‘t cause too much dirt, was not planning to use it as CNC...

[conny_g 251]

After redoing the Dremel side mounting plate so that the magnets stand out more and can touch the head side plate magnets it holds quite well. I can press from the bottom straight up with some 2-3kg of force and the Dremel does not move at all. Promising.
 

So I have now found a simple script converting Eagle drill files to CNC g-code (https://github.com/kdupreez/excellon2gcode) and modified it to generate UM2 g-code and interpret the coordinates correctly.

The first dry runs on basis of a real PCB layout look good, see video. Tomorrow I will test with real PCB and drill ?

Of course the Z speed/feedrate is now set higher to see better what happens through the camera (not at the printer at the moment, just observing it through the web cam).

For the real drilling this should probably be slower at first.

And I am ignoring tool changes at the moment, that's next in the script after the actual drill test succeded.
 

 

 

Edited July 5, 2018 by conny_g

[conny_g 251]

Short: it works!

 

I killed one drill bit with a wrong height setting (the very big hole front left, the drill is still in there...), the 2nd try worked.

Less than 2mins for 35 holes and the dirt/dust is rather limited, took 30 seconds to remove that with a hand vac. Still would add a vac tube to the drill holder later.

Now I need a proper PCB holder that can quickly be mounted to the build plate and some calibration procedure to set the drill height quickly each time and to map the hole coordinates to the actual place of the PCB.
The idea is to position the drill to two diagonal corners, use controls to correct/adjust the drill position and then use that offset to remap the coordinates.
 

 

 

 

Edited July 6, 2018 by conny_g

[conny_g 251]

Prototype of a PCB holder plate is printing.

Based on T-nut system. Inspired by https://www.thingiverse.com/thing:2202041, but trying to do it less complex/massive.

Actual T-nut holder in design process, possibly printing that later today.

Will get mounted on a laser cut UM2 printbed sized acrylic plate in two layers to achieve the 4-5mm of thickness and holes for the screw heads on the bottom side.

 

Edited July 15, 2018 by conny_g

[conny_g 251]

Works better than expected.
PCB is held in grooves 5-6mm above the T-nut plate, so the drill will not hurt the plate if calibrated correctly.

 

 

 

[conny_g 251]

PCB holder mount on an acrylic build plate replacement is done. I can now just swap the glass into the acrylic plate and start to drill.

Have done this for the UV laser exposure solution, too - if the measurements of the "build plate" is accurate there is actually little tolerance in the placement of the frame/holder. For the exposure plate it's only some 0,5mm I usually need to correct the home position and also just 0,5-1mm in angle error for the 160mm PCB length.

 

It's laser cut from scrap pieces of acrylic. Bottom side transparent 2mm, top side white 2mm, glued together. Resulting in the 4mm height also the glass plate has. So the glass plate mount works for this plate, too.
Melted thread inserts into the bottom of the PCB holder and cut a larger hole in the lower side acrylic so that the holder can be screwed from the bottom with the screw head countersunk into the acrylic.

That's the core reason for using the 2 layers of acrylic - aside of looking nice (could have used 4mm plywood, too) I can countersink the screw.

Now there is some more software work to do for the calibration process (home position, angle error).
And to enable multiple copies on one Euro PCB. The exposure scripts can do that to multiply the PNG export from eagle and position it X/Y multiple times, but the drill script currently only uses the one set of hole coordinates exported from eagle.

 

[geert_2 557]

Congratulations! This is a magnificient add-on: a relatively simple concept, but very versatile. This basic concept could also be useful for engraving (like nameplates), lasercutting, and similar things which require high accuracy but relatively light mechanical loads.

 

A question: when drilling with very small drills (0.5mm), how is stability of the drill head? Does it wobble? When last doing electronics (>10 years ago), I found that this was the biggest problem when drilling holes manually: drills could easily break at the slightest wobble, sometimes due to not holding the board well enough, sometimes due to the slight wobble in the drill head (standard but worn out Dremel).

 

[conny_g 251]

@geert_2,

I actually use this mount method already for a diode laser of 150mW for PCB exposure. But it could also be higher power for cutting or engraving.

Ah, and I have used the 150mW laser for cutting paper stencils for SMD PCBs. Requires to paint the paper black with a laser printer to make sure the low power laser can cut it (black absorbs the energy better and makes sure the laser can pierce through the paper at all) and it cuts veeeeery slow (like 1mm per second), but it works.
Still less work than putting the solder paste on the PCB manually.... don't care if the stencil takes 20-30mins to make.
If you are interested, the whole process is: export the pads as Gerber, use a python script to convert the Gerber to g-code for the 3D printer, black the paper with laser printer (ideally two times for shiny surface = less absorbent, easier to distribute the paste), cut the stencil with the laser, spray the black surface with transparent lacquer to make it less absorbent. After drying for 1 hour you can use it as a stencil, even more than once.

Engraving (not laser, regular engraving) I am not sure if that's going to work. The magnetic mount holds fine for light forces in Z direction, a few newton. But it will fall off with more force in that direction or even with lighter forces in X/Y.
But of course the mount could be strengthened - it's just a bit more redesign. My motivation was to do as little as possible of (re)design on top of the Mark2 fixture to add the Dremel. And that succeded for this purpose.

 

Regarding small drills and wobble. Have no experience yet, but the test in the video with 0.8mm drill bit was totally uncomplicated, no wobbling, no breaking of the bit. So I am not worried if it will work with 0.6mm - which is the smallest drills I use.
So let's see the next tests after I finished the software part.

I think the Dremels are quite good quality regarding wobbling. I tried a cheap $10 drill a few months ago and it's drill bit holder is a nightmare, wobbling like hell. Possibly one future iteration of this thing is to replace the Dremel flex extension with a decent quality DC motor with drill bit holder. Would make it more compact and it'd be still inexpensive.

Edited July 18, 2018 by conny_g

[francois-longin 5]

Very nice project. I began the same thing on my UM2+ with my own dremel mount.

Do you still use it ?

Do you still use the same method to generate gcode or another way ?

 

[conny_g 251]

On 12/31/2019 at 3:45 PM, francois-longin said:

Very nice project. I began the same thing on my UM2+ with my own dremel mount.

Do you still use it ?

Do you still use the same method to generate gcode or another way ?

 

 

Unfortunately I did not continue with that project back then. Work and other projects so it got stuck there. Generally worked little on electronics projects so that also the need for the solution was less. So I did not notice that the project is stuck 🙂