My MPCNC made in China

jeffeb3 · July 10, 2017, 6:22am

Have you used this?

Then, shim with tape.

So you think the widening of the path is caused by the gantry moving around? The outside path looks pretty good (and I think the total part ended up looking nice, but I get that it should be better). Is it possible there’s some kind of coating on the material that is making the surface harder? Are you sure the spindle isn’t wobbling w.r.t. the gantry?

What about making the first pass smaller, like 0.1mm? AFAIK, there isn’t a setting like that in estlcam, but you could add 0.4mm to the total depth, and just set your origin 0.4mm higher.

vicious1 · July 10, 2017, 8:08am

This is backwards, you are now fighting gravity and the direction of pull from the bit. The spring should add to this not subtract. I understand your thinking in this but I promise it should be pushing down. This could explain your first layer issues, you are preloading it and it can move that .5mm (which is pretty bad, you should replace your nut).

Don’t forget guys, there is 18" of z axis hanging out the top and a large vacuum system as well. Also this is on large linear bearings not the 608’s So I am not sure if there is any preload on them or if the initial material contact is the preload. Heavily modified, I want to help but there are tons of variables here.

forcerouge · July 10, 2017, 5:41pm

Yeah, this was just a quick and easy thing I’ve tried, I will replace the nut system by using two nuts, as I did it before.

Nope, my Z axis is a short as technically possible, I only have something like 6 or 7 cm of travel, the router is as close and as centered as possible to the gantry and I removed the vacuum system for the aluminum milling tests. I’m pretty sure I followed the best practices here.

[attachment file=37946]

But it is indeed possible that the linear bearings slide too well on the tubes, thus not dampering the vibrations, that’s a good point.

Anyway, I tried again yesterday, using a different approach in Estlcam (rough pass and the a finish pass). It worked ok until it reached the very last path and cut through the aluminum. At this point the whole thing vibrated like crazy and I had to stop the machine. It is not the first time I notice that the path where the bit starts to go through the material causes problems. Does anyone knows why?

forcerouge · July 10, 2017, 5:52pm

Nope, it seems like it is for ER11 systems.
But I could redesign it, no big deal, thanks.

Yeah, for some reason the outside path was a bit better. I don’t really know why.
I don’t think there is any special coating, maybe just a natural oxydation.
The spindle is wery tightly secured, when it moves the gantry moves too.

Tried that yesterday. I was surprised to see that I ended up with similar results.

forcerouge · July 16, 2017, 6:24pm

Gentlemens, welcome back to the shop, today we have a threat especial: the chinesium 3040 milling machine!

[attachment file=38540]

[attachment file=38541]

[attachment file=38542]

[attachment file=38543]

[attachment file=38544]

[attachment file=38545]

[attachment file=38546]

This one is entirely made of aluminum, with linear ball bearings and ball guided screws. Seems pretty rigid, not entirely perfect since the Z axis has a tiny little amount of play if you apply fairly good pressure on it, but it is far less than what I have on the MPCNC.
As many people were wondering how the MPCNC actually competes with other CNC structures, I think it could be instructive to post my new build here, so I can try to make a somewhat objective comparison between both systems I own!

So far, I’ve put all my energy this weekend to make it work, but I haven’t got time to actually try cutting something yet.

I’ve only purchased the chassis. For the electronics, I’m using a similar setup as for the MPCNC: Arduino + ramps + MPCNC firmware. The difference is that I don’t use the tiny crappy drivers, I’m using some beefier TB6560 ones, capable of 3A per phase, in order to drive the bigger steppers. The only issue I see there is that those drivers are limited to 1/16 steps instead of 1/32. I’m wondering if this actually makes a visible difference in terms of cut quality. I hope not because 1/16 makes the machine pretty fast and torquey, it would likely be quite slow and weaker in 1/32. I’ll see if this proves to be a real issue, I always wondered about that. 640 steps per mm seems to be enough, at least in my noob opinion… Makes 0.0015 mm per step, I m’n not sure there should be any real difference going for 1/32 which would make 0.0007mm per step. In my opinion, backlash and other issues will happen far before we hit those numbers. But Maybe someone can correct me if my assumptions are moronic, I’m really no expert.

I reused some steppers I had for other projects, X and Y are pretty normal ones except for the fact that they are rated at 84V, and the Z stepper is some pretty big ass stepper, it was actually the biggest and the most powerful one I could find at that time. Not necessary at all, but I had to use whatever materials I had. right now I supply them with 12V and it works fine, but I’ll try to find a 24-or 48V psu in a near future.

I spent quite a bit of time doing the wiring, and I added limit switches for all axis in order to do homing. I’ll add a probe system later.

It’s ready to work now, I’ve checked and adjusted the spindle squareness, set the proper step/mm values, jerk values, max speed and so on, and it seems to chooches fine.

I’ll probably make the first milling tests this week, as usual I’ll post pictures. I’ll also try to make comparative cuts between both machines, to see which machines performs best at doing what.
Of course, don’t take those as scientific experiences, and keep in myind that my MPCNC is probably a bit different than yours, for instance it is likely of different dimensions, different table structure, even different bearing systems. But anyway, it might be fun and interesting to see.

