[Mac.CNC]

My flight got canceled, I have to wait another day to get to Europe..
Having nothing else to do for the evening, I decided to try to cut some christmas ornaments, inspired by Michael over here:
http://www.cnczone.com/forums/momus_...christmas.html

I've always been intrigued by the really miniature tools, so that seemed to be a good idea to try some of that.
The endmill I chose is a 0.063 EM, and I had no idea what I was doing.
Gerry posted something a while ago that gave me a starting point:
http://www.cnczone.com/forums/woodwo..._end_mill.html

Suppose at 18,000rpm for a bit like that, you'll still have to maintain a high feedrate, otherwise it'll just get too hot and break off.
I was actually surprised how well this worked out. It sounded nice and there were no burn marks on the material.
The bit is still intact too. I might just cut a few more of these, lol.

Here's a short video on how that went, using a snowflake pattern:
Momus CNC X2S Milling Snowflake Christmas Ornaments - YouTube
--
Until next year,
Mac

[Mac.CNC]

Hello again and a happy new year to everyone :-)
The vacation was great and I got back with a lot of new energy to work some more on the Momus.

The alps carving was a huge hit with my parents for christmas and it's now hanging of their wall :-)
Some other people saw that and I guess I should make some more to pass out as presents.

I also made A LOT more plywood snowflakes that night before our flight left and passed them out to friends and family overseas.


WP_20131218_002 by mkloberg, on Flickr

Once we got back I got a little bored and make a quick fish out of a basswood slab I had laying around to see how fast the machine could go.
Turned out that anything higher than about 65ipm would make the belt stutter on the x, which will be my first major project for this year (2nd motor on the other side).


WP_20140108_006 by mkloberg, on Flickr

The finish on this thing is brown Briwax by the way, nothing else.

To catch up with projects, I turned to a thing where where a co-worker gave me nine ceramic tiles depicting their family name in letters on them, simply asking to do something with them when I had time, to put on their door. She said something about a Black Forest cake in return, so this is defenitely an important job, lol :-)
This is what I came up with:


WP_20140105_002 by mkloberg, on Flickr


WP_20140108_002 by mkloberg, on Flickr


WP_20140108_005 by mkloberg, on Flickr

She wants it stained to look like oakish, I've not decided what to finish it with, but it turned out pretty good (2 hours later cleaning it up with carving chisels and sandpaper)


WP_20140111_007 by mkloberg, on Flickr

It has pockets in its back to accept the tiles and a 1/8" thin back panel to hold it all in place.


WP_20140111_008 by mkloberg, on Flickr

Note the two keyholes, that was fun g-code program to write by hand (I found the keyhole cutter in an old router set coming from Sears).

The next challenge came from a friend that makes sausages and smoked meats I know from work.
He needed a replacement coupling part for his meat grinder. I asked him to make a drawing and since he's German, it turned out in metric:


Coupling-DWG by mkloberg, on Flickr

After learning a little more on the purpose of the part, I ordered a 1x3x12 block of black Delrin from McMasterCarr for this project.
I figured that should be just about the hardest plastic to stand up to the forces a drive shaft coupling would ever see in a meat grinder.
Here we go:


WP_20140110_001 by mkloberg, on Flickr


WP_20140110_005 by mkloberg, on Flickr


WP_20140110_010 by mkloberg, on Flickr

That wasn't so nice, although I had the dust colletor running at full throttle.
It just wasn't enough to pick up these fairly big Delrin chips - I'm seriously thinking about a bigger unity now (1hp currently).

The part turned out pretty good though and I hope it'll work out for my friend's meat grinder.


WP_20140110_013 by mkloberg, on Flickr


WP_20140111_002 by mkloberg, on Flickr

To recap, the Momus is a great machine in its original design. I made my mods and got it up and running.
The dust collection is working great and there is very little to complain about so far.
What I did see however during all these projects is that it can only take very light cuts because of the fact that the x-axis is only driven from one side.

I think I'm going to put a 2nd slave motor on the x axis to make it work better.
The past few weeks, I've been collecting parts for making that happen. I got the motor, the belt and the pulley already.
In the electronics department, the motor should just plug into the G540's 4th port to be slaved by Mach3...

