Reviewing the Shariff DMC2 Mini CNC Mill

Manufacturer: Shariff

Link: https://shariffdmc.com/product/dmc2-mini-cnc/

Price: $2,700 (kit); $4,500 (assembled)

Have you dreamed of a powerful CNC mill at a cost so low it’s crazy?

The DMC2 Mini is a beast of a CNC mill made by Shariff, a small company in Ontario, Canada. This is their second design and boasts 12″×7″×5½” work area, plus 6½ inches of internal Z height so you can use a vise for workholding, like the one that was included in our kit.

Phenomenal specs

Each axis is powered by a beefy 36V, closed-loop, NEMA 23 motor. It has a ferocious 2.2kW 24,000 RPM spindle, a flood cooling system, and a lubrication system for linear rails and ball screws.

I watched it race through a chunk of aluminum, and Shariff says it can handle plastics, tool steels, titanium, and more. And notably, at a cost that blows similar competition out of the water. The specs at this desktop scale are phenomenal, so what’s the catch? Well, you get out of it what you put into it.

Very hands-on

The $2,700 DIY kit is almost $2,000 cheaper than the assembled mill, so that’s the maker move. But that $2K accurately reflects the work required to build it. I put it together with the help of Make: alum Marty Marfin and believe me, when you put together every single part, you learn exactly what every part of the mill does.

Though he has set up machine shops and I’m at least a bit mechanically inclined, neither of us had built a mill before. When I visited him the first day, he was overhauling a rather disassembled SawStop.

Instead of a sand-blasted and anodized clean aluminum enclosure, this tool is unafraid to show off snaking wires and tubes. Organizing them is a project unfortunately at odds with my limited patience for cable management.

Since most of the customers choose to build this mill themselves, we defied common sense and opted to assemble the kit for an honest idea of the experience. Since Make is based in the US, it also made a certain kind of sense in this timeline of unpredictable tariffs.

None of the six boxes of components arrived preassembled. There were holes drilled and tapped in square steel tubes, but every one of the 200+ nuts and bolts had to be turned by hand.

There are 270+ pages of instructions. Thankfully, the documentation is good. There’s room for improvement, but with multiple photos and renders from different angles along the way, I was surprised how few mistakes we made.

If you enjoy assembling kits in your free time, good news: much of the build consists of straightforward, if tedious, bolting things together. I found it meditative. Marty said it had more screws than anything else he’s ever built.

Some parts weren’t packed in the most obvious spot, like the nipples for the oil system, but they installed easily once we had them located.

Altogether the build took us over 40 hours working together. It’s a long, but fulfilling process. Building a mill (as opposed to buying one) is empowering. It breaks down an incredibly complex device into small, comprehensible parts. It’s an education into how machines like this work. And raises the bar for what projects you’re capable of in the future.

Because it has a cooling system, there are parts that need sealing with silicone. I found that step kind of stinky and less glamorous than the mechanical assembly.

Still, there is a pride in building it yourself. The little mistakes we made along the way make this machine different from another DMC2 Mini.

One criticism about the structure is the skinny, spindly legs. The manual warns how fragile they are, and I believe it. They held up during our use, but we treated them very carefully.

There were some odd steps along the way that broke my engineering assumptions. I was surprised that we had to drill a hole in the precision probe — this was to bypass the batteries and wire it to the machine, but it caught me off guard. The inductive sensors come with tiny machine screws, but the build uses M4 screws that are obviously larger than their mounting holes. I have no doubts they can still sense things, but it felt more “renegade” than polished.

I don’t love that some of the wiring was referred to by partial descriptions or descriptions of the packages they were in, instead of labels with consistent names on them. But I didn’t end up cutting the wrong wire at any point.

We did have some confusion when the Z axis wouldn’t move, before tracing it down to a loose wire that I wasn’t explicitly instructed to plug in. There are worse failure modes.

The mill runs on Mach3, although you’re able to replace the controller board if you want to use different software. Ports are labeled to help you if you choose, so getting a new board is a suggested mod. While Mach3 is no longer in development (Mach4 has been out for over 6 years), it still has years of history as an industry standard and is certainly capable of controlling this machine.

Setting up the software is naturally as hands-on as the rest of the machine, with several pages of parameters to check. There was an issue with the E-stop button sending false triggers until the right setting was hunted down and corrected. It’s DIY in every sense. In fact, most of our time reviewing this mill was spent assembling it.

Accuracy depends on the builder. It isn’t the easiest mill to tram. Shariff states that “< 0.01mm Cutting Accuracy” is possible. We spent a few hours and the spindle was still off by more than a degree. I’m sure we could get it tighter with patience (more experienced reviewers have dialed it in further), but we had a print deadline and finite energy.

Tramming is performed by loosening, pushing, and tightening axes until everything is just right. The final step is facing the aluminum bed. One experienced machinist questioned why this was necessary, since the thick bed was already very flat, which was a fair point. Our bed was very flat but was unlevel along the X axis, which, for better or worse, was addressed by facing passes. Having spent time on smaller machines lately, I was shocked at how quickly the 10 mm bit shaved down the bed. Each pass took just a few minutes.

Our experience with the oil pump mirrored that of others — the pump had such high resistance that we couldn’t get oil to flow. Apparently, some pumps from their supplier came with stiffer springs than spec’ed. The solution is to order the correct springs and swap them out.

Still, I was thrilled to see what this thing can do.

Instead of trying to reinvent CAM software, Shariff provides a machine profile, tool libraries, and training videos for Fusion 360. So if you’re using this popular CAM software you’ll be a step ahead. Still, I was testing beginner-focused CAM software just before making my first cut, and had grown accustomed to the conservative default speed and feedrates. So when I set up and ran G-code for our first aluminum cut, we swiftly shaved away material for about 10 seconds before breaking the bit. Attention from the user is suggested.

I wouldn’t call this a beginner machine. But for what it’s worth, neither myself nor Marty had built a mill before and it works. Just be aware of what you’re stepping into.

Pros:

• The cost is very competitive.
• It’s very powerful for the price.
• It takes standard ER20 collets (so up to ½” shank!).
• It’s customizable.

Cons:

• It takes a while to put together.
• Our coolant pump wouldn’t trigger through Mach3. Other reviewers didn’t mention this problem, but I’m tempted to bypass the software and control it using the front panel.
• Safety. The E-stop operates through software and can take several excruciating seconds before halting. I’d make sure a power cutoff switch is within reach. 

A version of this article appears in Make: Vol. 96.