Hi Juan - Nice but it will need a cross brace or two. These machines do have some inertia and will move a table around.... Until you fill up the bottom with lots of heavy stuff... Peter
I've glued some of the pieces together, but I've run out of wood glue and so did the hardware store apparently. I've also tested the power supply and the stepper motors and drivers which have recently arrived and everything works as expected.
Last thing I've done is aplying epoxy to the surfaces the rails will mount to. Attached is a photo of the saddle, I'll do the others next. The wood absorbs the epoxy nicely, I think it will be fine. There will be some epoxy left to coat other parts of the frame as well.
On the Z axis brake, I initially thought of using a resistor across the stepper windings, however that would only make the spindle arm fall slower, it won't stop it. Instead I'll use a small solenoid whose plunger goes into a disc with some detents attached to the ballscrew. When not activated, the plunger will fall into a detent, and when energized it will let it rotate freely. The solenoid will be activated by the stepper enable signal. It won't maintain the exact position, yet that can´t be trusted anyways if motors have been disabled. Way cheaper than a commercial stepper brake.
Due to academic reasons (back to work!) any further development will slow down to a crawl until summer. Perfect times for letting glue/paint/resin dry .
Juan
Hi there,
I’m happy to report that I’m still alive and made it through allegedly the hardest semester in my degree (we’ll see). A little too busy with thermodynamics, fluid mechanics and what not to work in this project, which fortunately I can do now.
Since my last post I hadn’t done much more than pouring the epoxy on the surfaces where rails mount to, and unpacking the rails and ballscrews that arrived in April.
Those, as expected, are not the best quality. Even though they came from the same supplier, markings differ slightly, some ballscrews have centres and other don’t, and some cars run smother than others. Yet I think they should be sufficient for the project at hand.
I’ve spent the last three or so days drilling and tapping, which I found rather “relaxing”. The drill press has already paid for itself. And tapping wood is a breeze. I’ve attached some of the highlights below.
I also 3D printed a holder for the router, that clamps it to the arm by means of threaded rod which extends to the back of the arm. The ball nut is also secured by threaded rod for which I made some counterbores to fit the nuts so that they don’t interfere with the router holder.
Attached is also a mock-up of the complete machine. I have used just enough screws so that it does not fall apart under its own weight.
Next steps will be taking it apart, gluing the column to the base and then painting the whole thing, so it can be assembled again. I’ll make sure to use wide washers on the arm mounting screws to that the wood does not deform.
Another thing I have to figure out is how to mount the electronics, which will likely go in the back, and what I’m going to do with the table.
The screws holding the table in place are a bit too long, I’ll have to cut them. And I was thinking of counterboring the holes so that they stay below the surface, so I can clamp another piece of wood on top to serve as a work surface. Maybe the wood would be too thin after counterboring, I don’t know. I thing it should be fine.
Thanks for reading this long post.
Juan
Hi Juan - Looking good. Maybe countersink the holes vs counterbore. If counterbored size it to suit large washers... Peter
I counterbored the holes so as to have a bit of wiggle room for adjustment. Used washer-head screws.
------------------------
After a couple of days playing hide and seek with the pouring rain we’ve had over here, I managed to prime and paint all the pieces. I gave it two coats of polyurethane paint to protect the wood from moisture and such. It turned out fine or so I think.
When it dried, I assembled the machine, dropping some PVA glue where the screws thread into the wood. I indicated the rails on each axis parallel to each other to within 0,05 mm. Given that the frame is made out of wood, I expect them to move around slightly with time. Perpendicularity between axis is another subject…
What gave me most trouble were the couplings between the ballscrews and the steppers. The stepper’s shaft is 5 mm in diameter, while the bore in the couplings is 6,35 mm or 1/4” (don’t know why, the rest of the ballscrew’s dimensions are metric)
After trying some bent sheet metal and several pieces of tubing, I ended up machining some sleeves on the lathe and then cutting a slot through them.
I moved the axis around manually and made the first chips cutting an L shape. Still have to adjust the stepper current, and calibrate steps/mm, feeds, install limit switches, configure the homing cycle…
And it is apparent that I’ll need some covers for the rails, and some dust collection.
Juan
Well done Juan - some video and sound would be good when you get there. You've executed this very well. Now its time to cut the pudding... Peter
1/4" is a common standard for NEMA23 stepper shafts.
Looks like you are using NEMA17 motors. High chance they will be underpowered.
Thanks! Don’t worry, there will be video, hopefully this week. I have to do some reading on feeds and speeds, and get up to speed in CAM.
I see. Makes sense.