Since that constitutes a somewhat major teardown to a somewhat running machine, I carefully parked the z-axis on a block to hold the gantry up while trying to slide the entire left x-axis rail assembly out of the machine for modification.


WP_20140113_003 by mkloberg, on Flickr

Then I turned back into Sketchup to make the model and drawings for the mirrored parts to go on the left side of the gantry for this effort:


2014-01-13 20_12_35-Outliner by mkloberg, on Flickr

So far I took the left rail assembly off, got it marked up and started drilling pilot holes to accept the 2nd x-axis stepper motor:


WP_20140113_002 by mkloberg, on Flickr

There's a lot more to do than I thought actually. Putting x rollers on the left side also means that the interference problem between the lower bearings and the angle iron is coming back to haunt the rails now, meaning I have to grind down the entire left angle iron too.
I'll probably get that done tomorrow.
Fun stuff, stay tuned & happy CNC'ing... :-)
Mac

[Mac.CNC]

Hello again :-)
good news. After using the machine quite extensively and experiencing the problems caused by the deflection in the gantry on the x direction, I dediced to go through with the second slave motor for the x axis and started making the parts for this upgrade.
What I was hoping to accomplish by that was to give the gantry twice the power and eliminate the play/backlash that I think was at play when cutting certain toolpaths, resulting in not so perfect results.
This particular machine is much longer than the orignal Momus and I've frequently seen strange vibrations in the x axis travels during programs, which I attribute to the total length of the kevlar belt and its albeit minimal stretch (this loop is about 10 feet long).
To make a long story short, I was out of options to make this work by just adjusting things and decided to take the plunge and do it.
Here's how that went:

I started a usual, cutting and marking the new parts for the left side of the gantry, after having some of that modeled in Sketchup to understand the geometries better on the parts.

WP_20140118_002 by mkloberg, on Flickr

It wasn't all that difficult, as most of the parts are just modified mirror images of what goes on the right side of the gantry:
Part 6: Gantry outer right
Part 4: Gantry right top
Part 5: Gantry spacer
Part 3: Ganrty left top
Part 9: X Belt plate
Part 10: X Belt clamp
Part 33: X Belt pulley adjuster

Making all the new parts was a lot of fun once again and making them look shiny was fun too :-)


WP_20140119_002 by mkloberg, on Flickr

To install the upgrade, I didn't actually take the gantry off, I just parked the z axis on a block and slid the left part of the x-rail assembly out of the machine sideways.


WP_20140113_003 by mkloberg, on Flickr

Having a place to mount the second motor was the task from there, so I modified the left rail base plate to accept the motor and the four mounting screws:


WP_20140116_003 by mkloberg, on Flickr

Looking good so far, here's the gantry stripped off everything that goes on the left side with the new left rail base plate in place.


WP_20140118_001 by mkloberg, on Flickr

The left angle iron had to modified too, first of all it needed the same height grind down as the right one (remember that interference problem?) and then the arragement for the motor had to be made. All that went quite well.


WP_20140116_002 by mkloberg, on Flickr

Having all that done, I tried a dry assembly of the left truck on the gantry:


WP_20140122_003 by mkloberg, on Flickr

And tightening the belt up:

WP_20140122_005 by mkloberg, on Flickr

After that, I spent about two evenings with the T-bar and dial indicators to get everything back the way it was before the upgrade.
You might not believe this, but once two x-axis rails are governing what the gantry does on both ends, the game becomes a lot different and a little more complicated.
That was fun to say the least, and I'm still not sure I got it right. At a certain point (+-10 all the way around, I just gave up and said screw it, let's see how it cuts)

WP_20140122_006 by mkloberg, on Flickr

To test the whole thing, I just loaded the snowflake program from earlier and cut some air.
Interesting was that I couldn't hold the x axis before the ugrade at 100% of the program, here's a video of 200% the original feedrate.
The vibrations in the x are totally gone, hinting that idea may have actually worked to eliminate the stretch of the belt and the vibrations caused by the x axis roller
Tooling around with two x axis motors - YouTube

Can't wait to see what the resulting cuts look like from here.
Happy CNC'ing,
Mac

[Mac.CNC]

Hello again,
just thought you might find this interesting, remember that thing?