Cheers!

lionkev55 · July 16, 2017, 6:42pm

Looks kind of skookum to me…

Very curious to see how far you can push it. If for some reason my future electric longboard business plan goes right…I may make something similar to a 3040. That or if Ryan comes up with a brilliant pro version for those seeking better rigidity and features at a price that reflects it

Dui idk if you saw my youtube video but that can be a baseline of what the mpcnc can do if combined with correct cam and setup.

Ordered a 7 dollar tiny vice for my build, I sort of want to try experimenting with a router/mill combo, where most of the build area is taken for a router space, then a small chunk reserved for a vice higher up for milling thick but short stock material.

forcerouge · July 16, 2017, 7:16pm

Just saw that on the other thread, it’s really great!
(That’s right here, for those of you who may not know what we are talking about, you should really check this post: https://www.v1engineering.com/forum/topic/lionkevs-aluminum-attempts/page/2/#post-38463 )

Now, let’s see if a 100% aluminum machine is able to beat a mostly printed one !

I’ll need to buy a few good drill bits, any suggestion about diameter, number of flutes or whatever?
I’ll buy at least 2 of each and test on both machines. I’ll also have to tweak my MPCNC to have similar results as yours otherwise any comparison will be pointless, so I’ll need you to share your settings, or even better, directly send me a gcode test file of your own (if you don’t mind of course).

Again, great job!

lionkev55 · July 16, 2017, 11:17pm

Test file for dui, and anyone else but I wouldn’t use this lol. Careful, this gcode is a bit unpolished. Included is me trying to be professional, a drawing of the “part” For the rectangle, the nominal is 26mmx21mm, but actual is 6mm less (20mmx15mm), as the gcode is a 3mm wide bit.

3mmAluminumTest.gcode (140 KB)

TestPDFDrawing.pdf (107 KB)

forcerouge · July 17, 2017, 5:31pm

Ok, I just bought a set of aluminum bits of various sizes and types:
-one flute
-two flutes
-three flutes
-3 flutes slightly rounded
-ball
-rounded chamfer/fillet tool
-V bit
Almost all in 3mm, some in 4mm

I should be able to make almost any kind of tests with that, just hope that they are good!

vicious1 · July 17, 2017, 8:10pm

I am starting to like the single flutes more and more they work great for everything and seem to have a larger sweet spot.

lionkev55 · July 17, 2017, 9:28pm

It’s true, they do. When plunging holes, the singles are more forgiving because they have so much space for chip evacuation. My dual flute clogs up easier. The chips start to get stuck with plunge drilling but eventually come off. Whats interesting though is that more flutes technically increases your ability to remove material faster.

Seriously though make sure your aluminum if 6061 is the t6 or t6511, unless it will be all gummy and wont machine good at all regardless of what you try.

forcerouge · July 23, 2017, 5:59pm

Ok, so I finally got the 3040 CNC running and made my first tests.

So far, here is my opinion:

The 3040 is way more forgiving than the MPCNC. For some reason I’m still unable to identify, I still have quite a lot of chattering. I’m pretty sure now it is not coming from my drill bits, since they are brand new and designed for aluminum milling. So I suspect it could be the aluminum itself not being of an appropriate grade, or maybe my spindle running too slow. I hope it is not the spindle speed, since I just bought it and it was quite expensive, that would really be annoying. It is advertised at 12 000 rpm, hope this is enough…

But, despite the heavy chattering, I did not have one single failed attempt, and all jobs ended up completed. That would have been impossible with the MPCNC, it would have just given up due to step losses.
So, that’s a very good point for the 3040 here.

[attachment file=39158]

I tried a few bits:
-Single flute: horrible results, crazy chattering, I was afraid the whole thing would explode

[attachment file=39155]

-dual flute: great result, not perfect though

[attachment file=39156]

-3 flutes: similar to dual flute, but a bit more chattering, so dual flute remained the best.

[attachment file=39157]

I’ve only used trochoidal milling so far, at 3 to 3.5 mm per pass for the aluminum, 10mm/s to 20mm/s, and 4% to 10% trochoidal step. Seemed to work equally fine at any of those speed intervals. Maybe the best results were at 4% step and 15mm/s, but the difference was marginal.

So, globally, I’m not satisfied with how the milling goes so far, still way too much noise and chatter, but at least it goes through and the end result after the finishing pass it ok.
I do not consider this as a real test, since I still don’t have a decent working setup, but at least I can attest that, for sure, the 3040 gives more room for error without sacrificing a lot of stock material.

I’ve made 2 parts with it, a bracket for my future computer case, and a bit lenght probe sensor. I finally managed to get the probe to work, no idea why it wasn’t working before…
Now everything is working, so the next step will be to purchase some good quality aluminum and see if this makes any difference.

[attachment file=39159]

[attachment file=39160]

[attachment file=39161]

[attachment file=39162]

lionkev55 · July 23, 2017, 10:33pm

At 12,000rpm, a dual flute is probably the better option. The dw660 spins at a ridiculously fast 30,000 rpm no load. A single flute wouldn’t work as good unless you are cutting something very soft like most plastics. That explains the chatter you got with a single, your chipload was too high at 10mm/s at 12,000rpm.