We went through torque calculations and motor selection earlier in the thread, and these motors seemed appropriate. Running the machine, they have quite some force, as per my calibrated hand oposing the movement . Time will tell, it wouldn’t be too hard to replace them by NEMA 23s
Juan
Here is a short video of the very first chips made by the machine under computer control. For a first test, feeds and speeds are basically “made up”, will have to do some more reading and experimentation, I’m sure I could go quite a bit faster. I’m using Fusion 360 for the CAM.
As seen on the video, I added some rubber pieces to cover up the rails and ballscrews. Added limit switches on all axis as well. The emergency stop button has come in handy .
I also cut out a couple of gears out of some scrap plywood which turned out great. Couldn’t film that because my phone ran out of storage. No idea where all the Gb went…
It is apparent that it needs an enclosure. Rather not surprisingly, it is loud and it makes a mess. And I would feel safer with 4 mm of polycarbonate between the machine and me. Will be working on that.
Juan
Hi Juan - Well done and excellent journey. Peter
Thanks. Still a long way to go. Thak you for all your help.
Juan
Hi there,
I’ve began the construction of the containment vessel, aka the enclosure, out of MDF. This should keep dust out of the floor and broken endmills out of my face , hopefully. There will be a cut-out in the doors for a polycarbonate window. I’ll have to put some lights inside as well. The thing can be disassembled by means of some wing nuts for easier access.
I’ve also ordered some endmills and PCB drills. I ordered a 6 mm to 3.175 mm (1/8”) adapter and an ER-11 adapter as well. In the meantime, just for kicks I made myself and adapter on the lathe, drilling for 3 mm and then finishing with a 3.2 mm bit. Slits were cut with a hacksaw.
In the attached photo, from left to right the ER-11 adapter, the commercial 6 mm to 1/8” and the homemade 6 mm to 1/8”.
Playing around with the dial test indicator, here are the numbers:
Runout
Stock Makita collet with 6 mm pin 10-20 micron (0.0006”)
1/8” endmill on my homemade adapter 40 micron (0.0016”)
1/8” endmill on commercial adapter 60 micron (0.00236”)
4 mm endmill on ER-11 adapter 100 micron (0.0039”)
ER-11 taper 100 micron (0.0039”)
Surprisingly enough, the adapter I made out of “mystery” steel on my mini-lathe is the best one in terms of runout.
I’m no spindleologist, but I think that 100 micron runout is maybe too much. I can almost even see the endmill moving side to side. And to me its clear that the error is in the ER-11 adapter, as the taper seems to be offset 50 microns from the center of the shaft.
It occurred to me that I could maybe hold a 6 mm pin on the ER-11 collet, and hold the adapter upside down in the router spindle. Then I could somehow grind the shank true to the taper.
Just thinking out loud,
Juan
Scratch that last part. I was able to get 20 um runout on an end mill mounted on the ER -11. That seems more useable. I'll make sure I clean everything properly and measure again. Don't think I'd improve anything by modifying the existing adapter.
On another note, applying 50 N right on the tool with a baggage scale, I measured a deflection of 170 um on the y axis and 700 um on the x axis.
Juan
Hi Juan - as expected the stiffness is low but next machine will be better. Y 50N/170um = 0.29 N/um there are some small steel machines that get 1N/um whereas a modern VMC is about 150N/um... Peter
There will surely be a next machine, but I’ve got to have a lot of fun with this one first!
I may be able to improve rigidity if I bolt some steel squares on the sides or similar. We’ll see.
On the enclosure, I cut out the door window, mounted the electronics and tidied up the wiring. It is sort of “temporary”, but will do for now. I added some lights as well as it gets pretty dark inside.
The enclosure reduces a bit the noise and definitely helps with the dust. The thing makes a mess and I’m wearing a mask whenever I open the doors to clean up.
So far, I’ve machined a couple of pieces, an ER-11 collet holder out of plywood and a button support out of MDF, this last one in two setups and with three different tools. I have setup hold and resume buttons, and configured grbl so that when hold is pushed it not only stops moving, but also raises the z axis and stops the spindle, which comes in handy.
I can’t reach the computer and look at the machine at the same time, so I ordered a numeric keypad to use as a pendant with Universal Gcode Sender.
For CAM I’m using Fusion 360, but will try out the Path Workbench in FreeCAD so I have something to fall back to.
Juan
Hi Juan - Congrats - I'm sure there are many grand machines in you yet to come. Well done. "Improving" machines with add ons rarely is a good solution. The machines DNA has been set. Better off learning as much as possible then planning the next better machine. Machines make machines Peter