WP_20131216_001 by mkloberg, on Flickr

I cut this on the machine before the upgrade for my neighbour, and handed it over in a hurry - unfinished.
Turns out that he decided to silver leaf the piece to duplicate the finish of the Chandelier that was already hanging in the baby's room.
He did a really nice job and invited me over the other day, so I took some pictures.
looks really nice now that it's in place up on the ceiling:


WP_20140126_001 by mkloberg, on Flickr


WP_20140126_002 by mkloberg, on Flickr


WP_20140126_001 by mkloberg, on Flickr

Still working on getting back up and running with the second x-axis motor in place, homing issues, etc... but I'm almost there - stay tuned :-)
--
Mac

[Mac.CNC]

Hello again,
sorry for not posting anything lately, now that the machine is up and running failrly reliably - it's been keeping me busy with projects for co-workers and friends to cut and exploring new methods and technologies.
I found tat the Momus, even oversized as mine is a machine you cam make jewlery and pcb boards with!!! (assuming you have quality rails in place)

I also tried diving into pcb board milling, jewlery making and 3D scanning in the past 4 weeks.
Having embarked into these high precision areas of CNC arts, one of the main problems turned out to be how to zero Z to the top face of the workpiece each time the tool is changed.

Lol, sounds a little overboard, but that's really what it comes down to if you can hold x and y on the machine reasonably (I still can't in the X, but that's another story).

I have to say, that since I installed the slave x-axis motor, all the issues I had before are gone.
I mean, zero, zit, nothing on that axis that could possibly blamed on the Momus design from here.

To auto zero Z:
To a seasoned CNC machinist, this might sound silly and probably a no brainer, but I thought a solution might be if I put a zeroing surface type of UFO in the left front area of the machine, just behind the area where the vertical cutting takes place.
This would then serve as the absolute z-zero point to touch off from for (ever, I guess).

Currently, I've only worked with mostly flat stock (wood). It's either 0.25, 0.5 or 0.75 in thickness. Often it's not exactely that, but always marginally neglectable in variances.
I was looking for a way to say: Type into Mach3 somewhere how thick that thing is and then hit a button to reference the tool I just put in the collet to zero in reference of the top of the surface of that piece of wood by clicking just one button. GO!

I found this awesome touch off pod online that has a spring loaded brass plate and wired it into the machine.


WP_20140227_001 by mkloberg, on Flickr

I haven't done it yet, but I was thinking of installing an emergency stop button type swith inside the spring that holds up the plate in case the touch off goes wrong.
So far I have not found anything suitable to go in there, but that would be a killer feature, to have.

Next to the grounding issues we've talked about before in this forum, the next problem I had next was that I also wanted to use a touch probe to digitize things.
The fundamental difference in these is that the touch probes you can buy are normally-closed designs, where touching off a plate with a tool in the collet is a normally open situation.

Normally open, normally closed, shesh... That signal has to go into Mach3 somehow as a trigger, but just one way.
I already have an idea on how to solve that with a custom curcuit board, involving a transistor to reverse the signal from the nc touch probe.

Stay tuned... :-)
Mac

[Mac.CNC]

Hi guys,
about tooling touch off and a touch probe on the same Mach3 input:
This is going a little bit deeper into the electronics department and probably not relevant to most Momus builders, but I thought it might be a good idea to show you anyways :-)

The problem at hand was that I wanted to be able to touch off chucked tools off a fixed plate that sits somewhere in the machine to determine the current z-offset in reference to the workpiece.
In addition I wanted to be able to use a touch probe to scan things in if I had to.
Earlier, I bought this probe: Desk CNC from IMService

It wasn't a surprise, but I came to realize that this probe consists of a normally closed switch, meaning whenever the probe touches something, the switch inside of it opens.
With Mach3 that is now a problem, because it's probe sensor (can't have 2 as far as I know, without messing around in the pins and ports), requires a high signal to trigger the input.

