MPCNC with Double Belt Setup


I am new to this form and have built a few CNC Routers stemming form the buildyourcnc world. I discovered the same issue around the holding tabs and think it is from the deflection of the spindle under load. When it stops to lift it routs into the piece as it pulls up.

I was using a CAD/CAM program called CamBAm and it has an option for triangular holding tabs. with these the linear motion is continuous and the deflection stays the same. Also they are easier to cut out. Not sure about your cam software but I bet there is something like that around.


Hello Mike,

Fusion 360 has exactly the same setting and I am going to try this as soon as possible. Thanks a lot!

@Mike Triangular holding tabs really help a lot. Thanks again!


So far I still haven’t found a way to plunge into steel without an excessive amount of chatter. Predrilling the entry positions makes it a lot more enjoyable.

I also switched from a small DOC and (relatively) big WOC to a bigger DOC (5.2mm) and small WOC(0.2mm). This produces a much better surface finish as well as less vibration:


Sorry Ryan, you fixed the video while I was also editing it, but I crashed it again. Could you fix it one more time please.

Just cut and paste the share link, no need for the embed stuff.

[attachment file=“youtube embed.png”]


Wow nice cut settings, that’s 5mm DOC in steel? I need to try steel out soon…


That CamBam program allowed me to do a ramp in on a cut. It would slowly ramp the bit down into the material so that it was not a straight plunge.

Thanks again for the advice Mike, but I already tried that and all the other settings Fusion offers me (helical and “zigzag“ ramp). However I tested a single flute end mill and it ramped into the steel nearly flawlessly. But it is uncoated and I feel like the coating on the other end mill (I think it’s TiAlN) helps a lot. Cutting with the uncoated one creates much more heat and not so good looking chips. A coated single flute end mill would be great, but I can’t find one (for little money :wink: ).

Evan, make sure that your vacuum catches all the steel chips or use an enclosure, those chips are razor sharp. I am very interested in your experiments with different tools on steel and aluminum. Good luck with that.

Glad to hear you’re finding luck with the single flute. I’ve also been searching for coated single flute tools without much luck. All I’ve found is a TiN coated ebay special ($6) which I am guessing is junk grade carbide but I’ll try it anyways, and Datron has a proprietary coating but it’s aimed at cutting bronze.

I’ve been using a vacuum so far that does alright but getting serious chip evacuation issues in 3/8" material, so I bought a compressor and have some loc-line that should arrive today. I need to put some plexi shields around the machine to keep the mess localized.

If we’re having real issues with uncoated tools I can ask my tool guy about getting someone to coat some of the Kyocera ones we use. I don’t know how expensive it will be or what kind of quantities are needed to make it viable. I’ve just been looking through come catalogs & Microcut makes some swanky looking single flutes, they might be willing to do coated specials. Even the uncoated ones are $30-40 though. Hard to beat the cost and quality of the Kyocera tools, all the high end stuff I’m looking at is rather pricey for small tools.

Amazing results, bravo!

If you have trouble with plunging, what you could do is to first use a regular drill bit for making a hole, and then come back with a milling bit to do the rest, starting from this hole.

I think that’s pretty much the only method unless you find some kind of fancy bit.

Actually, I think that’s what he’s doing now, that’s where I got the idea to try it out on my own machine.

I didn’t found much time in the last weeks to work on my MPCNC, so sorry for the lack of updates.


Exactly, except that I can’t use regular drill bits as these are too long for my machine. However single flute end mills seem to work just fine for that.


Atleast I did some progress over the last weeks, the z-axis finally got a cable chain:

[attachment file=75794]


And the electronics found a home in an old computer case:

[attachment file=75792]


The front and back panel are 0,8mm sheet metal, obviously cut on the MPCNC. I did one test cut with thick paper first and then adjusted some distances. The final part fits beautifully:

[attachment file=75793]




I am currently going for the Mafell as Spindle since they optimized it for 71db loudness. Which tool holder did you use? The Mafell is a bit wider than the rest so I feel like most 43mm tool holders might not fit…

thanks in advance



I have operated a CNCrouterparts CNCPro machine milling 6061 aluminium sheet metal (2mm-10mm thick). Our magic formula for success:

