Alu-profile-based enclosed 250x150x100 desktop CNC

[Jacek]

As a follow-up to my former post about CNC design:
https://cnczone.nl/viewtopic.php?p=242922#p242922
I decided to make a smaller and chaper DIY machine. This time it went out from CAD :-)

Initial goals were to have a small, desktop size, fully enclosed one, that can be also built and used at home, powerful enough to cut alu. Structural frame and enclosure is based on alu profiles from ITEM. Working volume of 250x150x100 I managed to fit within the enclosure of 400x360x525, including all electronics. I used steppers integrated with trapezoidal screws (1,68A, Tr8x2). 10mm linear rails fully supported. SourceRabbit main board with USB. Daedalus spindle, 12kRPM, 0,5kW - one of a few cheap ones available that has a taper directly in motor shaft. Hall sensors on all axes in both directions. Electronics cabinet from alu sheets 1,5mm cut to size at Hornbach, glued together. Milled parts ordered from Poland, some parts 3d-printed, e.g. backlash tuning. The machine runs 4,5m/min after calibration. I spent half a year of evenings to design and build it. It had cost about 3kEUR total. The biggest struggle was with 10mm linear rails supported. Most shops sell crap that cannot be even assembled and is out of spec. wrt. their own drawings.

Video:
https://www.facebook.com/10000442884953 ... 9725840242

1.mp4

How do you like it? I will be in Wijk en Aalburg this weekend with the machine. Hope somebody can help me with some feeds & speeds for alu.

2.jpg

3.jpg

4.jpg

5.jpg

6.jpg

[kevertje]

Nice machine! Looking forward to see it in action Saturday.

[Kars-cnc]

Very neat concept, this compact and fully closed machine. Suitable as "Desktop CNC" to be operated in an office environment I guess.

Seen the color of the windows you anticipate the use of a laser as well?

[Jacek]

With orange looks more attractive vs. transparent.

[serum]

Nice work! I thought the same about the orange and the laser. Aesthetics are an important weigh factor. Do you bring a laptop too? Speeds and feeds in aluminum guidance; there is no clear guideline for that, only ideal parameters according to the diameter of the bit, feed per tooth and SFM. But these values are hugely depending on the machine. Spindle RPM, spindle power, stability.. In other words, you have to interpret strongly. But the good thing is, if you are an observant/analytical type of person (judging by the OCD that went into tidying this machine together the way you did, you'll probably get the hang of it soon enough). Aluminium is very forgiving material and can be machined with parameters way from their optimum values.

That being said; i would start slowly with a low chipload (0,01) just make sure that when a two flute cutter that a 0,01 chipload with a runout on the spindle (which is more of a a certainty than possibility) it would result in a 0,02mm chipload. (i thought these spindles use a bolt-on-motor clamp for the collet)

Just start machining, take it easy and go up from there. Just buy a hand full of cutters on Alie (2 flutes, single flutes) and get to know your machine.

Do you want to machine some aluminium on location? Do you have Fusion360 installed on your laptop or so? I can bring you some scraps of aluminium and i thought i have some smaller cutters laying here. What size of collets do you have?

Looking at the construction, i think the machine could benefit from aluminium screwed on panels on the left, right, rear and bottom. It will increase the structural strength/overall stiffness.

[Jacek]

I am going to have the laptop with Fusion, some small alu plates, plexi and wood, some flat-end endmills, chinesium, and some better ones as well + ER collets up to 7 mm.

I have some good online calculators that take into account the spindle power, but still they guide me to go too fast.

The 500W Daedalus spindle is the only one in their collection that has the taper directly in motor shaft. It was disassembled by somebody online and it was clearly visible, that's why I chose this one.

Same scrape alu would be helpful on site. Seems like we will be playing a bit :-)

[Jacek]

And some pictures and vidoes how it performs. These are first cuts, taken at the CNCzone.NL Day in Wijk en Aalburg.

[Kars-cnc]

Nice results Jacek! It was good to see your machine in reality yesterday, and I fully agree with Serum that it is an inspiring concept.

[serum]

What's that scuff in the left lower screw? You mean you are able to machine stainless steel with that machine?!