Being a little bit rusty on the electronics side, I decided to use a simulator to work this out: There's a free program for that kind of stuff you can get called LTspice.
I designed an initial circuit with three transistors in there and tested it, but made a logical error.
It wasn't until I cut the board, soldered everything in there and tested it until the error surfaced, so back to the drawing board.
The circuit that turned out to be working looked like this in LTspice (please ignore the powersources on the left and right, these are just for simulation):


2014-03-14 23_57_41-LTspice IV - [ProbeCircuit by mkloberg, on Flickr

To generate the G-Code for cutting a two sided circuit board, I first moved the schematic into another free software called EaglePCB:


2014-03-15 00_00_27-1 Schematic by mkloberg, on Flickr

After moving some components in the software around, EaglePCB then generated the board layout and using this totally awsome add-on to Eagle/script called PCB-Gcode, I was able to go out to the local Radio Shack, buy a two sided copper clad circuit board for a few bucks and went to work.

To give you a better idea what this thing does: You probably notice the relay :-)
That relay is powered by an output out of Mach3 and denotes the mode that we're touching things.

If the relay is:
ON: Touch plate mode (acts as normally open)
OFF: Touch probe mode (acts as normally closed)
In my case I can now control that output via Mach3 using script command: ActivateSignal(OUTPUT4) or DeActivateSignal(OUTPUT4)
I also made a gcode macro for controlling the relay: M101 to turn it on and M102 to turn it off.

The relay itself, I got from radio shack, it has a pair of ncno contacts.
I used one contact to connect the system GND to earth (because the machine chassis is only connected to earth), thus making it possible to establish a signal when contact is made.
The other contact switches the mode around on the board, connecting the Mach3 probe input to the relevant receiving end of the the input (probe or plate).



Looking at this thing, I thought, the first thing that needs to happen is a jig, to reliably carpet tape the pcb blank into a cradle of some kind to hold it down, so I made this:


WP_20140223_002 by mkloberg, on Flickr

There are flutes underneath so that we can grab the blank with an alligator clip for touch off, reliably. At this point, the clip is connected directly to the probe input of the G540.
This helps a lot, because now I could touch off the pcb directly, by grounding the spindle.

This is the touch off pod, fully installed and wired:


WP_20140223_004 by mkloberg, on Flickr

The first pcb routing attempt, pleas ignore - there was a logical error in it.


WP_20140224_005 by mkloberg, on Flickr


WP_20140224_014 by mkloberg, on Flickr

This is the probe, by the way:


WP_20140225_001 by mkloberg, on Flickr

After learning some more, I cut the final version of the board:

WP_20140226_003 by mkloberg, on Flickr

using these tools (grab bag set from harbour freight):

WP_20140226_008 by mkloberg, on Flickr

Much simpler now, I made a tool change mistake though (I'm sure you can see it, lol):


WP_20140226_013 by mkloberg, on Flickr

After soldering the components in, here it is: CNC Touch Probe Driver Board Rev.02

WP_20140226_014 by mkloberg, on Flickr

I installed it in the machine near the plate for the connector hole of the probe (sorry for the bad picture):

WP_20140302_003 by mkloberg, on Flickr

I did have to modify some things in Mach3 (button scripts and screenset) to really make it do what I wanted, but now I'm quite happy with it all.
I can now more or less automattically touch off anything I want using a variety of probes without having to change anything the configuration of Mach3 - by just plugging in the particular probe:


WP_20140302_004 by mkloberg, on Flickr

As you can imagine, the learning curve was quite steep to make this happen, between EaglePCB, LTSpice and how to mount the board, but it was a lot of fun doing it :-)
If you end up doing something along these lines, please let me know and I'll be happy to answer any questions you may have...
--
Mac

[ger21]

It's a little late, but there is a method to use both an active low and active high probe in Mach3 at the same time, with no additional electronics.

Assign one of them to the Probe input.
Assign the other to an unused Input, say Input 3 for example.
Create a Brain in Mach3 that triggers the Probe input when Input 3 is triggered.

The only issue with this method, is that there is probably a very slight delay for the Probe input to be triggered. But I suspect that it's very small (microseconds?) and the resulting positional error is probably less than .001" at slow probing rates.