There is definitely a harmonic sweet spot. When I do full radial passes, sometimes I hear this ringing, almost screeching sound. I wouldn’t call it chatter. Chatter has a lower pitch rumble. At one point, it goes silent and you can tell that the settings and everything were just spot on. All I can hear is the sound of the dewalt like it is cutting through nothing but air. I also think it has something to do with the thickness of your workpiece.

I am going to begin cutting some serious parts out of .25" and 3/8" 6061 t6511, will post my results.

forcerouge · July 30, 2017, 7:45pm

Hi guys,

I’haven’t done much on the MPCNC on last week, (had to rebuild my electric scooter), but I’ll try to do some more experiments this week.

I removed the router motor and put back the printing head, since I had to print a part for a friend. So I’m back to experimenting 3D print again for a while.

I think I may have found a very nice solution to make a heated bed for 3D printing PLA. Found those electric heated floor tiles. Didn’t know those things existed.
The only issue is that, according to the manufacturer, they can only reach about 50 Degree Celsius. not sure it will be enough, but I’ll try and see. I really hope it works, because this solution gives me the opportunity to have a very hard and perfectly flat surface, ideal for the first layer. Plus it is exactly the same size as my mirror plate.

Link here:
https://item.taobao.com/item.htm?spm=a1z0d.6639537.1997196601.4.7be0196DBHKZr&id=549290988230

I’m not sure if there is a way to make this thing hotter if 50 degree is not enough, but I’ll see. At least it surely will be better than nothing, I’ve already printed on my glass when it was less than 15 degree C and it was working relatively ok, so I guess it can only be better at 50…
I’m also worrying that it may take hours to reach 50 degree, since it is given at 135 W per square meter, which is very low, but experience will tell…

Hopefully I should be able to try that this week.

forcerouge · July 30, 2017, 7:49pm

By the way, I think that regular floor tiles could do a great table basis to have a dead flat table for CNC routing. Just put a waste board on top of it and you’re pretty much guaranteed it will be flat and super rigid all over the surface. Plus it will add weight, which is always a good thing here. And you can add several tiles on top of another if one is not enough…

Just a thought, I think it might worth a try since it is a very cheap and possibly very effective solution.

lionkev55 · July 30, 2017, 8:57pm

Good idea, wonder how that will work. Otherwise is it possible to make a heated bed with an aluminum plate, a thermocouple/thermistor, and some sort of heat source? I assume that of course the bed would have to be heated evenly. Maybe multiple thermistors or multiple heaters.

50C probably isn’t enough for PLA, most go for 60C. Better than nothing I suppose.

forcerouge · July 30, 2017, 10:02pm

It’s possible in theory to use a big aluminum plate and stuff, the issue is how to build it, specifically. Finding a good heating source, if possible flat, is an issue, plus the price of a thick aluminum plate may be very expensive. I’m thinking about how to do this heating bed for a while now, and this is the only easily doable solution I came up with.
Other options included:
-Silicon heating pad: very expensive and not entirely flat.
-Some kind of water heated bed: complicated to do, lots of welding and water tightening involved. Complicates the bed leveling
-electric resistors + aluminum plate: expensive since the aluminum plate needs to be thick. Uneven bottom surface, so leveling will be difficult.
-Several 300x300mm PCB heating beds: Expensive, need of many power supplies, having a flat surface will be a nightmare
-Electric floor heating mats: Not flat, not rigid, probably can’t heat more than 50C.
-Heater floor tiles: perfect for everything, except that it can only reach 50C… Actually, if the resistors inside could be more powerful it would be the ideal solution… Dead flat, almost no bending, heavy mass, decent thermal conductivity to the glass since it is polished, lots of heat inertia… Too bad they don’t do those with a bit more power.

So I hope this very cheap solution will work. I know that 50C is probably not enough, but I’ll see first how it works and if it doesn’t I’ll see if it is possible to tune it. No idea how yet, maybe by increasing the voltage. It may be possible also to stack two of those on top of each other and increase the heating property, by insulating the sides and bottom… Not sure.

I’m not sure yet it will work for printing, but I’m pretty sure it can be a great help for milling. Plunging into the material usually bends the table a little bit, (unless maybe if your table is made of super rigid wood) creating vibrations and all sorts of bad things. Having a non bendable material there should really help with this.

barry99705 · July 31, 2017, 3:45am

You’d be surprised at how not flat they really are. Unless you’re using real marble, which may be flat-ish on top, but it won’t be parallel between the sides.

jeffeb3 · July 31, 2017, 3:59am

Is the 50C a logical limit? Is there some temperature feedback?

At any rate, I’ve printed pla at 45C. It will be a lot better than nothing. It will all just depend on the quality of your first layer thickness.

forcerouge · July 31, 2017, 4:23pm

Well, I haven’t received yet so I can’t tell, but I have similar tiles in my living room and they are absolutely flat. At least hundred times more than my current CNC MDF table top… I’ll see once the tile gets there.

I don’t think there is a logical limit, I think it is just a performance limit. So it will also probably depend of the outside temperature.