  • Amana tools 51373 single 'O' flute cutter 1/8 inch dia, 5/16 inch cutting height, 1/4 inch shank (ToolsToday USA) [!] Highly recommended (am not affiliated with either Amana nor ToolsToday).
  • 18000RPM spindle speed.
  • 1300mm/min (50 inch/min) feedrate. Increase to 1600mm/min (65 inch/min) if stable. Probably want to halve these values for a MPCNC.
  • Too slow feedrates will dull a cutter. Too high feedrates will break a cutter. I'd rather dull a bit over time than break it straight away, so best to go slow at first.
  • Lead in/out feedrate = 50% of feedrate.
  • Plunging feedrate = 25% of feedrate.
  • 2mm depth of cut.
  • 1 or 2 finishing passes.
  • [!] Isopropyl alcohol mister (use a simple spray bottle to continuously spray a mist over the cutter whilst cutting for cooling as aluminium gets very sticky/gummy when hot and tends to weld itself to the cutter).
  • [!] Fusion 360 > 2D Contour operation > Geometry > Select "Tabs" > Triangles instead of rectangles (triangle slope = less plunge = less wear on cutter).
  • [!] Fusion 360 > 2D Contour operation > Passes tab > Select "Passes" > Left (Climb milling). Do not select Right (Conventional milling)
  • [!] Fusion 360 > 2D Contour operation > Linking tab > Select "Ramp" > Ramping angle: 20 degrees.
  • Screw down sheet onto spoilboard around perimeter (50-100mm distance between screws) and in center of some of the larger shapes being cut out as otherwise chips will make their way underneath sheetmetal and lift the sheet causing excessive vibration if not careful.

Hi all, first post here and just to let you know I tried to make Tom Sanladerer aware of this double belt system. I feel like this really could be something good without having any experiences with it.

After doing that I started thinking… the belt that is basically just double sided taped on to the angle iron is flat. It never bends. It is also held by the clamps at each end, right? (Could use a tensioning device)

And the main argument against steel core belts is they don’t like bending back and forth right?

So I was thinking about a total re-design. Why not tension / tape a steel reinforced belt onto a 1 X 1/2” box section. Use 12mm wide belt on the 1/2” side. On the other side you have a set of bearings. Run a 15mm wide motor output pully wheel directly on the top. This is the 4 sides done. Still use the same 25mm stainless steel tube core setup. The box section is actually cheaper then the stainless tubing, fairly smooth and has some weight to it. No belts get bent so can be steel core and tensioned to the max.

Your thoughts welcome… will do a drawing when time permits.

So basically a poor man’s rack and pinion setup? How do you keep it tight to the belt? That’s why we wrap the belt around the gear now, so it can’t slip.

Yes, a bearing or two on the other side should take care of that. As long as the box section is fairly uniform in size it wouldn’t skip on the belt. The last bunch I had in my workshop for a completely different purpose was very smooth.

See the quick and dirty sketch.

With a large pulley I can see this working, but I agree with Barry, at least for a small pulley this might be trouble.

As it happens, I am contemplating alternatives to reduce belt stretch, so this is something I’m interested in.

You might be able to make a little “tank tread” with a short closed-loop belt running inside-out. This is an awful drawing but trying to show the bumpy side is the side with teeth. There would be three idlers. The drive gear has a healthy amount of wrap, and the track has plenty of engagement with the tank tread.

[attachment file=112213]



it took some time, but here’s the latest update of my MPCNC:


[attachment file=112242]

Enclosure made with 20mm aluminum profiles and 4mm acrylic (still with protective cover). Table welded from 50x50x3mm steel.


[attachment file=112244]

Electronics cabinet features two RAMPS-controlled 230V outlets for the spindle and compressor. The RAMPS/Arduino board can be connected to an external laptop or to the included Raspberry Pi which runs cncjs.


[attachment file=112243]

I think adding an air blast is one of the best things one can do, at least if the machine is used mainly for aluminum. It’s really nice to be able to let the machine do its job without babysitting it the whole time and vacuuming chips.
If you look closely at the carriages, you can spot some wipers on the steel rails. I machined a lot of plastic and the chips tend to get stuck on the rails, where they would be compressed by the bearings. This causes the motion to get really rough and left marks in the parts. The wipers kept away the majority of those chips and are essentially just a small plastic part with some cloth on the inside.


As you may have noticed, I removed the double belt. It worked ok for some time, but then the glued belt started to curl up at some places. Additionally, as I have already mentioned in a previous post, it’s not really easy to tension the belt. Too little tension -> loose belt. Too much tension -> the angle iron will be pushed down and the two belts don’t engage each other anymore.
I did try a few different designs, but in the end I just wanted to be able to use the CNC and not tinker with it anymore. Therefore it now has a single 9mm belt. And to be honest, I don’t see any difference to the double belt. Backlash is the same (0,03-0,04mm) and rigidity also didn’t changed noticeably.
I still think that a double belt is an interesting drive mechanism, but the MPCNC just isn’t the right use case. Just look what Bell-Everman is doing with their double belt system.


Just throwing this out there. You want the belt on the bottom. Chips will settle on it otherwise.