[Mac.CNC]

Hi Gerry,
thanks, I wish I'd known this, that's pretty slick. Well, what I have now works really well and there's nearly no delay / positional errors.
I did a series of tests with a 60deg v-bit off the fixed plate and found the repeatability to be in the range of +-0.0002", which was surprisingly better than what I would have thought it might be.
Also, I really didn't want machine ground (earth) permanently connected to the GND potential of the power supplies, the relay method only makes that connection during touch off cycles.
Good to know though, that there's another way to do it. :-)
--
Mac

[DY123]

Where did you find the Z height touch off pod? Are those three legs removable? I need something low profile.

[Mac.CNC]

That one came from CncRouterParts:
Auto Z Touch Plate | CNCRouterParts

It doesn't come with the legs, these are just aluminum standoffs that I added to raise it up a little bit.
--
Mac

[ger21]

It's a little late, but there is a method to use both an active low and active high probe in Mach3 at the same time, with no additional electronics.

Assign one of them to the Probe input.
Assign the other to an unused Input, say Input 3 for example.
Create a Brain in Mach3 that triggers the Probe input when Input 3 is triggered.

After telling someone else about this method, it doesn't seem to beworking. Although I thought it was working fine when I originally wrote the Brain.
More testing is needed.

[Mac.CNC]

Hi Gerry,
thanks for the followup - sorry to hear that it didn't work as expected using the brain mod method, I'm still interested to see how that works out.
For disclosure, here's the button script I'm runing currently, to get to the +-0.0002 repeatability. Not entirely my works, I built it on top of an existing masterpiece.

There are a bunch of new DRO's I added to the settings screen, importantly one for the input of the distance between the plate and the machine bed (DRO#1180).
The other one is to give the user a way to input the workpiece thickness (DRO#1181).

Begin Dialog TouchOffDialog 180,97,"Auto Tool Zero"
PushButton 12,8,72,60, "Touch-Off Plate"
PushButton 96,8,72,60, "Touch-Off Part"
CancelButton 12,75,156,14
End Dialog

Dim Button
Dim Dlg1 As TouchOffDialog
Button = Dialog (Dlg1)

Dim abort
Dim jogDownHeight

CurrentFeed = GetOemDRO(818) 'Get the current feedrate to return to later
CurrentAbsInc = GetOemLED(48) 'Get the current G90/G91 state
CurrentGmode = GetOemDRO(819) 'Get the current G0/G1 state
PlateThickness = GetUserDRO(1151) 'Z-plate thickness DRO
ToolChangeXpos = GetUserDRO(1200) 'Tool change y position
ToolChangeYpos = GetUserDRO(1201) 'Tool change y position
ToolChangeZpos = GetUserDRO(1202) 'Tool change z position
PlateToBedOffset = GetUserDRO(1180) 'Height differential between fixed plate and spoilboard surface
PartThickness = GetUserDRO(1181) 'Thickness of the workpiece
TouchPlateThickness = GetUserDRO(1182) 'Thickness of the plug-in touch plate

If Button <> 0 Then
'Turn the touch off plate relay on
ActivateSignal(OUTPUT4)
'Pause and wait for the relay to come on
Sleep(500)

'Check to see if the touch-off sensor is already grounded or faulty
If GetOemLed(825)=0 Then
If Button = 1 Then
'Bring the z-axis up all the way
Code "G53 G0 Z-0.03"
While IsMoving()
Wend

'Zero the Z axis so any moves will start from here
Call SetOEMDRO(802,0)
Sleep(250)

'Move right over the touch off plate (in machine coordinates)
Code "G53 G0 X-0.4350 Y-0.0675"
While IsMoving()
Wend

'We don't know how far the cutting tool extends, so do a probing jog move down to above the plate.
jogDownHeight = -4.5
Code "G90 G31Z" & jogDownHeight & " F100"
While IsMoving() 'wait while it happens
Wend

'Check if we touched anything yet (which shouldn't happen normally)
zProbePos = GetVar(2002) 'Get the point we might have touched already
If zProbePos <> jogDownHeight Then
MsgBox "Error: Touch-Off contact occured during rapid down move. Please check and try again, or adjust the tool start length."
abort = 1
End If
End If