By coincidence, I talked about this with Kars yesterday, these little accidents always happen in a rush.. (Bob Ross calls them 'happy misstakes where painting is probably more forgiving than machining at the wrong place in space).. When visiting Kars a few meeting before, i tried ruining a brick ornament in Kars his garden with the back of our car) Not my best attempt. Ornament took it well though..

You're a quick learner, (good observer/listener) I'm confident that you work out another machine sooner or later with interesting features and tweaks.

But a Z offset probe, even the most rudimentary version will be more convenient than zeroing with paper/rolling a cutter. (or make your own with aluminium and some 3d printed isolator) The controller has an input for that.

[hfjbuis]

I checked the width of 2 cuts and they where within 0.02 mm. Pretty repeatable and better than I expected.
Jacek checked the length of a cut and that was (as far as I remember) within 0.02 mm.
I was impressed.

[Jacek]

UPDATES:

In the recent months I disassembled half of the machine and assembled again after "fine tuning" the Chinese supported shafts and carraiges. "Fine tuning" required a lot of filing and hammering. In the end I was able to adjust the "pre-load" on the carriages and make the machine much more stiff. I replaces several small parts and the back cover of the machine, designed and 3d-printed some dial gauge holders that are helping me with "squaring and tramming". And the machine is now back to operation.

Today I did the surfacing of the whole table. The flatness runout is within 0,02 mm now. Next step is to "square the spindle". Left/right it's OK, but it is leaning forward a bit, hence not ideal surface finish of the table.

And here it is a fresh question to those gurus of feeds and speeds. I was surfacing with 3mm, 1-flute flat carbide endmill, @12kRPM, AP 0,5mm, AE 2mm. From any F&S calculator I am using the recommended feedrate shall be ~240 mm/min. I tried with this value just to realise shitload of chatter and irregular noises. Started increasing feed incrementally to find a sweet spot at 600 mm/min!! What is my worry? With 240 mm/min the chipload is within commonly recommneded ranges, but with F600, the chipload is 0,05 mm, which seems VERY high for alu and 3mm endmill. Shall I expect problems from this?

Second topic is: available power. My spindle max is 12000 RPM and 500W. All those cuts that i did during the event and later on, according to calculators, shall not consume more than 0,1kW of power. However, I was able to stall the machine a couple of times, e.g. with a 6mm endmill, 2-flute, AE6, AP1, F1000. The question is: If there is quite a huge power margin available, is it possible that machine stalls just because of chip load is too high? As per intuition, the spindle shall utilise the powet margin and break/cut the chip instead of stall, right? This is confusing me a bit. Certainly i need to play a bit more with alu. Important learning: for desktop CNC machines most of the cuts do not require high spindle power, but to be able to go faster, max RPM is what really matters!!

Thanks for your support.

flatness.mp4

F600.mp4

[hfjbuis]

At 12000 RPM, your bearings and fan probably consume most of the power. You can measure the amps to the motor (milling air) to get an idea of how much energie is lost in the motor itself.

Could it be you reduced the chatter by putting an extra load on the end mill?

[Jacek]

This weekend I spent some time pushing my little machine to the limits. I wanted to find the boundary of it's capability in terms of Material Removal Rate. I ran a set of experiments with 6mm endmill: milling adaptive pockets with a fixed (quite low) chip-load and varying depths from 1 to 6 mm. I found 4 mm depth to be best. Then I milled multiple 4mm pockets with varying feed and I found a "sweet spot". I am much more than happy because this is what I achieved:

2-flute 6mm flat VHM endmill @12kRPM
Ap=4mm, Ae=1mm
F=1200mm/min
That equals to Material Removal Rate of 4,8cm^3/min.

Some cuts had 6cm^3/min MRR, but were unstable on sound. I guess I could push it even further with 7mm endmill... It also went through the material at F 1300 and F 1400, but with unpleasant vibration. All trials were running with a big "dent" in one of the endmill flutes - hence the surface finish is crap.

How do you like the power of this little beast now? Video attached and linked.

https://fb.watch/imtahJrdnk/

[hfjbuis]

Using a 6mm end mill and moving aluminium that fast is a really good achievement.
The finish could and should be better but if you run it at finishing conditions, I am sure, the finish will be better.
The next question is how repeatable the pockets are made.

Overall a good result.