'Touch-Off
If abort = 0 Then
DoOEMButton (1010) 'zero the Z axis so the probe move will again start from here
Code "G90 G31Z-6.25 F4" 'probing move, can set the feed rate here as well as how far to move
While IsMoving() 'wait while it happens
Wend

zProbePos = GetVar(2002) 'get the axact point the probe was hit
Code "G1 Z" &zProbePos + 0.02 &" F4"
Code "G90 G31Z-6.25 F0.2" 'probing move, can set the feed rate here as well as how far to move
While IsMoving() 'wait while it happens
Wend

zProbePos = GetVar(2002) 'get the axact point the probe was hit
Code "G0 Z" &zProbePos 'go back to that point, always a very small amount of overrun
While IsMoving ()
Wend
If Button = 1 Then
Call SetDro (2, -1 * (PlateToBedOffset + PartThickness)) 'set the Z axis DRO
Sleep(250)
End If
If Button = 2 Then
Call SetDro (2, PlateThickness) 'set the Z axis DRO
Sleep(250)
End If

'Move z up back to home and turn the touch off plate relay off
DeActivateSignal(OUTPUT4)
Code "G53 G0 Z-0.03"
While IsMoving ()
Wend

Code "(Z axis is now zeroed)" 'puts this message in the status bar
Code "F" &CurrentFeed 'Returns to prior feed rate
Else
'There was some kind of problem, bring the z-axis up all the way
Code "G53 G0 Z0"
While IsMoving()
Wend

'Turn the touch off plate relay off
DeActivateSignal(OUTPUT4)
End If
Else
Code "(Touch-off sensor has already made contact, check connection and try again.)"
End If

If CurrentAbsInc = 0 Then 'if G91 was in effect before then return to it
Code "G91"
End If

If CurrentGMode = 0 Then 'if G0 was in effect before then return to it
Code "G0"
End If

If Button = 1 Then
Code "G53 G0 X" & ToolChangeXpos & " Y" & ToolChangeYpos
While IsMoving()
Wend
End If
End If

--
Mac

[Mac.CNC]

Hello again,
it's been a while, and a lot has happened...
The machine is performing great and I've made a lot of things on it so far, it's been running pretty good almost every evening, making something.



I also made a tool carousel style tool holder that I really like by now, it was just a bunch of 3/4 plywood disks with holes in it and a 1/8th center hold that I stuck a long dowel through. Each level rotates on the dowel and it's been a real helper so far.
All my collets and mills sit in this carousel now and I made it so that the dowel sticks out the bottom by a little and hooks into one of the dog holes in the workbench, so that it can't fly off the table when someone brushes it by walking by. That'll be a rather expensive accident if it ends up on the floor... ;-)


The office desk sign for my GF's mom got some more love and I finally finished that with the right lettering on it:


That lead to an attempt to get into jewlery making (gf's mom suggested to carve the high schools fb team logo into thin wood slices).
I looked around and found these Tagua slices on Amazon. It's a vegetable nut that is as hard as Ivory and you can get these slices all over the place.
Problem is, as I came to find out - the slices are usually not flat...
That these slices totally unfit for V-carving.

Having about 50 of these on my workbench, I resisted to give up and plugged the touch probe in.
I made a raster plate and carpet taped a handfull of these slices to the plate and scanned them.
I the carving software that I'm using, I used that scan as a base plane and overlayed the v-cvare logo onto the scanned surface.
It took a while but the result speaks for itself:



While going through all this, I found myself grabbing a flashlight all the time, to peek through the top plate of the dust shoe to see what the machine was doing.
With the brush skirt all around even at optimal lighting conditions, there was no way to see what the machine was cutting without shining a flash light in there.
The other issue was that the mounting plate of my dust collection system had the slot that clamps it to the spindle on the wrong side for visibility in videos, etc.
I've been mulling over for a while now and finally decided to take the plunge to fix it.
For me, that meant to flip the plate and incorporate a bright light source into it.

That's been in my head for a while now, I just wasn't sure what to use for lights - until I found this thing:
T10 21-LED White Light Car Angel Eye (70mm Diameter) - Free Shipping - DealExtreme

It's basically a PCB with a lot of bright LED's on it, originally to go around car headlights in the car tuning community.
The 70mm diameter lended itself for my spindle, because fits pretty good to go around the hole in the dust shoe.

Said, done - I designed a cut a new top plate for the dust collection shoe to incorporate the LED ring and a matching power plug to come out the top.
Turned out that the dx.com site is chinese and I didn't get the LED ring I ordered until three weeks later. The little envelope had a Chinese postage stamp on it.
Anyways, they delivered in the end and that's what the ring looked like (12V):


I incorporated a plug and socket, drawing 12V power from the Z-axis limit switch that was already there.
All it took was a thin 2 wire cable coming down from the junktion where the Z-axis switch was, follwing the motor cable to meet the collector plate.
Here's the acrylic stock after cutting the idea:


The LED ring had resistors on the underside, so I had to make some pockets to accomodate these:


Plate with the ring and the plug installed:


I ended up supergluing the components into the plate, here's the whole thing from the top side:


Installed it back into the machine, view from underneath:


Bright! no doubt






At that point, I decided to cut a second dust collector skirt plate for a 3" brush since I had the materials already.
...and to check out the lights during cutting of course...




Almost spooky, having all that light moving around:


Well, right after that toolpath I heard a loud "CRACK" while moving the gantry into the back to take the part out.
I moved things back and forth a few times and finally found out what happened:

The back y bearing on the left upper side of the gantry cracked.


Grrr, that's a little frustrating, things were going so good - and now this, mehhh...

Right now, I have no idea on how to replace the bearing without taking the carriage off the gantry - meaning the gantry has to come out of the machine.
Dang, things were going pretty good so far, but this one is probably the worst possibly conceivable failure to deal with in this design (any other bearing you could probably replace in place).
I thought about drilling a maintenance hole through the tube, the size of the bolt head - somewhere , to drop the bold through it out the bottom...

Anyone got any better ideas?
--
Mac

[Hawkeye67]

Hello Mac

Is it not possible to run the gantry partly out over the end, if you dismantle the timing belt and then drop down the bolt?
I enjoy to follow your thread here and all your good ideas and great pictures, which are good inspiration.

best regards
Mogens

[jim_r53]

Hi Mac,

I noticed that you have installed some filters over the air ports for your electronics but I can't find any info in your thread about them.

The light ring is a great find!

[Mac.CNC]

Well, getting that one repaired was interesting.j
I thought about this for a long time this morning, considering what it would take in terms of realigning everything to at least 100th of an inch.
I didn't really feel like going trough all of the steps again to adjust every bearing on the gantry and came up with a shortcut.
The problem at hand was just to provide enough clearance to get the back y axis bearing bolt out of there, remove the old bearing, slide a new bearing in and move on.

I'm not even sure why I'm telling you this, but perhaps - if something like that happens to your Momus you might find this process usefull.
It took just about 45 minutes to replace the bearing, I didn't disconnect any of the belts and there was no realignment steps to do afterwards.
Lol, sounds like magic I would have thought last night, but it wasn't.
Here's what I did:

- I parked the gantry to left of the y axis and put some blocks underneath, to give the z axis some support.
- Next I marked the areas where the y-axis rails meet the x-axis trucks with blue marking paint and scribed some lines.



- This was to get the rail into the same orientation and height as it was before what comes next.
I started to replace the bolts that hold the y-axis rail to the tube with longer bolts, carefully removing the short bolts out the back.
I have to admit I was very concerned with that at first, but after dropping the first one inside the tube (ooohh, noo, it'll be ratteling in there), I found a way to extract lost parts from inside the tube relatively easy.
A strong magnet! Remember the tube is aluminum, anyting metal in it will be attracted.
The one I have is a rare earth type and basically attracts anyting up that is inside the tube, so that worked out really well.

After that, I was able to move the gantry back a bit, without the y-rail and the carriage sitting on it - suspending that with big wooden wedge - just enough the get that dreaded bolt out.



I removed the bearing, put a new one in, and followed these steps in reverse, jacking up the gantry to where I had the lines scribed earlier.


After some tests, it wasn't really where it needed to be in terms of z-axis height in the 100th's, but at least the problem component was replaced:



Yay!

@Mogens: Good idea, but I have all these electrical connections coming off the gantry, into the carriage in terms of IGUS energy tracks on both the y and x.
The original problem was too to get to the bolt on the y, so sliding the gantry off, would not have done the trick yet.

@Jim
The vent grates look pretty cool, don't they? lolz.. :-) Sorry, I'm I little bit proud of these grates myself and had others around the shop ask about that too.
I did never mention them in this thread because I thought it wasn't that big of a deal to replace painters taped cheese cloth intake ports that have been seen in various videos and pictures with something decent.I'm glad you ask though:
The inspiration for these grills came from the old star ship Enterpise tv series.
They had this divider in engine room where Kirk often talked to Scottie, resembling a honey comb arrangement.
I took that idea and made covers for the intake holes for my machine using the same pattern.
What it was in the end were two pieces on each side: The grate and an outer frame to go underneath it.
Inside that stack, I placed a piece of hvac filter material, cutout with cisors to match the envelope.
Long story short, it's mainly decorative more than useful - I just needed soemthing to contain the filter fabric in front of the ports, instead of the cheesecloth and painters tape.

I'd be happy to send to you the stls for these if interested :-)
--
Happy CnC'ing,
Mac

[Hawkeye67]

Good to see you have solved the problem. I did misunderstand what bearing it was. I was looking on the bearing on the left side for the X-axis - Sorry :-)
What is it that crack the bearings? Is it to much tension or to big load?

best regards
Mogens

[the_tool_man]

...What is it that crack the bearings? Is it to much tension or to big load?...

This has come up before. These bearings aren't really meant to be used as rollers. They are meant to be mounted inside another piece, which would distribute any radial loads over a large area of the outer race. As they are being used, the concentrated rolling contact load is enough to fatigue crack the outer race. This is an inherent consequence of using these bearings this way. I'm not being critical of the design, or of Mac's build. That's just a result of using a low-cost solution.

Mac, I suspect that you'll start seeing this type of failure on the remaining bearings. This one was only the first to go. It would be interesting to explore methods for replacing each bearing individually, to see if the alignment settings can be held. If you end up disassembling the gantry or carriage, you might consider replacing all of the hard-to-reach ones as a preventative measure.

Also, I love the light ring! Good work!

[Mac.CNC]

Hello again,
thanks again, tool_man for explaining the issue. This is actually my 2nd cracked bearing now...
I have to say that I keep them a little tighter than recommended, and they are running on hardened precision ground steel rails on all three axis.
On the rails, over time - I can see a difference in surface property (shinyness) on all the rails, meaning that the bearings are indenting themselfes into the rails.
I could imagine, that over time - this taxes them a little more as opposed to be running on cold rolled steel - which is a lot softer.
At this point, I'm still super happy with the performance of the machine, don;t mind a cracked bearing every other month and will take your advice, swap as many bearings when there's a "crack" as I can when there is an opportunity.

I also have to say I almost think I'm making the machine do something that it probably wasn't designed for.
Pretty much every night for the past few weeks, it's been 3D surfacing and profiling 3/4 thick parts in poplar, to make "something", that is usually 12 to 29 inches long.
I guess so far I got away with taxing the machine to such extremes, and I'm going to do so until something major breaks and fix it again, lol...

For the numbers, I'm using a 0.25" upcut endmill at 60ipm/9600rpm, half diameter pass depth (0.125) to cut large things out of 3/4 poplar or oak.
Over time, that seemed to be the what the machine "likes" at this point. Cranking up the ipm by 10 or 20, I found results in bad surfaces most of the time especially when going against the grain of the wood. No sure, what to do to make it more efficient.
--
Mac

[the_tool_man]

...I also have to say I almost think I'm making the machine do something that it probably wasn't designed for.
Pretty much every night for the past few weeks, it's been 3D surfacing and profiling 3/4 thick parts in poplar, to make "something", that is usually 12 to 29 inches long.
I guess so far I got away with taxing the machine to such extremes, and I'm going to do so until something major breaks and fix it again, lol...

I, for one, am glad you're exploring the limits of this design. Quite frankly, your experience should give others confidence that this is a viable machine that can be extended a bit to grow with the user. It is relatively inexpensive to build, fairly simple to maintain, and easy to repair if something breaks. Good work!